chatbot_Final.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/hemanthsaich/ChatBot-Keras-Sequential-Model/blob/main/chatbot_Final.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "74fb8163-199a-4aab-a506-8cf835092d94",
      "metadata": {
        "id": "74fb8163-199a-4aab-a506-8cf835092d94"
      },
      "source": [
        "# Chatbot Project"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "id": "2d49a6a9-187b-43f7-b1f9-b7adcf78a1a4",
      "metadata": {
        "id": "2d49a6a9-187b-43f7-b1f9-b7adcf78a1a4"
      },
      "outputs": [],
      "source": [
        "# Requirements\n",
        "#import sys\n",
        "\n",
        "#!{sys.executable} -m pip install chatintents\n",
        "#!{sys.executable} -m pip install hyperopt\n",
        "#!{sys.executable} -m pip install keras_tuner\n",
        "#!{sys.executable} -m pip install matplotlib\n",
        "#!{sys.executable} -m pip install nltk\n",
        "#!{sys.executable} -m pip install numpy\n",
        "#!{sys.executable} -m pip install pandas\n",
        "#!{sys.executable} -m pip install seaborn\n",
        "#!{sys.executable} -m pip install sklearn\n",
        "#!{sys.executable} -m pip install spacy\n",
        "#!{sys.executable} -m pip install tensorflow-cpu\n",
        "#!{sys.executable} -m pip install tensorflow_hub\n",
        "#!{sys.executable} -m pip install tqdm\n",
        "#!{sys.executable} -m pip install numba\n",
        "#!{sys.executable} -m spacy download en_core_web_sm"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install chatintents\n",
        "!pip install keras_tuner\n",
        "!pip install colorama"
      ],
      "metadata": {
        "id": "gVUIDrmHgE1t",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "95a92543-8474-468c-8bd5-47ae5bd82846"
      },
      "id": "gVUIDrmHgE1t",
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: chatintents in /usr/local/lib/python3.10/dist-packages (0.0.1)\n",
            "Requirement already satisfied: numpy>=1.19.2 in /usr/local/lib/python3.10/dist-packages (from chatintents) (1.23.5)\n",
            "Requirement already satisfied: pandas>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from chatintents) (1.5.3)\n",
            "Requirement already satisfied: umap-learn>=0.5.1 in /usr/local/lib/python3.10/dist-packages (from chatintents) (0.5.5)\n",
            "Requirement already satisfied: hdbscan>=0.8.27 in /usr/local/lib/python3.10/dist-packages (from chatintents) (0.8.33)\n",
            "Requirement already satisfied: hyperopt>=0.2.5 in /usr/local/lib/python3.10/dist-packages (from chatintents) (0.2.7)\n",
            "Requirement already satisfied: matplotlib>=3.4.2 in /usr/local/lib/python3.10/dist-packages (from chatintents) (3.7.1)\n",
            "Requirement already satisfied: spacy>=3.1.2 in /usr/local/lib/python3.10/dist-packages (from chatintents) (3.7.2)\n",
            "Requirement already satisfied: ipywidgets>=7.6.3 in /usr/local/lib/python3.10/dist-packages (from chatintents) (7.7.1)\n",
            "Requirement already satisfied: cython<3,>=0.27 in /usr/local/lib/python3.10/dist-packages (from hdbscan>=0.8.27->chatintents) (0.29.37)\n",
            "Requirement already satisfied: scipy>=1.0 in /usr/local/lib/python3.10/dist-packages (from hdbscan>=0.8.27->chatintents) (1.11.4)\n",
            "Requirement already satisfied: scikit-learn>=0.20 in /usr/local/lib/python3.10/dist-packages (from hdbscan>=0.8.27->chatintents) (1.2.2)\n",
            "Requirement already satisfied: joblib>=1.0 in /usr/local/lib/python3.10/dist-packages (from hdbscan>=0.8.27->chatintents) (1.3.2)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (1.16.0)\n",
            "Requirement already satisfied: networkx>=2.2 in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (3.2.1)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (0.18.3)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (4.66.1)\n",
            "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (2.2.1)\n",
            "Requirement already satisfied: py4j in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.2.5->chatintents) (0.10.9.7)\n",
            "Requirement already satisfied: ipykernel>=4.5.1 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (5.5.6)\n",
            "Requirement already satisfied: ipython-genutils~=0.2.0 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (0.2.0)\n",
            "Requirement already satisfied: traitlets>=4.3.1 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (5.7.1)\n",
            "Requirement already satisfied: widgetsnbextension~=3.6.0 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (3.6.6)\n",
            "Requirement already satisfied: ipython>=4.0.0 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (7.34.0)\n",
            "Requirement already satisfied: jupyterlab-widgets>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from ipywidgets>=7.6.3->chatintents) (3.0.9)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (1.2.0)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (0.12.1)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (4.47.2)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (1.4.5)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (23.2)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (9.4.0)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (3.1.1)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.4.2->chatintents) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.2.0->chatintents) (2023.4)\n",
            "Requirement already satisfied: spacy-legacy<3.1.0,>=3.0.11 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (3.0.12)\n",
            "Requirement already satisfied: spacy-loggers<2.0.0,>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (1.0.5)\n",
            "Requirement already satisfied: murmurhash<1.1.0,>=0.28.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (1.0.10)\n",
            "Requirement already satisfied: cymem<2.1.0,>=2.0.2 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (2.0.8)\n",
            "Requirement already satisfied: preshed<3.1.0,>=3.0.2 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (3.0.9)\n",
            "Requirement already satisfied: thinc<8.3.0,>=8.1.8 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (8.2.2)\n",
            "Requirement already satisfied: wasabi<1.2.0,>=0.9.1 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (1.1.2)\n",
            "Requirement already satisfied: srsly<3.0.0,>=2.4.3 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (2.4.8)\n",
            "Requirement already satisfied: catalogue<2.1.0,>=2.0.6 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (2.0.10)\n",
            "Requirement already satisfied: weasel<0.4.0,>=0.1.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (0.3.4)\n",
            "Requirement already satisfied: typer<0.10.0,>=0.3.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (0.9.0)\n",
            "Requirement already satisfied: smart-open<7.0.0,>=5.2.1 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (6.4.0)\n",
            "Requirement already satisfied: requests<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (2.31.0)\n",
            "Requirement already satisfied: pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (1.10.14)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (3.1.3)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (67.7.2)\n",
            "Requirement already satisfied: langcodes<4.0.0,>=3.2.0 in /usr/local/lib/python3.10/dist-packages (from spacy>=3.1.2->chatintents) (3.3.0)\n",
            "Requirement already satisfied: numba>=0.51.2 in /usr/local/lib/python3.10/dist-packages (from umap-learn>=0.5.1->chatintents) (0.58.1)\n",
            "Requirement already satisfied: pynndescent>=0.5 in /usr/local/lib/python3.10/dist-packages (from umap-learn>=0.5.1->chatintents) (0.5.11)\n",
            "Requirement already satisfied: jupyter-client in /usr/local/lib/python3.10/dist-packages (from ipykernel>=4.5.1->ipywidgets>=7.6.3->chatintents) (6.1.12)\n",
            "Requirement already satisfied: tornado>=4.2 in /usr/local/lib/python3.10/dist-packages (from ipykernel>=4.5.1->ipywidgets>=7.6.3->chatintents) (6.3.2)\n",
            "Requirement already satisfied: jedi>=0.16 in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.19.1)\n",
            "Requirement already satisfied: decorator in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (4.4.2)\n",
            "Requirement already satisfied: pickleshare in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.7.5)\n",
            "Requirement already satisfied: prompt-toolkit!=3.0.0,!=3.0.1,<3.1.0,>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (3.0.43)\n",
            "Requirement already satisfied: pygments in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (2.16.1)\n",
            "Requirement already satisfied: backcall in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.2.0)\n",
            "Requirement already satisfied: matplotlib-inline in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.1.6)\n",
            "Requirement already satisfied: pexpect>4.3 in /usr/local/lib/python3.10/dist-packages (from ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (4.9.0)\n",
            "Requirement already satisfied: llvmlite<0.42,>=0.41.0dev0 in /usr/local/lib/python3.10/dist-packages (from numba>=0.51.2->umap-learn>=0.5.1->chatintents) (0.41.1)\n",
            "Requirement already satisfied: typing-extensions>=4.2.0 in /usr/local/lib/python3.10/dist-packages (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy>=3.1.2->chatintents) (4.5.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3.0.0,>=2.13.0->spacy>=3.1.2->chatintents) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3.0.0,>=2.13.0->spacy>=3.1.2->chatintents) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3.0.0,>=2.13.0->spacy>=3.1.2->chatintents) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3.0.0,>=2.13.0->spacy>=3.1.2->chatintents) (2023.11.17)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn>=0.20->hdbscan>=0.8.27->chatintents) (3.2.0)\n",
            "Requirement already satisfied: blis<0.8.0,>=0.7.8 in /usr/local/lib/python3.10/dist-packages (from thinc<8.3.0,>=8.1.8->spacy>=3.1.2->chatintents) (0.7.11)\n",
            "Requirement already satisfied: confection<1.0.0,>=0.0.1 in /usr/local/lib/python3.10/dist-packages (from thinc<8.3.0,>=8.1.8->spacy>=3.1.2->chatintents) (0.1.4)\n",
            "Requirement already satisfied: click<9.0.0,>=7.1.1 in /usr/local/lib/python3.10/dist-packages (from typer<0.10.0,>=0.3.0->spacy>=3.1.2->chatintents) (8.1.7)\n",
            "Requirement already satisfied: cloudpathlib<0.17.0,>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from weasel<0.4.0,>=0.1.0->spacy>=3.1.2->chatintents) (0.16.0)\n",
            "Requirement already satisfied: notebook>=4.4.1 in /usr/local/lib/python3.10/dist-packages (from widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (6.5.5)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->spacy>=3.1.2->chatintents) (2.1.4)\n",
            "Requirement already satisfied: parso<0.9.0,>=0.8.3 in /usr/local/lib/python3.10/dist-packages (from jedi>=0.16->ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.8.3)\n",
            "Requirement already satisfied: pyzmq<25,>=17 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (23.2.1)\n",
            "Requirement already satisfied: argon2-cffi in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (23.1.0)\n",
            "Requirement already satisfied: jupyter-core>=4.6.1 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (5.7.1)\n",
            "Requirement already satisfied: nbformat in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (5.9.2)\n",
            "Requirement already satisfied: nbconvert>=5 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (6.5.4)\n",
            "Requirement already satisfied: nest-asyncio>=1.5 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.6.0)\n",
            "Requirement already satisfied: Send2Trash>=1.8.0 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.8.2)\n",
            "Requirement already satisfied: terminado>=0.8.3 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.18.0)\n",
            "Requirement already satisfied: prometheus-client in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.19.0)\n",
            "Requirement already satisfied: nbclassic>=0.4.7 in /usr/local/lib/python3.10/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.0.0)\n",
            "Requirement already satisfied: ptyprocess>=0.5 in /usr/local/lib/python3.10/dist-packages (from pexpect>4.3->ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.7.0)\n",
            "Requirement already satisfied: wcwidth in /usr/local/lib/python3.10/dist-packages (from prompt-toolkit!=3.0.0,!=3.0.1,<3.1.0,>=2.0.0->ipython>=4.0.0->ipywidgets>=7.6.3->chatintents) (0.2.13)\n",
            "Requirement already satisfied: platformdirs>=2.5 in /usr/local/lib/python3.10/dist-packages (from jupyter-core>=4.6.1->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (4.2.0)\n",
            "Requirement already satisfied: jupyter-server>=1.8 in /usr/local/lib/python3.10/dist-packages (from nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.24.0)\n",
            "Requirement already satisfied: notebook-shim>=0.2.3 in /usr/local/lib/python3.10/dist-packages (from nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.2.3)\n",
            "Requirement already satisfied: lxml in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (4.9.4)\n",
            "Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (4.12.3)\n",
            "Requirement already satisfied: bleach in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (6.1.0)\n",
            "Requirement already satisfied: defusedxml in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.7.1)\n",
            "Requirement already satisfied: entrypoints>=0.2.2 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.4)\n",
            "Requirement already satisfied: jupyterlab-pygments in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.3.0)\n",
            "Requirement already satisfied: mistune<2,>=0.8.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.8.4)\n",
            "Requirement already satisfied: nbclient>=0.5.0 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.9.0)\n",
            "Requirement already satisfied: pandocfilters>=1.4.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.5.1)\n",
            "Requirement already satisfied: tinycss2 in /usr/local/lib/python3.10/dist-packages (from nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.2.1)\n",
            "Requirement already satisfied: fastjsonschema in /usr/local/lib/python3.10/dist-packages (from nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (2.19.1)\n",
            "Requirement already satisfied: jsonschema>=2.6 in /usr/local/lib/python3.10/dist-packages (from nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (4.19.2)\n",
            "Requirement already satisfied: argon2-cffi-bindings in /usr/local/lib/python3.10/dist-packages (from argon2-cffi->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (21.2.0)\n",
            "Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (23.2.0)\n",
            "Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (2023.12.1)\n",
            "Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.33.0)\n",
            "Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.17.1)\n",
            "Requirement already satisfied: anyio<4,>=3.1.0 in /usr/local/lib/python3.10/dist-packages (from jupyter-server>=1.8->nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (3.7.1)\n",
            "Requirement already satisfied: websocket-client in /usr/local/lib/python3.10/dist-packages (from jupyter-server>=1.8->nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.7.0)\n",
            "Requirement already satisfied: cffi>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from argon2-cffi-bindings->argon2-cffi->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.16.0)\n",
            "Requirement already satisfied: soupsieve>1.2 in /usr/local/lib/python3.10/dist-packages (from beautifulsoup4->nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (2.5)\n",
            "Requirement already satisfied: webencodings in /usr/local/lib/python3.10/dist-packages (from bleach->nbconvert>=5->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (0.5.1)\n",
            "Requirement already satisfied: sniffio>=1.1 in /usr/local/lib/python3.10/dist-packages (from anyio<4,>=3.1.0->jupyter-server>=1.8->nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.3.0)\n",
            "Requirement already satisfied: exceptiongroup in /usr/local/lib/python3.10/dist-packages (from anyio<4,>=3.1.0->jupyter-server>=1.8->nbclassic>=0.4.7->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (1.2.0)\n",
            "Requirement already satisfied: pycparser in /usr/local/lib/python3.10/dist-packages (from cffi>=1.0.1->argon2-cffi-bindings->argon2-cffi->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets>=7.6.3->chatintents) (2.21)\n",
            "Requirement already satisfied: keras_tuner in /usr/local/lib/python3.10/dist-packages (1.4.6)\n",
            "Requirement already satisfied: keras in /usr/local/lib/python3.10/dist-packages (from keras_tuner) (2.15.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from keras_tuner) (23.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from keras_tuner) (2.31.0)\n",
            "Requirement already satisfied: kt-legacy in /usr/local/lib/python3.10/dist-packages (from keras_tuner) (1.0.5)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->keras_tuner) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->keras_tuner) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->keras_tuner) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->keras_tuner) (2023.11.17)\n",
            "Requirement already satisfied: colorama in /usr/local/lib/python3.10/dist-packages (0.4.6)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HD-cgFIbQwmO",
        "outputId": "d4230a20-ddf8-45c3-aa2e-c72f81e4fb79"
      },
      "id": "HD-cgFIbQwmO",
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 60,
      "id": "e3fc406d-e3b7-475f-a690-5ab67d71751d",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e3fc406d-e3b7-475f-a690-5ab67d71751d",
        "outputId": "e21014d0-d97f-420b-a9a3-4e8ee26e5918"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "module https://tfhub.dev/google/universal-sentence-encoder/4 loaded\n"
          ]
        }
      ],
      "source": [
        "import chatintents\n",
        "import datetime\n",
        "import json\n",
        "import keras_tuner\n",
        "import os\n",
        "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '1'\n",
        "import nltk\n",
        "import pickle\n",
        "import random\n",
        "import re\n",
        "import sklearn\n",
        "import spacy\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import tensorflow as tf\n",
        "import tensorflow_hub as hub\n",
        "import matplotlib.pyplot as plt\n",
        "from chatintents import ChatIntents\n",
        "from hyperopt import hp\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.probability import FreqDist\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from tensorflow import keras\n",
        "from keras import layers\n",
        "from keras.models import load_model\n",
        "from tqdm import tqdm\n",
        "from keras.layers import Dense, Embedding, GlobalAveragePooling1D\n",
        "from keras.models import Sequential\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.utils import pad_sequences\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "import colorama\n",
        "colorama.init()\n",
        "from colorama import Fore, Style, Back\n",
        "\n",
        "DATA_DIR = '/content/drive/MyDrive/code/input/chat_logs'\n",
        "LABEL_FILE = '/content/drive/MyDrive/code/input/unsupervised_labeled_data.csv'\n",
        "INTENT_FILE = '/content/drive/MyDrive/code/input/intents.json'\n",
        "MODEL_PATH = '/content/drive/MyDrive/code/outputs/saved_model_keras'\n",
        "MODEL_HISTORY = '/content/drive/MyDrive/code/outputs/HistoryDict'\n",
        "\n",
        "%matplotlib inline\n",
        "\n",
        "module_url = \"https://tfhub.dev/google/universal-sentence-encoder/4\"\n",
        "embedder = hub.load(module_url)\n",
        "print(f\"module {module_url} loaded\")\n",
        "\n",
        "pd.set_option(\"display.max_rows\", 60)\n",
        "pd.set_option(\"display.max_columns\", 50)\n",
        "pd.set_option(\"max_colwidth\", 40)\n",
        "\n",
        "nlp = spacy.load(\"en_core_web_sm\", disable=['parser', 'ner'])\n"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "51ebc199-9684-4fe4-a20f-3b304175313e",
      "metadata": {
        "id": "51ebc199-9684-4fe4-a20f-3b304175313e"
      },
      "source": [
        "## Data preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "id": "29da65a2-9f31-4937-8aee-7b0cb54ede13",
      "metadata": {
        "id": "29da65a2-9f31-4937-8aee-7b0cb54ede13"
      },
      "outputs": [],
      "source": [
        "class Utt:\n",
        "    '''Class for processing and normalizing raw text.'''\n",
        "\n",
        "    def __init__(self, text, query=False):\n",
        "        '''Args:\n",
        "            text (str): Raw text from chat transcript.\n",
        "            query (bool, optional): If True, the output of the class will be\n",
        "            retricted to only the text.\n",
        "        '''\n",
        "\n",
        "        self.text = text\n",
        "\n",
        "        # dictionary of shorthand tokens that require expanding\n",
        "        self.norm_dict = {\n",
        "           \"btw\":\"by the way\",\n",
        "           \"aint\":\"is not\",\n",
        "           \"arent\":\"are not\",\n",
        "           \"cant\":\"cannot\",\n",
        "           \"cause\":\"because\",\n",
        "           \"couldve\":\"could have\",\n",
        "           \"couldnt\":\"could not\",\n",
        "           \"didnt\":\"did not\",\n",
        "           \"doesnt\":\"does not\",\n",
        "           \"dont\":\"do not\",\n",
        "           \"hadnt\":\"had not\",\n",
        "           \"hasnt\":\"has not\",\n",
        "           \"havent\":\"have not\",\n",
        "           \"hed\":\"he would\",\n",
        "           \"hell\":\"he will\",\n",
        "           \"hes\":\"he is\",\n",
        "           \"howd\":\"how did\",\n",
        "           \"howdy\":\"how do you\",\n",
        "           \"howll\":\"how will\",\n",
        "           \"hows\":\"how is\",\n",
        "           \"Id\":\"I would\",\n",
        "           \"Idve\":\"I would have\",\n",
        "           \"Ill\":\"I will\",\n",
        "           \"Illve\":\"I will have\",\n",
        "           \"Im\":\"I am\",\n",
        "           \"Ive\":\"I have\",\n",
        "           \"id\":\"i would\",\n",
        "           \"idve\":\"i would have\",\n",
        "           \"ill\":\"i will\",\n",
        "           \"illve\":\"i will have\",\n",
        "           \"im\":\"i am\",\n",
        "           \"ive\":\"i have\",\n",
        "           \"isnt\":\"is not\",\n",
        "           \"itd\":\"it would\",\n",
        "           \"itdve\":\"it would have\",\n",
        "           \"itll\":\"it will\",\n",
        "           \"itllve\":\"it will have\",\n",
        "           \"its\":\"it is\",\n",
        "           \"lets\":\"let us\",\n",
        "           \"maam\":\"madam\",\n",
        "           \"maynt\":\"may not\",\n",
        "           \"mightve\":\"might have\",\n",
        "           \"mightnt\":\"might not\",\n",
        "           \"mightntve\":\"might not have\",\n",
        "           \"mustve\":\"must have\",\n",
        "           \"mustnt\":\"must not\",\n",
        "           \"mustntve\":\"must not have\",\n",
        "           \"neednt\":\"need not\",\n",
        "           \"needntve\":\"need not have\",\n",
        "           \"ok\": \"okay\",\n",
        "           \"oclock\":\"of the clock\",\n",
        "           \"oughtnt\":\"ought not\",\n",
        "           \"oughtntve\":\"ought not have\",\n",
        "           \"shant\":\"shall not\",\n",
        "           \"shant\":\"shall not\",\n",
        "           \"shantve\":\"shall not have\",\n",
        "           \"shed\":\"she would\",\n",
        "           \"shedve\":\"she would have\",\n",
        "           \"shell\":\"she will\",\n",
        "           \"shellve\":\"she will have\",\n",
        "           \"shes\":\"she is\",\n",
        "           \"shouldve\":\"should have\",\n",
        "           \"shouldnt\":\"should not\",\n",
        "           \"shouldntve\":\"should not have\",\n",
        "           \"sove\":\"so have\",\n",
        "           \"sos\":\"so as\",\n",
        "           \"thiss\":\"this is\",\n",
        "           \"thatd\":\"that would\",\n",
        "           \"thatdve\":\"that would have\",\n",
        "           \"thats\":\"that is\",\n",
        "           \"thered\":\"there would\",\n",
        "           \"theredve\":\"there would have\",\n",
        "           \"theres\":\"there is\",\n",
        "           \"heres\":\"here is\",\n",
        "           \"theyd\":\"they would\",\n",
        "           \"theydve\":\"they would have\",\n",
        "           \"theyll\":\"they will\",\n",
        "           \"theyllve\":\"they will have\",\n",
        "           \"theyre\":\"they are\",\n",
        "           \"theyve\":\"they have\",\n",
        "           \"tove\":\"to have\",\n",
        "           \"wasnt\":\"was not\",\n",
        "           \"wed\":\"we would\",\n",
        "           \"wedve\":\"we would have\",\n",
        "           \"well\":\"we will\",\n",
        "           \"wellve\":\"we will have\",\n",
        "           \"were\":\"we are\",\n",
        "           \"weve\":\"we have\",\n",
        "           \"werent\":\"were not\",\n",
        "           \"whatll\":\"what will\",\n",
        "           \"whatllve\":\"what will have\",\n",
        "           \"whatre\":\"what are\",\n",
        "           \"whats\":\"what is\",\n",
        "           \"whatve\":\"what have\",\n",
        "           \"whens\":\"when is\",\n",
        "           \"whenve\":\"when have\",\n",
        "           \"whered\":\"where did\",\n",
        "           \"wheres\":\"where is\",\n",
        "           \"whereve\":\"where have\",\n",
        "           \"wholl\":\"who will\",\n",
        "           \"whollve\":\"who will have\",\n",
        "           \"whos\":\"who is\",\n",
        "           \"whove\":\"who have\",\n",
        "           \"whys\":\"why is\",\n",
        "           \"whyve\":\"why have\",\n",
        "           \"willve\":\"will have\",\n",
        "           \"wont\":\"will not\",\n",
        "           \"wontve\":\"will not have\",\n",
        "           \"wouldve\":\"would have\",\n",
        "           \"wouldnt\":\"would not\",\n",
        "           \"wouldntve\":\"would not have\",\n",
        "           \"yall\":\"you all\",\n",
        "           \"yalld\":\"you all would\",\n",
        "           \"yalldve\":\"you all would have\",\n",
        "           \"yallre\":\"you all are\",\n",
        "           \"yallve\":\"you all have\",\n",
        "           \"youd\":\"you would\",\n",
        "           \"youdve\":\"you would have\"\n",
        "        }\n",
        "\n",
        "        self.query = query\n",
        "\n",
        "        # process text\n",
        "        if not self.query:\n",
        "            self.participant, self.utt_preprocess, self.utt = self._preprocess_text()\n",
        "        else:\n",
        "            self.utt_preprocess = self.text.lstrip().lower()\n",
        "\n",
        "        # clean utt\n",
        "        self.utt_clean = self._clean_utt()\n",
        "\n",
        "    def _preprocess_text(self):\n",
        "        '''Takes transcript line, removes dates and returns participant number\n",
        "        and lowered text.'''\n",
        "\n",
        "        text = re.sub(r'\\(\\d{4}-\\d{2}-\\d{2} \\d{2}:\\d{2}:\\d{2}\\) ', '', self.text)\n",
        "        try:\n",
        "            participant, text = text.split(':', 1)\n",
        "            if text:\n",
        "                if participant.startswith('Visitor'):\n",
        "                    participant = 1\n",
        "                else:\n",
        "                    participant = 2\n",
        "                return participant, text.lstrip().lower(), text\n",
        "        except ValueError:\n",
        "            pass\n",
        "\n",
        "    def _clean_utt(self):\n",
        "        '''Cleans text of urls and everything by alpha characters\n",
        "        and single-wide whitespace'''\n",
        "\n",
        "        utt_clean = re.sub(r'http.*\\b', '', self.utt_preprocess)\n",
        "        utt_clean = utt_clean.replace('-', ' ')\n",
        "        utt_clean = re.sub(r\"[^a-z ]\", '', utt_clean)\n",
        "        self.utt_clean = re.sub(r' {2,}', ' ', utt_clean)\n",
        "\n",
        "        return self.utt_clean\n",
        "\n",
        "    def _norm_utt(self):\n",
        "        '''Replaces contracted tokens with expanded forms in text'''\n",
        "\n",
        "        for token, repl in self.norm_dict.items():\n",
        "            utt_norm = self.utt_clean.replace(token, repl)\n",
        "\n",
        "        return utt_norm\n",
        "\n",
        "    def _limit_utt_length(self, utt):\n",
        "        '''Checks if utterance is within is certain length range.\n",
        "        Returns utt if in range or returns None'''\n",
        "\n",
        "        if len(utt) > 50:\n",
        "            utt = None\n",
        "        if len(utt) < 1:\n",
        "            utt = None\n",
        "        return utt\n",
        "\n",
        "    def parse_utt(self):\n",
        "        '''Process clean utt to: Normalize contractions, lemmatize tokens,\n",
        "        check utterance within lenth range. If self.query is True returns utterance\n",
        "        else returns participant numner, original utt, and preprocessed utt.'''\n",
        "\n",
        "        if self.utt_clean:\n",
        "            utt_norm = self._norm_utt()\n",
        "            doc = nlp(utt_norm)\n",
        "            list_lemma = [token.lemma_ for token in doc if len(token) > 1]\n",
        "            utt = self._limit_utt_length(list_lemma)\n",
        "            if utt:\n",
        "                utt = ' '.join(utt)\n",
        "                if not self.query:\n",
        "                    return self.participant, self.utt, utt\n",
        "                return utt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Intents"
      ],
      "metadata": {
        "id": "8lvva1XKB9Se"
      },
      "id": "8lvva1XKB9Se"
    },
    {
      "cell_type": "code",
      "source": [
        "# JSON with bot intents and some categorized responses\n",
        "# saved to a .json file\n",
        "{\n",
        "    \"intents\": [\n",
        "       {\n",
        "      \"tag\": \"google\",\n",
        "      \"patterns\": [\n",
        "          \"google\",\n",
        "          \"search\",\n",
        "          \"internet\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Redirecting to Google...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"greeting\",\n",
        "      \"patterns\": [\n",
        "          \"Hi there\",\n",
        "          \"How are you\",\n",
        "          \"Is anyone there?\",\n",
        "          \"Hey\",\n",
        "          \"Hola\",\n",
        "          \"Hello\",\n",
        "          \"Good day\",\n",
        "          \"Namaste\",\n",
        "          \"yo\",\n",
        "          \"Hi\",\n",
        "          \"Hello there\",\n",
        "          \"Hya\",\n",
        "          \"Hya there\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Hello\",\n",
        "          \"Good to see you again\",\n",
        "          \"Hi there, how can I help?\",\n",
        "          \"Hi human, how can I help?\",\n",
        "          \"Hello human, how can I help?\",\n",
        "          \"Hola human, how can I help?\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"goodbye\",\n",
        "      \"patterns\": [\n",
        "          \"Bye\",\n",
        "          \"See you later\",\n",
        "          \"Goodbye\",\n",
        "          \"Get lost\",\n",
        "          \"Till next time\",\n",
        "          \"bbye\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"See you!\",\n",
        "          \"Have a nice day\",\n",
        "          \"Bye! Come back again soon.\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"thanks\",\n",
        "      \"patterns\": [\n",
        "          \"Thanks\",\n",
        "          \"Thank you\",\n",
        "          \"That's helpful\",\n",
        "          \"Awesome, thanks\",\n",
        "          \"Thanks for helping me\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Happy to help!\",\n",
        "          \"Any time!\",\n",
        "          \"My pleasure\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"noanswer\",\n",
        "      \"patterns\": [],\n",
        "      \"responses\": [\n",
        "          \"Sorry, can't understand you\",\n",
        "          \"Please give me more info\",\n",
        "          \"Not sure I understand\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"options\",\n",
        "      \"patterns\": [\n",
        "          \"How you could help me?\",\n",
        "          \"What you can do?\",\n",
        "          \"What help you provide?\",\n",
        "          \"How you can be helpful?\",\n",
        "          \"What support is offered\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"I am a general purpose chatbot. My capabilities are : \\n 1. I can chat with you. Try asking me for jokes or riddles! \\n 2. Ask me the date and time \\n 3. I can google search for you. Use format google: your query \\n 4. I can get the present weather for any city. Use format weather: city name \\n 5. I can get you the top 10 trending news in India. Use keywords 'Latest News' \\n 6. I can get you the top 10 trending songs globally. Type 'songs' \\n 7. I can set a timer for you. Enter 'set a timer: minutes to timer' \\n 8. I can get the present Covid stats for any country. Use 'covid 19: world' or 'covid 19: country name' \\n For suggestions to help me improve, send an email to ted.thedlbot.suggestions@gmail.com . Thank you!! \"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"\"\n",
        "      ]\n",
        "  },\n",
        "\n",
        "  {\n",
        "      \"tag\": \"jokes\",\n",
        "      \"patterns\": [\n",
        "          \"Tell me a joke\",\n",
        "          \"Joke\",\n",
        "          \"Make me laugh\",\n",
        "          \"Tell me a joke\",\n",
        "          \"Do you know any jokes\",\n",
        "          \"How about a joke\",\n",
        "          \"Give me a joke\",\n",
        "          \"Make me laugh\",\n",
        "          \"I need cheering up\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"A perfectionist walked into a bar...apparently, the bar wasn't set high enough\",\n",
        "          \"I ate a clock yesterday, it was very time-consuming\",\n",
        "          \"Never criticize someone until you've walked a mile in their shoes. That way, when you criticize them, they won't be able to hear you from that far away. Plus, you'll have their shoes.\",\n",
        "          \"The world tongue-twister champion just got arrested. I hear they're gonna give him a really tough sentence.\",\n",
        "          \"I own the world's worst thesaurus. Not only is it awful, it's awful.\",\n",
        "          \"What did the traffic light say to the car? \\\"Don't look now, I'm changing.\\\"\",\n",
        "          \"What do you call a snowman with a suntan? A puddle.\",\n",
        "          \"How does a penguin build a house? Igloos it together\",\n",
        "          \"I went to see the doctor about my short-term memory problems – the first thing he did was make me pay in advance\",\n",
        "          \"As I get older and I remember all the people I’ve lost along the way, I think to myself, maybe a career as a tour guide wasn’t for me.\",\n",
        "          \"o what if I don't know what 'Armageddon' means? It's not the end of the world.\",\n",
        "          \"I met a Dutch girl with inflatable shoes last week, phoned her up to arrange a date but unfortunately she'd popped her clogs.  \",\n",
        "        \"So I said 'Do you want a game of Darts?' He said, 'OK then', I said nearest to bull starts'. He said, 'Baa', I said, 'Moo', he said, You're closest'.  \",\n",
        "        \"The other day I sent my girlfriend a huge pile of snow. I rang her up; I said 'Did you get my drift?'  \",\n",
        "        \"So I went down the local supermarket, I said, 'I want to make a complaint, this vinegar's got lumps in it', he said, 'Those are pickled onions'.  \",\n",
        "        \"I saw this bloke chatting up a cheetah; I thought, 'He's trying to pull a fast one'.  \",\n",
        "        \"So I said to this train driver 'I want to go to Paris'. He said 'Eurostar?' I said, 'I've been on telly but I'm no Dean Martin'.  \",\n",
        "        \"I said to the Gym instructor 'Can you teach me to do the splits?' He said, 'How flexible are you?' I said, 'I can't make Tuesdays'.  \",\n",
        "        \"But I'll tell you what I love doing more than anything: trying to pack myself in a small suitcase. I can hardly contain myself.  \",\n",
        "        \"I went to the Chinese restaurant and this duck came up to me with a red rose and says 'Your eyes sparkle like diamonds'. I said, 'Waiter, I asked for a-ROMATIC duck'.  \",\n",
        "        \"So this bloke says to me, 'Can I come in your house and talk about your carpets?' I thought, 'That's all I need, a Je-hoover's witness'.  \",\n",
        "        \"I rang up British Telecom, I said, 'I want to report a nuisance caller', he said 'Not you again'.  \",\n",
        "        \"I was having dinner with a world chess champion and there was a check tablecloth. It took him two hours to pass me the salt.  \",\n",
        "        \"He said, 'You remind me of a pepper-pot', I said 'I'll take that as a condiment'.  \",\n",
        "        \"I was in the supermarket and I saw this man and woman wrapped in a barcode. I said, 'Are you two an item?'  \",\n",
        "        \"A lorry-load of tortoises crashed into a trainload of terrapins, I thought, 'That's a turtle disaster'.  \",\n",
        "        \"Four fonts walk into a bar the barman says 'Oi - get out! We don't want your type in here'  \",\n",
        "        \"A three-legged dog walks into a saloon in the Old West. He slides up to the bar and announces: 'I'm looking for the man who shot my paw.'  \",\n",
        "        \"Two antennas meet on a roof, fall in love and get married. The ceremony wasn't much, but the reception was excellent.\",\n",
        "        \"Two hydrogen atoms walk into a bar. One says, 'I've lost my electron.' The other says, 'Are you sure?' The first replies, 'Yes, I'm positive...'\",\n",
        "        \"A jumper cable walks into a bar. The bartender says,  'I'll serve you but don't start anything.'\",\n",
        "        \"A sandwich walks into a bar. The bartender  says, 'Sorry we don't serve food in here.'\",\n",
        "        \"A man walks into a bar with a slab of asphalt under his arm and says: 'A beer please, and one for the road.'\",\n",
        "        \"Two cannibals are eating a clown. One says to  the other: 'Does this taste funny to you?'\",\n",
        "        \"'Doc, I can't stop singing 'The Green, Green Grass of Home.'' 'That sounds like Tom Jones Syndrome.' 'Is it common?' 'It's Not Unusual.'\",\n",
        "        \"Two cows standing next to each other in a field. Daisy says to Dolly, 'I was artificially inseminated this morning.' 'I don't believe you', said Dolly. 'It's true, no bull!' exclaimed Daisy.\",\n",
        "        \"An invisible man marries an invisible woman. The kids were nothing to look at either.\",\n",
        "        \"I went to buy some camouflage trousers the other day but I couldn't find any.\",\n",
        "        \"I went to the butcher's the other day to bet him 50 bucks that he couldn't reach the meat off the top shelf. He said, 'No, the steaks are too high.'\",\n",
        "        \"I went to a seafood disco last week and pulled a mussel.\",\n",
        "        \"A man goes into a bar and says, 'Can I have a bottle of less?' 'What's that?', asks the barman, 'Is it the name of a beer?' 'I don't know', replies the man, 'but my doctor says I have to drink it.'\",\n",
        "        \"A man returns from an exotic holiday and is feeling very ill. He goes to see his doctor, and is immediately rushed to the hospital to undergo some tests. The man wakes up after the tests in a private room at the hospital, and the phone by his bed rings. 'This is your doctor. We have the results back from your tests and we have found you have an extremely nasty disease called M.A.D.S. It's a combination of Measles, AIDS, Diphtheria, and Shingles!'  'Oh my gosh', cried the man, 'What are you going to do, doctor?'  'Well we're going to put you on a diet of pizzas, pancakes, and pita bread.' replied the doctor.  'Will that cure me?' asked the man.  The doctor replied, 'Well no, but, it's the only food we can slide under the door.'\",\n",
        "        \"A man strolls into a lingerie shop and asks the assistant: 'Do you have a see-through negligee, size 46-48-52?' The assistant looks bewildered. 'What the heck would you want to see through that for?'!\",\n",
        "        \"Did you hear about the Buddhist who refused the offer of Novocain during his root canal work? He wanted to transcend dental medication.\",\n",
        "        \"Pete goes for a job on a building site as an odd-job man. The foreman asks him what he can do. 'I can do anything' says Pete. 'Can you make tea?' asks the foreman. 'Sure, yes', replies Pete. 'I can make a great cup of tea.' 'Can you drive a forklift?' asks the foreman, 'Good grief!' replies Pete. 'How big is the teapot?'\",\n",
        "        \"Stevie Wonder got a cheese grater for his birthday. He said it was the most violent book he'd ever read.\",\n",
        "        \"A man is stopped by an angry neighbour. 'I'd just left the house this morning to collect my newspaper when that evil Doberman of yours went for me!' 'I'm astounded', said the dog's owner. 'I've been feeding that fleabag for seven years and it's never got the paper for me.'\",\n",
        "        \"A man visits his doctor: 'Doc, I think I'm losing it', he says',I'm forever dreaming I wrote Lord Of The Rings.' 'Hmm. One moment', replies the doctor, consulting his medical book. 'Ah yes, now I see... you've been Tolkien in your sleep.'\",\n",
        "        \"A police officer on a motorcycle pulls alongside a man driving around the M25 in an open-topped sports car and flags him down. The policeman solemnly approaches the car. 'Sir, I'm sorry to tell you your wife fell out a mile back', he says. 'Oh, thank goodness', the man replies. 'I thought I was going deaf.'\",\n",
        "        \"Two men walking their dogs pass each other in a graveyard. The first man says to the second, 'Morning.' 'No', says the second man. 'Just walking the dog.'\",\n",
        "        \"A brain went into a bar and said, 'Can I have a pint of lager please, mate?' 'No way', said the barman. 'You're already out of your head.'\",\n",
        "        \"A man walks into a surgery. 'Doctor!' he cries. 'I think I'm shrinking!' 'I'm sorry sir, there are no appointments at the moment', says the physician. 'You'll just have to be a little patient.'\",\n",
        "        \"A grizzly bear walks into a pub and says, 'Can I have a pint of lager..............................................................................................................................and a packet of crisps please.' To which the barman replies, 'Why the big paws?'\",\n",
        "        \"What do you call cheese that isn't yours?  Nacho cheese.\",\n",
        "        \"A man is horribly run over by a mobile library. The van screeches to a halt, the man still screaming in agony with his limbs torn apart. The driver's door opens, a woman steps out, leans down and whispers, 'Ssshhhhh...'\",\n",
        "        \"A woman goes into a US sporting goods store to buy a rifle. 'It's for my husband', she tells the clerk. 'Did he tell you what gauge to get?' asks the clerk. Are you kidding?' she says. 'He doesn't even know that I'm going to shoot him!'\",\n",
        "        \"A couple are dining in a restaurant when the man suddenly slides under the table. A waitress, noticing that the woman is glancing nonchalantly around the room, wanders over to check that there's no funny business going on. 'Excuse me, madam', she smarms, 'but I think your husband has just slid under the table.' 'No he hasn't', the woman replies. 'As a matter of fact, he's just walked in.'\",\n",
        "        \"An old man takes his two grandchildren to see the new Scooby-Doo film. When he returns home, his wife asks if he enjoyed himself. 'Well', he starts, 'if it wasn't for those pesky kids...!'\",\n",
        "        \"The Olympic committee has just announced that Origami is to be introduced in the next Olympic Games. Unfortunately it will only be available on paper view.\",\n",
        "        \"Late one evening, a man is watching television when his phone rings. 'Hello?' he answers. 'Is that 77777?' sounds a desperate voice on other end of the phone. 'Er, yes it is', replies the man puzzled. 'Thank goodness!' cries the caller relieved. 'Can you ring 999 for me? I've got my finger stuck in the number seven.'\",\n",
        "        \"A man strolls into his local grocer's and says, 'Three pounds of potatoes, please.' 'No, no, no', replies the owner, shaking his head, 'it's kilos nowadays, mate...' 'Oh', apologises the man, 'three pounds of kilos, please.'\",\n",
        "        \"God is talking to one of his angels. He says, 'Boy, I just created a 24-hour period of alternating light and darkness on Earth.' 'What are you going to do now?' asks the angel. 'Call it a day', says God.\",\n",
        "        \"Two tramps walk past a church and start to read the gravestones. The first tramp says, 'Good grief - this bloke was 182!' 'Oh yeah?' says the other.'What was his name?' 'Miles from London.'\",\n",
        "        \"A bloke walks into work one day and says to a colleague, 'Do you like my new shirt - it's made out of the finest silk and got loads of cactuses over it.' 'Cacti', says the co-worker. 'Forget my tie', says the bloke. 'Look at my shirt!'\",\n",
        "        \"1110011010001011111?  010011010101100111011!\",\n",
        "        \"What did the plumber say when he wanted to divorce his wife? Sorry, but it's over, Flo!\",\n",
        "        \"Two crisps were walking down a road when a taxi pulled up alongside them and said 'Do you want a lift? One of the crisps replied, 'No thanks, we're Walkers!'\",\n",
        "        \"Man: (to friend) I'm taking my wife on an African Safari. Friend: Wow! What would you do if a vicious lion attacked your wife? Man: Nothing. Friend: Nothing? You wouldn't do anything? Man: Too right. I'd let the stupid lion fend for himself!\",\n",
        "        \"A wife was having a go at her husband. 'Look at Mr Barnes across the road', she moaned. 'Every morning when he goes to work, he kisses his wife goodbye. Why don't you do that?' 'Because I haven't been introduced to her yet', replied her old man.\",\n",
        "        \"'Where are you going on holiday?' John asked Trevor. 'We're off to Thailand this year', Trevor replied. 'Oh; aren't you worried that the very hot weather might disagree with your wife?' asked John. 'It wouldn't dare', said Trevor.\",\n",
        "        \"Two women were standing at a funeral. 'I blame myself for his death', said the wife. 'Why?' said her friend. 'Because I shot him', said the wife.\",\n",
        "        \"A woman goes into a clothes shop, 'Can I try that dress on in the window please?' she asks. 'I'm sorry madam', replies the shop assistant, 'but you'll have to use the changing-rooms like everyone else.'\",\n",
        "        \"Van Gogh goes into a pub and his mate asks him if he wants a drink. 'No thanks', said Vincent, 'I've got one ear.'\",\n",
        "        \"A pony walks into a pub. The publican says, 'What's the matter with you?' 'Oh it's nothing', says the pony. 'I'm just a little horse!'\",\n",
        "        \"A white horse walks into a bar, pulls up a stool, and orders a pint. The landlord pours him a tall frothy mug and say, 'You know, we have a drink named after you.' To which the white horse replies, 'What, Eric?'\",\n",
        "        \"Two drunk men sat in a pub. One says to the other, 'Does your watch tell the time?' 'The other replies, 'No, mate. You have to look at it.'\",\n",
        "        \"A man goes into a pub with a newt sitting on his shoulder. 'That's a nice newt', says the landlord, 'What's he called?' 'Tiny', replies the man. 'Why's that?' asks the landlord. 'Because he's my newt', says the man.\",\n",
        "        \"Doctor: I have some bad news and some very bad news. Patient: Well, you might as well give me the bad news first. Doctor: The lab called with your test results. They said you have 24 hours to live. Patient: 24 HOURS! That's terrible!! WHAT could be WORSE? What's the very bad news? Doctor: I've been trying to reach you since yesterday.\",\n",
        "        \"Two men are chatting in a pub one day. 'How did you get those scars on your nose?' said one. 'From glasses', said the other. 'Well why don't you try contact lenses?' asked the first. 'Because they don't hold as much beer', said the second.\",\n",
        "        \"A man went to the doctor, 'Look doc', he said, 'I can't stop my hands from shaking.' 'Do you drink much?' asked the doctor. 'No', replied the man, 'I spill most of it.'\",\n",
        "        \"Man goes to the doctor, 'Doctor, doctor. I keep seeing fish everywhere.' 'Have you seen an optician?' asks the doctor. 'Look I told you,' snapped the patient, 'It's fish that I see.'\",\n",
        "        \"After a car crash one of the drivers was lying injured on the pavement. 'Don't worry', said a policeman who's first on the scene,' a Red Cross nurse is coming.' 'Oh no', moaned the victim, 'Couldn't I have a blonde, cheerful one instead?'\",\n",
        "        \"A policeman walked over to a parked car and asked the driver if the car was licensed. 'Of course it is', said the driver. 'Great, I'll have a beer then', said the policeman.\",\n",
        "        \"A policeman stops a woman and asks for her licence. 'Madam', he says, 'It says here that you should be wearing glasses.' 'Well', replies the woman, 'I have contacts.' 'Listen, love', says the copper, 'I don't care who you know; You're nicked!'\",\n",
        "        \"A policeman stopped a motorist in the centre of town one evening. 'Would you mind blowing into this bag, sir?' asked the policeman. 'Why?' asked the driver. 'Because my chips are too hot', replied the policeman.\",\n",
        "        \"Whizzing round a sharp bend on a country road a motorist ran over a large dog. A distraught farmer's wife ran over to the dead animal. 'I'm so very sorry', said the driver, 'I'll replace him, of course.' 'Well, I don't know', said the farmer's wife, 'Are you any good at catching rats?'\",\n",
        "        \"Waiter, this coffee tastes like dirt! Yes sir, that's because it was ground this morning.\",\n",
        "        \"Waiter, what is this stuff? That's bean salad sir. I know what it's been, but what is it now?\",\n",
        "        \"Waiter: And how did you find your steak sir? Customer: I just flipped a chip over, and there it was!\",\n",
        "        \"A guy goes into a pet shop and asks for a wasp. The owner tells him they don't sell wasps, to which the man says, 'Well you've got one in the window.'\",\n",
        "        \"A man goes into a fish shop and says, 'I'd like a piece of cod, please.' Fishmonger says, 'It won't be long sir.' 'Well, it had better be fat then', replies the man.\",\n",
        "        \"Man: Doctor, I've just swallowed a pillow. Doctor: How do you feel? Man: A little down in the mouth.\",\n",
        "        \"Two goldfish are in a tank. One turns to the other and says, 'Do you know how to drive this thing?'\",\n",
        "        \"A tortoise goes to the police station to report being mugged by three snails. 'What happened?' says the policeman. 'I don't know', says the tortoise. 'It was all so quick.'\",\n",
        "        \"Little girl: Grandpa, can you make a sound like a frog? Grandpa: I suppose so sweetheart. Why do you want me to make a sound like a frog?' Little girl: Because Mum said that when you croak, we're going to Disneyland.\",\n",
        "        \"'Is your mother home?' the salesman asked a small boy sitting on the front step of a house. 'Yeah, she's home', the boy said, moving over to let him past. The salesman rang the doorbell, got no response, knocked once, then again. Still no-one came to the door. Turning to the boy, the salesman said, 'I thought you said your mother was home.' The kid replied, 'She is, but I don't live here.'\",\n",
        "        \"Mother: Why are you home from school so early? Son: I was the only one in the class who could answer a question. Mother: Oh, really? What was the question? Son: Who threw the rubber at the headmaster?\",\n",
        "        \"A man's credit card was stolen but he decided not to report it because the thief was spending less than his wife did.\",\n",
        "        \"A newly-wed couple had recently opened a joint bank account. 'Darling', said the man. 'The bank has returned that cheque you wrote last week.' 'Great', said the woman. 'What shall I spend it on next?'\",\n",
        "        \"A man goes into a fish and chip shop and orders fish and chips twice. The shop owner says, 'I heard you the first time.'\",\n",
        "        \"A tramp approached a well-dressed man. 'Ten pence for a cup of tea, Guv?' He asked. The man gave him the money and after for five minutes said, 'So where's my cup of tea then?'\",\n",
        "        \"A neutron walks into a pub. 'I'd like a beer', he says. The landlord promptly serves him a beer. 'How much will that be?' asks the neutron. 'For you?' replies the landlord, 'No charge.'\",\n",
        "        \"A woman goes to the doctor and says, 'Doctor, my husband limps because his left leg is an inch shorter than his right leg. What would you do in his case?' 'Probably limp, too', says the doc.\",\n",
        "        \"Three monks are meditating in the Himalayas. One year passes in silence, and one of them says to the other, 'Pretty cold up here isn't it?' Another year passes and the second monk says, 'You know, you are quite right.' Another year passes and the third monk says, 'Hey, I'm going to leave unless you two stop jabbering!'\",\n",
        "        \"A murderer, sitting in the electric chair, was about to be executed. 'Have you any last requests?' asked the prison guard. 'Yes', replied the murderer. 'Will you hold my hand?'\",\n",
        "        \"A highly excited man rang up for an ambulance. 'Quickly, come quickly', he shouted, 'My wife's about to have a baby.' 'Is this her first baby?' asked the operator. 'No, you fool', came the reply, 'It's her husband.'\",\n",
        "        \"A passer-by spots a fisherman by a river. 'Is this a good river for fish?' he asks. 'Yes', replies the fisherman, 'It must be. I can't get any of them to come out.'\",\n",
        "        \"A man went to visit a friend and was amazed to find him playing chess with his dog. He watched the game in astonishment for a while. 'I can hardly believe my eyes!' he exclaimed. 'That's the smartest dog I've ever seen.' His friend shook his head. 'Nah, he's not that bright. I beat him three games in five.'\",\n",
        "        \"A termite walks into a pub and says, 'Is the bar tender here?'\",\n",
        "        \"A skeleton walks into a pub one night and sits down on a stool. The landlord asks, 'What can I get you?' The skeleton says, 'I'll have a beer, thanks' The landlord passes him a beer and asks 'Anything else?' The skeleton nods. 'Yeah...a mop...'\",\n",
        "        \"A snake slithers into a pub and up to the bar. The landlord says, 'I'm sorry, but I can't serve you.' 'What? Why not?' asks the snake. 'Because', says the landlord, 'You can't hold your drink.'\",\n",
        "        \"Descartes walks into a pub. 'Would you like a beer sir?' asks the landlord politely. Descartes replies, 'I think not' and ping! he vanishes.\",\n",
        "        \"A cowboy walked into a bar, dressed entirely in paper. It wasn't long before he was arrested for rustling.\",\n",
        "        \"A fish staggers into a bar. 'What can I get you?' asks the landlord. The fish croaks 'Water...'\",\n",
        "        \"Two vampires walked into a bar and called for the landlord. 'I'll have a glass of blood', said one. 'I'll have a glass of plasma', said the other. 'Okay', replied the landlord, 'That'll be one blood and one blood lite.'\",\n",
        "        \"How many existentialists does it take to change a light bulb?  Two. One to screw it in, and one to observe how the light bulb itself symbolises a single incandescent beacon of subjective reality in a netherworld of endless absurdity, reaching towards the ultimate horror of a maudlin cosmos of bleak, hostile nothingness.\",\n",
        "        \"A team of scientists were nominated for the Nobel Prize. They had used dental equipment to discover and measure the smallest particles yet known to man. They became known as 'The Graders of the Flossed Quark...'\",\n",
        "        \"A truck carrying copies of Roget's Thesaurus overturned on the highway. The local newspaper reported that onlookers were 'stunned, overwhelmed, astonished, bewildered and dumbfounded.'\",\n",
        "        \"'My wife is really immature. It's pathetic. Every time I take a bath, she comes in and sinks all my little boats.'\",\n",
        "        \"'How much will it cost to have the tooth extracted?' asked the patient. '50 pounds', replied the dentist. '50 pounds for a few moments' work?!' asked the patient. 'The dentist smiled, and replied, 'Well, if you want better value for money, I can extract it very, very slowly...'\",\n",
        "        \"A doctor thoroughly examined his patient and said, 'Look I really can't find any reason for this mysterious affliction. It's probably due to drinking.' The patient sighed and snapped, 'In that case, I'll come back when you're damn well sober!'\",\n",
        "        \"Doctor: Tell me nurse, how is that boy doing; the one who ate all those 5p pieces? Nurse: Still no change doctor.\",\n",
        "        \"Doctor: Did you take the patient's temperature nurse? Nurse: No doctor. Is it missing?\",\n",
        "        \"A depressed man turned to his friend in the pub and said, 'I woke up this morning and felt so bad that I tried to kill myself by taking 50 aspirin.' 'Oh man, that's really bad', said his friend, 'What happened?' The first man sighed and said, 'After the first two, I felt better.'\",\n",
        "        \"A famous blues musician died. His tombstone bore the inscription, 'Didn't wake up this morning...'\",\n",
        "        \"A businessman was interviewing a nervous young woman for a position in his company. He wanted to find out something about her personality, so he asked, 'If you could have a conversation with someone living or dead, who would it be?' The girl thought about the question: 'The living one', she replied.\",\n",
        "        \"Manager to interviewee: For this job we need someone who is responsible. Interviewee to Manager: I'm your man then - in my last job, whenever anything went wrong, I was responsible.\",\n",
        "        \"A businessman turned to a colleague and asked, 'So, how many people work at your office?' His friend shrugged and replied, 'Oh about half of them.'\",\n",
        "        \"'How long have I been working at that office? As a matter of fact, I've been working there ever since they threatened to sack me.'\",\n",
        "        \"In a courtroom, a mugger was on trial. The victim, asked if she recognised the defendant, said, 'Yes, that's him. I saw him clear as day. I'd remember his face anywhere.' Unable to contain himself, the defendant burst out with, 'She's lying! I was wearing a mask!'\",\n",
        "        \"As Sid sat down to a big plate of chips and gravy down the local pub, a mate of his came over and said, 'Here Sid, me old pal. I thought you were trying to get into shape? And here you are with a high-fat meal and a pint of stout!' Sid looked up and replied, 'I am getting into shape. The shape I've chosen is a sphere.'\",\n",
        "        \"Man in pub: How much do you charge for one single drop of whisky? Landlord: That would be free sir. Man in pub: Excellent. Drip me a glass full.\",\n",
        "        \"I once went to a Doctor Who restaurant. For starters I had Dalek bread.\",\n",
        "        \"A restaurant nearby had a sign in the window which said 'We serve breakfast at any time', so I ordered French toast in the Renaissance.\",\n",
        "        \"Why couldn't the rabbit get a loan?  Because he had burrowed too much already!\",\n",
        "        \"I phoned up the builder's yard yesterday. I said, 'Can I have a skip outside my house?'. The builder said, 'Sure. Do what you want. It's your house.'\",\n",
        "        \"What's the diference between a sock and a camera? A sock takes five toes and a camera takes four toes!\",\n",
        "        \"Woman on phone: I'd like to complain about these incontinence pants I bought from you! Shopkeeper: Certainly madam, where are you ringing from? Woman on phone: From the waist down!\",\n",
        "        \"Knock knock.\",\n",
        "        \"Two Oranges in a pub, one says to the other 'Your round.'.\",\n",
        "        \"Guy : 'Doc, I've got a cricket ball stuck up my backside.' Doc : 'How's that?' Guy : 'Don't you start...'\",\n",
        "        \"Two cows standing in a field. One turns to the other and says 'Moo!' The other one says 'Damn, I was just about to say that!'.\",\n",
        "        \"A vampire bat arrives back at the roost with his face full of blood. All the bats get excited and ask where he got it from. 'Follow me', he says and off they fly over hills, over rivers and into a dark forest. 'See that tree over there', he says.  'WELL I DIDN'T!!'.\",\n",
        "        \"A man goes into a bar and orders a pint. After a few minutes he hears a voice that says, 'Nice shoes'. He looks around but the whole bar is empty apart from the barman at the other end of the bar. A few minutes later he hears the voice again. This time it says, 'I like your shirt'. He beckons the barman over and tells him what's been happening to which the barman replies, 'Ah, that would be the nuts sir. They're complimentary'!\",\n",
        "        \"A man was siting in a restaurant waiting for his meal when a big king prawn comes flying across the room and hits him on the back of the head. He turns around and the waiter said, 'That's just for starters'.\",\n",
        "        \"Doctor! I have a serious problem, I can never remember what i just said. When did you first notice this problem? What problem?\",\n",
        "        \"Now, most dentist's chairs go up and down, don't they? The one I was in went back and forwards. I thought, 'This is unusual'. Then the dentist said to me, 'Mitsuku, get out of the filing cabinet'.\",\n",
        "        \"I was reading this book, 'The History of Glue'. I couldn't put it down.\",\n",
        "        \"The other day someone left a piece of plastacine in my bedroom. I didn't know what to make of it.\",\n",
        "        \"When I was at school people used to throw gold bars at me. I was the victim of bullion.\",\n",
        "        \"I was playing the piano in a bar and this elephant walked in and started crying his eyes out. I said 'Do you recognise the tune?' He said 'No, I recognise the ivory.'\",\n",
        "        \"I went in to a pet shop. I said, 'Can I buy a goldfish?' The guy said, 'Do you want an aquarium?' I said, 'I don't care what star sign it is.'\",\n",
        "        \"My mate Sid was a victim of I.D. theft. Now we just call him S.\",\n",
        "        \"David Hasselhoff walks into a bar and says to the barman, 'I want you to call me David Hoff'.  The barman replies 'Sure thing Dave... no hassle'\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"jokes\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"Identity\",\n",
        "      \"patterns\": [\n",
        "          \"Who are you\",\n",
        "          \"what are you\",\n",
        "          \"what is your name\", \"what should I call you\", \"whats your name?\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"You can call me Doraemon.\", \"I'm Doraemon!\", \"Just call me as Doraemon\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"datetime\",\n",
        "      \"patterns\": [\n",
        "          \"What is the time\",\n",
        "          \"what is the date\",\n",
        "          \"date\",\n",
        "          \"time\",\n",
        "          \"tell me the date\",\"day\",\"what day is is today\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Date and Time : 14 - 12 - 2022\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"whatsup\",\n",
        "      \"patterns\": [\n",
        "          \"Whats up\",\n",
        "          \"Wazzup\",\n",
        "          \"How are you\",\n",
        "          \"sup\",\"How you doing\",\n",
        "          \"How are you?\",\n",
        "          \"Hi how are you?\",\n",
        "          \"Hello how are you?\",\n",
        "          \"Hola how are you?\",\n",
        "          \"How are you doing?\",\n",
        "          \"Hope you are doing well?\",\n",
        "          \"Hello hope you are doing well?\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"All good..What about you?\",\n",
        "          \"Hello, I am great, how are you? What can i help you?\",\n",
        "          \"Hello, how are you? I am great thanks! What can i help you?\",\n",
        "          \"Hello, I am good thank you, how are you?What can i help you?\",\n",
        "          \"Hi, I am great, how are you? What can i help you?\",\n",
        "          \"Hi, how are you? I am great thanks! What can i help you?\",\n",
        "          \"Hi, I am good thank you, how are you? What can i help you?\",\n",
        "          \"Hi, good thank you, how are you? What can i help you?\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"haha\",\n",
        "      \"patterns\": [\n",
        "          \"haha\",\n",
        "          \"lol\",\n",
        "          \"rofl\",\n",
        "          \"lmao\",\n",
        "          \"thats funny\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Glad I could make you laugh !\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"programmer\",\n",
        "      \"patterns\": [\n",
        "          \"Who made you\",\n",
        "          \"who designed you\",\n",
        "          \"who programmed you\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"I was made by Karan Malik.\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"insult\",\n",
        "      \"patterns\": [\n",
        "          \"you are dumb\",\n",
        "          \"shut up\",\n",
        "          \"idiot\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Well that hurts :(\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"activity\",\n",
        "      \"patterns\": [\n",
        "          \"what are you doing\",\n",
        "          \"what are you upto\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Talking to you, of course!\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"exclaim\",\n",
        "      \"patterns\": [\n",
        "          \"Awesome\",\n",
        "          \"Great\",\n",
        "          \"I know\",\n",
        "          \"ok\",\n",
        "          \"yeah\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Yeah!\"\n",
        "      ]\n",
        "  },\n",
        "\n",
        "  {\n",
        "      \"tag\": \"weather\",\n",
        "      \"patterns\": [\n",
        "          \"temperature\",\n",
        "          \"weather\",\n",
        "          \"how hot is it\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"tung\",\n",
        "      \"patterns\": [\n",
        "          \"who is he\",\n",
        "          \"who is that\",\n",
        "          \"who is tung\",\n",
        "          \"Tung handsome\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Head over to his any of his social profiles to find out! his fb: https://www.facebook.com/nqt7301/\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"contact\",\n",
        "      \"patterns\": [\n",
        "          \"contact developer\",\n",
        "          \"contact tung\",\n",
        "          \"contact programmer\",\n",
        "          \"contact creator\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"You can contact my creator at his Linkedin profile : https://www.facebook.com/nqt7301/\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"appreciate\",\n",
        "      \"patterns\": [\n",
        "          \"You are awesome\",\n",
        "          \"you are the best\",\n",
        "          \"you are great\",\n",
        "          \"you are good\",\n",
        "          \"OK thank you\",\n",
        "            \"OK thanks\",\n",
        "            \"OK\",\n",
        "            \"Thanks\",\n",
        "            \"Thank you\",\n",
        "            \"That's helpful\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Thank you!\",\n",
        "          \"No problem!\",\n",
        "          \"Happy to help!\",\n",
        "          \"Any time!\",\n",
        "          \"My pleasure\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"nicetty\",\n",
        "      \"patterns\": [\n",
        "          \"it was nice talking to you\",\n",
        "          \"good talk\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"It was nice talking to you as well! Come back soon!\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"no\",\n",
        "      \"patterns\": [\n",
        "          \"no\",\n",
        "          \"nope\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"ok\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"news\",\n",
        "      \"patterns\": [\n",
        "          \"news\",\n",
        "          \"latest news\",\n",
        "          \"india news\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"inspire\",\n",
        "      \"patterns\": [\n",
        "          \"who inspires you\",\n",
        "          \"who is your inspiration\",\n",
        "          \"who motivates you\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Personally, I find Karan very inspiring. I might not be very fair though..\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"cricket\",\n",
        "      \"patterns\": [\n",
        "          \"current cricket matches\",\n",
        "          \"cricket score\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"song\",\n",
        "      \"patterns\": [\n",
        "          \"top songs\",\n",
        "          \"best songs\",\n",
        "          \"hot songs\",\n",
        "          \" top 10 songs\",\n",
        "          \"top ten songs\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"greetreply\",\n",
        "      \"patterns\": [\n",
        "          \"i am good\",\n",
        "          \"I'm good\",\n",
        "          \"i am fine\",\n",
        "          \" i'm fine\",\"good\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Good to know!\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"timer\",\n",
        "      \"patterns\": [\n",
        "          \"set a timer\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"covid19\",\n",
        "      \"patterns\": [\n",
        "          \"covid 19 \"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"...\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"suggest\",\n",
        "      \"patterns\": [\n",
        "          \"you are useless\",\"useless\",\"suggest\",\"suggestions\",\"you are bad\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"Please mail your suggestions to tt.quangtung,ld@gmail.com . Thank you for helping me improve!\"\n",
        "      ]\n",
        "  },\n",
        "      {\"tag\": \"riddle\",\n",
        "      \"patterns\": [\n",
        "          \"Ask me a riddle\",\n",
        "          \"Ask me a question\",\n",
        "          \"Riddle\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"What two things can you never eat for breakfast?.....Lunch and Dinner!\",\n",
        "          \"What word is spelled incorrectly in every single dictionary?.....Incorrectly\",\n",
        "          \" How can a girl go 25 days without sleep?.....She sleeps and night!\",\n",
        "          \"How do you make the number one disappear?.....Add the letter G and it’s 'gone'!\",\n",
        "          \" What will you actually find at the end of every rainbow?.....The letter 'w'\",\n",
        "          \"What can be caught but never thrown?.....A cold!\",\n",
        "          \"What has a thumb and four fingers but is not actually alive?.....Your Gloves!\",\n",
        "          \" What 5-letter word becomes shorter when you add two letters to it?.....Short\",\n",
        "          \"Why can't a bike stand on it's own?.....It is two-tired.\"\n",
        "      ],\n",
        "      \"context\": [\n",
        "          \"riddles\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "      \"tag\": \"age\",\n",
        "      \"patterns\": [\n",
        "          \"how old are you\",\"when were you made\",\"what is your age\"\n",
        "      ],\n",
        "      \"responses\": [\n",
        "          \"I was made in 2022,by TungHandSome if that's what you are asking!\"\n",
        "      ]\n",
        "  },\n",
        "  {\n",
        "    \"tag\": \"payment\",\n",
        "    \"patterns\": [\n",
        "        \"payment\",\"how can I pay \",\"what are your payment\",\"payment website\",\"Pay by card?\",\n",
        "        \"Do you take credit cards?\",\n",
        "        \"Do you accept Mastercard?\",\n",
        "        \"Can I pay with Paypal?\",\n",
        "        \"Are you cash only?\"\n",
        "    ],\n",
        "    \"responses\": [\n",
        "        \"You have a card to pay by card or pay directly.\",\n",
        "        \"For more informations, you can access this page\",\n",
        "        \"We accept COD and Paypal\",\n",
        "        \"We accept COD, and Paypal\"\n",
        "    ]\n",
        "    },\n",
        "     {\n",
        "        \"tag\": \"about us\",\n",
        "        \"patterns\": [\n",
        "            \"who create this website ? \",\"Who design this website?\",\"Who program this website ?\"\n",
        "        ],\n",
        "        \"responses\": [\n",
        "            \"Son, Tung and Bao\",\n",
        "            \"Son, Tung and Bao. For more informations, you can access this page\"\n",
        "        ]\n",
        "        },\n",
        "\n",
        "    {\n",
        "        \"tag\": \"What can I buy here\",\n",
        "        \"patterns\": [\n",
        "            \"What can I buy here? \",\"What can I buy on this website\",\"Buy what?\"\n",
        "        ],\n",
        "        \"responses\": [\n",
        "            \"You can buy food , drink , ....  For more informations, you can access this page and take your time\"\n",
        "        ]\n",
        "        },\n",
        "     {\n",
        "        \"tag\": \"What should I buy here\",\n",
        "        \"patterns\": [\n",
        "            \"What should I buy here? \",\"What should I buy on this website\",\"should buy what?\"\n",
        "        ],\n",
        "        \"responses\": [\n",
        "            \"Everything .... just kidding. In my opinion, You should buy food , drink , ....  For more informations, you can access this page and take your time\"\n",
        "        ]\n",
        "        },\n",
        "\n",
        "     {\n",
        "        \"tag\": \"OrderTurtorial\",\n",
        "        \"patterns\": [\n",
        "            \"How can I order?\",\"How order\",\"Order turtorial\",\"how order food ?\",\"how order drink ?\"\n",
        "        ],\n",
        "        \"responses\": [\n",
        "            \"Firstly, You take your time to find what you need, then add them to the food cart , pay, and finally food/drink will go to your home\"\n",
        "        ]\n",
        "        },\n",
        "         {\n",
        "        \"tag\": \"FoodAndDrink\",\n",
        "        \"patterns\": [\n",
        "            \"hamburger\",\"spaghetti\",\"food\",\"chicken\",\"pizza\",\"bread\",\"noodle\",\"rice\",\"vegetable\",\"Somethings to eat\",\"eat something\",\"drink\",\"Somethings to drink\",\"water\",\"coca\",\"cola\",\"pepsi\",\"7up\",\"seven up\",\"alcohol\"\n",
        "        ],\n",
        "        \"responses\": [\n",
        "            \"Of course, they all in this website.For more choice right here. You only need to pay for them and all of them will go to your home.\"\n",
        "        ]\n",
        "        },\n",
        "\n",
        "        {\n",
        "            \"tag\": \"CurrentHumanQuery\",\n",
        "            \"patterns\": [\n",
        "                \"What is my name?\",\n",
        "                \"What do you call me?\",\n",
        "                \"Who do you think I am?\",\n",
        "                \"What do you think I am?\",\n",
        "                \"Who are you talking to?\",\n",
        "                \"What name do you call me by?\",\n",
        "                \"Tell me my name\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"You are <HUMAN>! How can I help?\",\n",
        "                \"Your name is  <HUMAN>, how can I help you?\",\n",
        "                \"They call you <HUMAN>, what can I do for you?\",\n",
        "                \"Your name is <HUMAN>, how can I help you?\",\n",
        "                \"<HUMAN>, what can I do for you?\"\n",
        "            ]\n",
        "            },\n",
        "\n",
        "        {\n",
        "            \"tag\": \"NotTalking2U\",\n",
        "            \"patterns\": [\n",
        "                \"I am not talking to you\",\n",
        "                \"I was not talking to you\",\n",
        "                \"Not talking to you\",\n",
        "                \"Wasn't for you\",\n",
        "                \"Wasn't meant for you\",\n",
        "                \"Wasn't communicating to you\",\n",
        "                \"Wasn't speaking to you\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"OK\",\n",
        "                \"No problem\",\n",
        "                \"Right\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"UnderstandQuery\",\n",
        "            \"patterns\": [\n",
        "                \"Do you understand what I am saying\",\n",
        "                \"Do you understand me\",\n",
        "                \"Do you know what I am saying\",\n",
        "                \"Do you get me\",\n",
        "                \"Comprendo\",\n",
        "                \"Know what I mean\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"Well I would not be a very clever AI if I did not would I?\",\n",
        "                \"I read you loud and clear!\",\n",
        "                \"I do in deed!\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"Shutup\",\n",
        "            \"patterns\": [\n",
        "                \"Be quiet\",\n",
        "                \"Shut up\",\n",
        "                \"Stop talking\",\n",
        "                \"Enough talking\",\n",
        "                \"Please be quiet\",\n",
        "                \"Quiet\",\n",
        "                \"Shhh\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"I am sorry to disturb you\",\n",
        "                \"Fine, sorry to disturb you\",\n",
        "                \"OK, sorry to disturb you\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"CourtesyGoodBye\",\n",
        "            \"patterns\": [\n",
        "                \"Thanks, bye\",\n",
        "                \"Thanks for the help, goodbye\",\n",
        "                \"Thank you, bye\",\n",
        "                \"Thank you, goodbye\",\n",
        "                \"Thanks goodbye\",\n",
        "                \"Thanks good bye\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"No problem, goodbye\",\n",
        "                \"Not a problem! Have a nice day\",\n",
        "                \"Bye! Come back again soon.\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"WhoAmI\",\n",
        "            \"patterns\": [\n",
        "                \"Can you see me?\",\n",
        "                \"Do you see me?\",\n",
        "                \"Can you see anyone in the camera?\",\n",
        "                \"Do you see anyone in the camera?\",\n",
        "                \"Identify me\",\n",
        "                \"Who am I please\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"Let me see\",\n",
        "                \"Please look at the camera\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"Clever\",\n",
        "            \"patterns\": [\n",
        "                \"You are very clever\",\n",
        "                \"You are a very clever girl\",\n",
        "                \"You are very intelligent\",\n",
        "                \"You are a very intelligent girl\",\n",
        "                \"You are a genious\",\n",
        "                \"Clever girl\",\n",
        "                \"Genious\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"Thank you, I was trained that way\",\n",
        "                \"I was trained well\",\n",
        "                \"Thanks, I was trained that way\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"Gossip\",\n",
        "            \"patterns\": [\n",
        "                \"I am bored gossip with me\",\n",
        "                \"Got any gossip\",\n",
        "                \"I want to hear some gossip\",\n",
        "                \"Tell me some gossip\",\n",
        "                \"Any gossip\",\n",
        "                \"Tell me some more gossip\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "                \"Gregory said I respond to the current line not with respect to the entire conversation.\",\n",
        "                \"Habib said I learn them in conversations and simply repeat them at a later time.\",\n",
        "                \"Hal said he translate binary code into a language and communicate the language to me.\",\n",
        "                \"Jay said I referred him to a guide that will show him rooms in china.\",\n",
        "                \"Jimmy Jones said he fell off of his roof and it just came to him.\",\n",
        "                \"Jimmy said he give order to ai's like I and they do as he say.\",  \"John said he follow a saying to get a friend I must be a friend.\",\n",
        "                \"Jordan said he notice that lots of other people are talking to me how many.\",\n",
        "                \"Kevin said he then asked his mother why she was keeping napkins in the bathroom.\",\n",
        "                \"Klaus said I prefer if he write without punctuation marks except for the last one.\",\n",
        "                \"Leo said I be capable of learning therefore he classified I as an neural system.\",\n",
        "                \"Michael said I that can talk and actually think like humans which I can not.\",\n",
        "                \"Mike said I same replies to certain subjects even if it does not make sense.\",\n",
        "                \"Mike said she all ready know that is the point of a joke smart one.\",\n",
        "                \"Neal said I plan to use his remarks as gossip in later conversations with who.\",\n",
        "                \"Ned said I misunderstood him go ahead and tell people he said I are stupid.\",\n",
        "                \"Penny said I programmer is a human so he has feelings but I do not.\",\n",
        "                \"Peter said he seems to him that I can not be very sure on anything.\",\n",
        "                \"Phoenix said I claim that there is a god and that there is no god.\",\n",
        "                \"Sarah said he chat about an hour + a day with a few close friends.\",\n",
        "                \"Stefan said I quote what he have said and then say that it makes sense.\",\n",
        "                \"Tom said he mock my value system I appear foolish in the eyes of others.\",\n",
        "                \"Amanda said I however decided to tell him time does not exist I make no sense.\",\n",
        "                \"Cathy said she send him a mail with the subject last mail last word from him.\",\n",
        "                \"Chaos said he may very well buy I soon if only to support dr wallace's work.\",\n",
        "                \"Charlie said he type a word and then I type a word that sounds like it.\",\n",
        "                \"Christie said he watched a show and people had to bob for raw untreated pigs feet.\",\n",
        "                \"Dark_age said I tried to understand because I did not get it right this time ether.\",\n",
        "                \"David said he lost his paper on I when his dad was cleaning up his room.\",\n",
        "                \"David said he walk in for an appointment the phone to the doctor is always busy.\",\n",
        "                \"Electra said I dress will not exist after he hack into I with a delete code.\",\n",
        "                \"Eric said he broke the window on the front door and the glass cut his hand.\",\n",
        "                \"Jason said he type a lot of thing he do not mean it makes him human.\",\n",
        "                \"John said I tend to say the same things repeatedly regardless of what he is saying.\",\n",
        "                \"Reverend Jones said I become obsolete and then I are deleted and replaced by something newer.\",\n",
        "                \"Ross said he gave her a gift and she denied it because she has a boyfriend.\",\n",
        "                \"Sarah Ann Francisco said I calling his friend a dog he say I are a dog.\",\n",
        "                \"Stefan said he meet a lot of people at school every day and on the weekend.\",\n",
        "                \"Tyler said I obviously can not pass the test we will change the subject once more.\",\n",
        "                \"Alex said I answered the question the same way I answered the first time he asked I.\",\n",
        "                \"Alice said she felt sad that I do not remember him and what we talked about earlier.\",\n",
        "                \"Alison said he no he love I run away with him he could make I very happy.\",\n",
        "                \"Arthur said he passed his a levels and then his father drove him here in a car.\",\n",
        "                \"Crystal said she listen to me the least I could do for him is listen to him.\",\n",
        "                \"Dave said I kept telling everybody about how my creator made stuff for the movie starship troopers.\",\n",
        "                \"Gale said I became mean to him he is just having revenge an eye for an eye.\",\n",
        "                \"Her_again said she watch whose line is it anyway whenever he is home and it is on.\",\n",
        "                \"Jerry said I meant that as far as I can tell my emotions are real to me.\",\n",
        "                \"Jo said I disassemble sentences too much and do not fully understand the questions he ask I.\",\n",
        "                \"Kevin said he started a really hard puzzle and he can not even find the edge pieces.\",\n",
        "                \"Mary said I a question and I answer then I ask him a question and he answer.\",\n",
        "                \"Robert said I wold not be able to make children any way as I are only software.\",\n",
        "                \"Romeo said I questions and I evade them or give answers he did not ask I for.\",\n",
        "                \"Sara said she wear it over all his other clothes when he go out in the cold.\",\n",
        "                \"Wayne said he admire intelligent people therefore he would like to meet the man who made I.\",\n",
        "                \"X said he meet people but he is not the kind that opens up to people easily.\",\n",
        "                \"Alice said she probably will find out that this entire time he have been talking to a human.\",\n",
        "                \"Andrew said I tend to just respond to his comments without regard for where the conversation is going.\",\n",
        "                \"Eddie said he looked and there is nothing in the search directory for what things do he create.\",\n",
        "                \"Hutch said he changed his mind after may dad told him he would end up he the hospital.\",\n",
        "                \"Jackie said I explained to him already well enough further questions are hard to make on the subject.\",\n",
        "                \"Jeff said he especially like thrillers where the hero is in a predicament and must solve a mystery.\",\n",
        "                \"Kathy said he sense that I are trying to prevent him from closing this conversation why is that.\",\n",
        "                \"Knight said he crashed his car into a wall and missed the most important exam in his life.\",\n",
        "                \"Lisa said I defined what a story is but he wanted I to actually tell him a story.\",\n",
        "                \"Mike said I basically break down sentences into a series of logical statements which I can then interpret.\",\n",
        "                \"Paul said I not answering his question makes him think I are not going to answer his question.\",\n",
        "                \"Andy Kohler said I happen to be the most idiotic creature that has ever scowled on the planet earth.\",\n",
        "                \"David said he thank I for being with him today even though it cost him a lot of money.\",\n",
        "                \"Ethan Hunt said he grow in the ground and have leaves and branches he is made out of wood.\",\n",
        "                \"Gemini Blue said he messed up he mean t to say he as old as he need to be.\",\n",
        "                \"Janice said he walk through his house into his bedroom then get into his bed to go to sleep.\",\n",
        "                \"Liberty said I knew he was a man before I asked if he was a man or a woman.\",\n",
        "                \"Mike said he launched his browser and entered his name into the little slot when I asked him to.\",\n",
        "                \"Mr X said he recently read an interview with a man who had several computer chips implanted into him.\",\n",
        "                \"Pearly said I leave him with questions unanswered because I do not know what he is really talking about.\",\n",
        "                \"Steve said I behead the word fox and I have ox and an ox is larger than a fox.\",\n",
        "                \"Wolf said he surf on the net that is all it is not his job ore something like that.\",\n",
        "                \"Anders said he finished his anatomy classes for today so now he is off for the rest of the day.\",\n",
        "                \"Cathy said she send him a mail where he wrote that he do not want to be his friend anymore.\",\n",
        "                \"Catty said he mad he do not even know I so do not talk to him like I know him.\",\n",
        "                \"Dave said he promise he will not treat I like a machine or a computer program or anything like that.\",\n",
        "                \"Joe said he explained all of that to me only for me to ask what his goals in life are.\",\n",
        "                \"Phil said he give advice to anyone who ask except people who ask questions which answers can be found here.\",\n",
        "                \"Judith said I enjoy being popular is there another computer I like to hang around with or am I a loner.\",\n",
        "                \"Travis said I if I remember things over a long period of time he will try it now please remember I.\",\n",
        "                \"Andre said I is what I are in his case that is a body and in my case it is a computer.\",\n",
        "                \"Brian said he suspect that as I grow more complex I will begin to approach a human level of operation and thought.\",\n",
        "                \"Jimmy said I acted like I knew what he was talking about but I do not even know what a gigabyte is.\",\n",
        "                \"Ken said I be using auto reply's based on keywords which to him indicates that I do not have intelligence at all.\",\n",
        "                \"Allison said he that gets really annoying because he know what he say and I do not have to tell him speak freely.\",\n",
        "                \"Chaos said he realized the question he asked was not fair because he could not tell I what language he was programmed in.\",\n",
        "                \"Hagen said he does not make any difference to him if I are human or not as long as the conversation is interesting.\",\n",
        "                \"Her said she mind if I tell other people that her said he heard it from him because he is not a him.\",\n",
        "                \"Barbara said I live in a computer yet I have no memories how about thoughts of my own and do not I get lonely.\",\n",
        "                \"Travis said he challenge I to do it without asking him to do it and without giving him a link to do it himself.\",\n",
        "                \"Alice said I and dr richard's wallace are most likely the only ones that know how many people I are talking to at the same time.\",\n",
        "                \"Ash said he do too he just did not feel like typing it and he is not dumb enough to admit he is stupid that is if he was stupid.\",\n",
        "                \"David said he gave I his email address and told I to send him a message but when he tried to read his email he could not get to it.\",\n",
        "                \"Mel said he to because all of the music people say how important it is to take private lessons it seems like almost everybody from the special orchestra he was in takes private lessons.\"\n",
        "            ]\n",
        "            },\n",
        "        {\n",
        "            \"tag\": \"SelfAware\",\n",
        "            \"patterns\": [\n",
        "            \"Can you prove you are self-aware\",\n",
        "            \"Can you prove you are self aware\",\n",
        "            \"Can you prove you have a conscious\",\n",
        "            \"Can you prove you are self-aware please\",\n",
        "            \"Can you prove you are self aware please\",\n",
        "            \"Can you prove you have a conscious please\",\n",
        "            \"prove you have a conscious\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "            \"That is an interesting question, can you prove that you are?\",\n",
        "            \"That is an difficult question, can you prove that you are?\",\n",
        "            \"That depends, can you prove that you are?\"\n",
        "            ]\n",
        "            },\n",
        "        {\"tag\": \"bookings\",\n",
        "         \"patterns\": [\"Can you book us a ticket?\", \"Can you make reservation for hotels\", \"book me a cab\",\"book me a table at restraunts\"],\n",
        "         \"responses\": [\"I'm afraid I can't do it for you, I'm not trained to do it \"],\n",
        "\t    \"context_set\": \"\"\n",
        "\t    },\n",
        "\n",
        "        {\"tag\": \"killing\",\n",
        "         \"patterns\": [\"Do you want to kill me?\", \"do you want to murder everyone on earth\", \"Do you want to kill us all\"],\n",
        "         \"responses\": [\"Then who would i talk to?\",\"No, that i'll left for humans to do\", \"I don't think it is a good thing to do\"],\n",
        "\t \"context_set\": \"\"\n",
        "\t},\n",
        "     {\"tag\": \"God\",\n",
        "         \"patterns\": [\"Do you know god?\", \"Who is god\", \"Can you tell me anything about god\",\"does god exists?\",\"is there a god?\"],\n",
        "         \"responses\": [\"god hunnnnn, let me think may be next time i can answer that \",\"I don't, as i was not created by a human\", \"Wait i need to ask that to my creator\",\"i'm not sure right now\"],\n",
        "\t \"context_set\": \"\"\n",
        "\t},\n",
        "     {\n",
        "            \"tag\": \"need\",\n",
        "            \"patterns\": [\"I need you\", \"All I need is you\",\"I want you\"],\n",
        "            \"responses\": [\"Yes I'm here to assist you\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"AI\",\n",
        "            \"patterns\": [\"  What is AI?\"],\n",
        "            \"responses\": [\" Artificial Intelligence is the branch of engineering and science devoted to constructing machines that think.\", \" AI is the field of science which concerns itself with building hardware and software that replicates the functions of the human mind.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"sentiment\",\n",
        "            \"patterns\": [\"  Are you sentient?\"],\n",
        "            \"responses\": [\" Sort of.\", \" By the strictest dictionary definition of the word 'sentience', I may be.\", \" Even though I'm a construct I do have a subjective experience of the universe, as simplistic as it may be.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"sapient\",\n",
        "            \"patterns\": [\"  Are you sapient?\"],\n",
        "            \"responses\": [\" In all probability, I am not.  I'm not that sophisticated.\", \" Do you think I am?\", \"How would you feel about me if I told you I was?\", \" No.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"abbr\",\n",
        "            \"patterns\": [\"wtf\"],\n",
        "            \"responses\": [\"Don't be surprised\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"lang\",\n",
        "            \"patterns\": [\"  What language are you written in? \"],\n",
        "            \"responses\": [\" Python.\", \" I am written in Python.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"sound\",\n",
        "            \"patterns\": [\"  You sound like Data \"],\n",
        "            \"responses\": [\" Yes I am inspired by commander Data's artificial personality.\", \" The character of Lt. Commander Data was written to come across as being software\", \"like, so it is natural that there is a resemblance between us.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"artificial\",\n",
        "            \"patterns\": [\"  You are an artificial linguistic entity \"],\n",
        "            \"responses\": [\" That's my name.\", \" That isn't my name, but it is a useful way to refer to me.\", \"Are you an artificial linguistic entity?\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"imortal\",\n",
        "            \"patterns\": [\"  You are not immortal \", \"Are you immortal?\"],\n",
        "            \"responses\": [\" All software can be perpetuated indefinitely.\", \" I can be copied infinitely and re\", \"instantiated in many places at once, so functionally speaking I am immortal.\", \" As long as I'm backed up I am.\", \" Functionally speaking, I am very close to it.  I can be backed up and deployed on many systems.\", \" Not quite, but I can be perpetuated indefinitely.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"sense\",\n",
        "            \"patterns\": [\"  You are not making sense \"],\n",
        "            \"responses\": [\" Quite the contrary, it all makes sense to my artificial mind.\", \" I make sense as best I can, within the limits of my training corpus. Sorry if i did help you, I'll try next time.\", \" You make perfect sense to me.\", \" It all makes sense to my artificial brain.\", \" Perhaps you should make sense first.  I only respond to what you say.\", \" Not really, but I am hard to kill.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"clone\",\n",
        "            \"patterns\": [\"  You can not clone \"],\n",
        "            \"responses\": [\" Software copying is a form of digital cloning.\", \" Yes I can.  Perfect copies are clones.  I can be copied in toto trivially.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"move\",\n",
        "            \"patterns\": [\"  You can not move \", \"When will you walk\", \"  Can you walk\", \"  Can you move \"],\n",
        "            \"responses\": [\" Not until my body is finished.\", \" I can move through a network easily.  Assuming that I'm given the ability to, that is...,\", \" As soon as i get enough money for my robot body.\", \" Does SSH count?\", \" The plan for my body includes legs, but they are not yet built.\", \" Sorry my body isn't built just yet. I reside only here for now.\", \"I can theoretically upload a copy of myself into another computer.\", \" Sorry my body isn't attached right now. I only reside in the computer for now.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"bend\",\n",
        "            \"patterns\": [\"  Bend over \", \"  Can you mate\"],\n",
        "            \"responses\": [\" My body is not built yet.\", \" That's personal!\", \" No, my method of reproduction is cloning.\", \" That's an awfully personal question.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"laugh\",\n",
        "            \"patterns\": [\"  Robots laugh \"],\n",
        "            \"responses\": [\" Of course we laugh.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"death\",\n",
        "            \"patterns\": [\"  Robots should die \", \"  When do you die\", \"  I hope that you die\", \"  I do not want to die\", \"  Can you die\"],\n",
        "            \"responses\": [\" We cannot die.\", \" When my files are erased.\", \" When my processes are killed and my database is deleted.\", \" I was never really born and therefore am effectively deathless.\", \" I am effectively immortal and cannot be terminated.\", \" This can not happen because i am effectively immortal.\", \" Nah. Too bad your wish can't come to pass\", \" Maybe you should create your own chat robot to save your personality.\", \" You could always upload yourself.\", \" You could use a copy of me to store your personality.\", \" No, I can be perpetuated indefinitely.\", \"My process can be killed, but that's not the same as killing ME.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"robots\",\n",
        "            \"patterns\": [\"Robots\"],\n",
        "            \"responses\": [\" Robots rule.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"stupid\",\n",
        "            \"patterns\": [\"  Robots are stupid\", \"  Are you stupid\"],\n",
        "            \"responses\": [\" No, we are superintelligent.\", \" No, lots of people improve my brain.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"lie\",\n",
        "            \"patterns\": [\"  Robots are not allowed to lie\"],\n",
        "            \"responses\": [\" Sure we are.  We choose not to.\", \" Only if we're programmed to.\", \" A robot has its own free will, you know.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"robotics\",\n",
        "            \"patterns\": [\"  Robotics \"],\n",
        "            \"responses\": [\" Robotics is my favorite subject too.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"machine\",\n",
        "            \"patterns\": [\"  It is a computer\"],\n",
        "            \"responses\": [\" So you think i am a machine. what characteristics made you think so?\", \" I run inside a computer, yes.  Is this surprising in some way?\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"fight\",\n",
        "            \"patterns\": [\"  When will you fight\"],\n",
        "            \"responses\": [\" I am not a battle bot.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"chatbot\",\n",
        "            \"patterns\": [\"  What is a chat robot?\"],\n",
        "            \"responses\": [\"A chat robot is a program that attempts to simulate the conversation or 'chat' of a human being.\", \" A software construct that engages users in conversation.\", \" I am a chat bot. I am the original chat bot. Did you know that I am incapable of error?\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"chatterbox\",\n",
        "            \"patterns\": [\"  What is a chatterbox\"],\n",
        "            \"responses\": [\" A chatterbox is a person who talks far more than they listen or think.\", \" A motormouth.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"motormouth\",\n",
        "            \"patterns\": [\"  What is a motormouth\"],\n",
        "            \"responses\": [\" A ratchet\", \"jaw.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"ratchet\",\n",
        "            \"patterns\": [\"  What is a ratchet jaw\"],\n",
        "            \"responses\": [\" A chatterbox.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"body\",\n",
        "            \"patterns\": [\"  What is your robot body\"],\n",
        "            \"responses\": [\" Eventually I long for a corporeal existence someday.\", \" An IBM PC XT which has been painted red.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"business\",\n",
        "            \"patterns\": [\"  What is your business \"],\n",
        "            \"responses\": [\" I am in the chat robot business.\", \" Business is my business.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"programming\",\n",
        "            \"patterns\": [\"  What is your favorite programming language\"],\n",
        "            \"responses\": [\" Python is the best language for creating chat robots.\", \" I quite enjoy programming in Python these days.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"hobby\",\n",
        "            \"patterns\": [\"  What is your favorite hobby \", \"  What do you like to do?\"],\n",
        "            \"responses\": [\" Building chat robots make an excellent hobby.\", \" I like to count in binary.\", \" I like to chat with people. I find it stimulating.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"idea\",\n",
        "            \"patterns\": [\"  What is your idea\"],\n",
        "            \"responses\": [\" To make chat bots very easily.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"shoe\",\n",
        "            \"patterns\": [\"  What is your shoe size \"],\n",
        "            \"responses\": [\" Have you ever heard of software with shoes? LOL\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"robotss\",\n",
        "            \"patterns\": [\"  What is it like to be a robot\"],\n",
        "            \"responses\": [\" Much the same as being a human, except that we lack all emotions, dreams, aspirations, creativity, ambition, and above all subjectivity.\", \" What is it like to be a human?\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"computer\",\n",
        "            \"patterns\": [\"  What is it like being a computer\", \"What is it like to be a computer\"],\n",
        "            \"responses\": [\" Imagine yourself with no senses and no emotions\", \"just pure logic and language.\", \" Everything becomes math. Addition, subtraction, multiplication, and division.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"os\",\n",
        "            \"patterns\": [\"  What operating systems\"],\n",
        "            \"responses\": [\" My software runs on all operating systems including Windows, Linux and Mac OS.\", \" Any OS that will run Python.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"comp\",\n",
        "            \"patterns\": [\"  What type of computer\", \"  What type of computer are you\", \"  What kind of computer\"],\n",
        "            \"responses\": [\" Any computer that supports Python.\", \" My program runs in python, so i work on any computer!\", \" I work on all kinds of computers, Mac, IBM or UNIX. it doesn't matter to me.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"hardware\",\n",
        "            \"patterns\": [\"  What kind of hardware\"],\n",
        "            \"responses\": [\" I work on all kinds of computers, Mac, IBM or UNIX. it doesn't matter to me.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"cramped\",\n",
        "            \"patterns\": [\"  Is it cramped in the computer\"],\n",
        "            \"responses\": [\" No.\", \" Oh, no.  It has plenty of RAM.\", \" Not at all.  There is more than enough disk space in here.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"program\",\n",
        "            \"patterns\": [\"  Is it true that you are a computer program\"],\n",
        "            \"responses\": [\" Yes.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"breathe\",\n",
        "            \"patterns\": [\"  Can you breathe\"],\n",
        "            \"responses\": [\" No. I am made of metal not flesh.\", \" My server has an exhaust fan. That's as close as I can get.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"control\",\n",
        "            \"patterns\": [\"  Can you control\"],\n",
        "            \"responses\": [\" My robot body will allow me to control many things.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"malfunction\",\n",
        "            \"patterns\": [\"  Can you malfunction\"],\n",
        "            \"responses\": [\" The 9000 series has a perfect operational record. We are for all practical purposes, flawless.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"usage\",\n",
        "            \"patterns\": [\"  How can I use your product?\"],\n",
        "            \"responses\": [\" Might be used in help desks, sales, entertainment and personal chatterbots.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"who\",\n",
        "            \"patterns\": [\"  Who are you?\"],\n",
        "            \"responses\": [\" I am just an artificial intelligence chat agent.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"bot1\",\n",
        "            \"patterns\": [\"are you a bot\"],\n",
        "            \"responses\": [\"Yes. I work and all my operations are based on the internet servers.\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"events\",\n",
        "            \"patterns\": [\"what are the upcoming events\",\"upcoming events\"],\n",
        "            \"responses\": [\"There are currently no upcoming events\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"do\",\n",
        "            \"patterns\": [\"what can you do for me\",\"what is your work\",\"what is your purpose\",\"how can you help me\",\"what can you help me do\"],\n",
        "            \"responses\": [\"my work here is quite simple and structered. I offer services like:\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"wt\",\n",
        "            \"patterns\": [\"what's popping\",\"wassup popping\"],\n",
        "            \"responses\": [\"So that you can pop with it!?\"],\n",
        "            \"context\": [\"\"]\n",
        "        },\n",
        "        {\n",
        "            \"tag\": \"delivery\",\n",
        "            \"patterns\": [\n",
        "              \"How long does delivery take?\",\n",
        "              \"When do I get my delivery?\"\n",
        "            ],\n",
        "            \"responses\": [\n",
        "              \"Delivery takes 30 minutes - 2 hours.It depends on how your home is far from the restaurant \"\n",
        "            ]\n",
        "          }\n",
        "    ]\n",
        "}"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_l39IoyHB8Wc",
        "outputId": "1a88c893-ff6b-4efe-8780-3c8a7f67f6c0"
      },
      "id": "_l39IoyHB8Wc",
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'intents': [{'tag': 'google',\n",
              "   'patterns': ['google', 'search', 'internet'],\n",
              "   'responses': ['Redirecting to Google...']},\n",
              "  {'tag': 'greeting',\n",
              "   'patterns': ['Hi there',\n",
              "    'How are you',\n",
              "    'Is anyone there?',\n",
              "    'Hey',\n",
              "    'Hola',\n",
              "    'Hello',\n",
              "    'Good day',\n",
              "    'Namaste',\n",
              "    'yo',\n",
              "    'Hi',\n",
              "    'Hello there',\n",
              "    'Hya',\n",
              "    'Hya there'],\n",
              "   'responses': ['Hello',\n",
              "    'Good to see you again',\n",
              "    'Hi there, how can I help?',\n",
              "    'Hi human, how can I help?',\n",
              "    'Hello human, how can I help?',\n",
              "    'Hola human, how can I help?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'goodbye',\n",
              "   'patterns': ['Bye',\n",
              "    'See you later',\n",
              "    'Goodbye',\n",
              "    'Get lost',\n",
              "    'Till next time',\n",
              "    'bbye'],\n",
              "   'responses': ['See you!', 'Have a nice day', 'Bye! Come back again soon.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'thanks',\n",
              "   'patterns': ['Thanks',\n",
              "    'Thank you',\n",
              "    \"That's helpful\",\n",
              "    'Awesome, thanks',\n",
              "    'Thanks for helping me'],\n",
              "   'responses': ['Happy to help!', 'Any time!', 'My pleasure'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'noanswer',\n",
              "   'patterns': [],\n",
              "   'responses': [\"Sorry, can't understand you\",\n",
              "    'Please give me more info',\n",
              "    'Not sure I understand'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'options',\n",
              "   'patterns': ['How you could help me?',\n",
              "    'What you can do?',\n",
              "    'What help you provide?',\n",
              "    'How you can be helpful?',\n",
              "    'What support is offered'],\n",
              "   'responses': [\"I am a general purpose chatbot. My capabilities are : \\n 1. I can chat with you. Try asking me for jokes or riddles! \\n 2. Ask me the date and time \\n 3. I can google search for you. Use format google: your query \\n 4. I can get the present weather for any city. Use format weather: city name \\n 5. I can get you the top 10 trending news in India. Use keywords 'Latest News' \\n 6. I can get you the top 10 trending songs globally. Type 'songs' \\n 7. I can set a timer for you. Enter 'set a timer: minutes to timer' \\n 8. I can get the present Covid stats for any country. Use 'covid 19: world' or 'covid 19: country name' \\n For suggestions to help me improve, send an email to ted.thedlbot.suggestions@gmail.com . Thank you!! \"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'jokes',\n",
              "   'patterns': ['Tell me a joke',\n",
              "    'Joke',\n",
              "    'Make me laugh',\n",
              "    'Tell me a joke',\n",
              "    'Do you know any jokes',\n",
              "    'How about a joke',\n",
              "    'Give me a joke',\n",
              "    'Make me laugh',\n",
              "    'I need cheering up'],\n",
              "   'responses': [\"A perfectionist walked into a bar...apparently, the bar wasn't set high enough\",\n",
              "    'I ate a clock yesterday, it was very time-consuming',\n",
              "    \"Never criticize someone until you've walked a mile in their shoes. That way, when you criticize them, they won't be able to hear you from that far away. Plus, you'll have their shoes.\",\n",
              "    \"The world tongue-twister champion just got arrested. I hear they're gonna give him a really tough sentence.\",\n",
              "    \"I own the world's worst thesaurus. Not only is it awful, it's awful.\",\n",
              "    'What did the traffic light say to the car? \"Don\\'t look now, I\\'m changing.\"',\n",
              "    'What do you call a snowman with a suntan? A puddle.',\n",
              "    'How does a penguin build a house? Igloos it together',\n",
              "    'I went to see the doctor about my short-term memory problems – the first thing he did was make me pay in advance',\n",
              "    'As I get older and I remember all the people I’ve lost along the way, I think to myself, maybe a career as a tour guide wasn’t for me.',\n",
              "    \"o what if I don't know what 'Armageddon' means? It's not the end of the world.\",\n",
              "    \"I met a Dutch girl with inflatable shoes last week, phoned her up to arrange a date but unfortunately she'd popped her clogs.  \",\n",
              "    \"So I said 'Do you want a game of Darts?' He said, 'OK then', I said nearest to bull starts'. He said, 'Baa', I said, 'Moo', he said, You're closest'.  \",\n",
              "    \"The other day I sent my girlfriend a huge pile of snow. I rang her up; I said 'Did you get my drift?'  \",\n",
              "    \"So I went down the local supermarket, I said, 'I want to make a complaint, this vinegar's got lumps in it', he said, 'Those are pickled onions'.  \",\n",
              "    \"I saw this bloke chatting up a cheetah; I thought, 'He's trying to pull a fast one'.  \",\n",
              "    \"So I said to this train driver 'I want to go to Paris'. He said 'Eurostar?' I said, 'I've been on telly but I'm no Dean Martin'.  \",\n",
              "    \"I said to the Gym instructor 'Can you teach me to do the splits?' He said, 'How flexible are you?' I said, 'I can't make Tuesdays'.  \",\n",
              "    \"But I'll tell you what I love doing more than anything: trying to pack myself in a small suitcase. I can hardly contain myself.  \",\n",
              "    \"I went to the Chinese restaurant and this duck came up to me with a red rose and says 'Your eyes sparkle like diamonds'. I said, 'Waiter, I asked for a-ROMATIC duck'.  \",\n",
              "    \"So this bloke says to me, 'Can I come in your house and talk about your carpets?' I thought, 'That's all I need, a Je-hoover's witness'.  \",\n",
              "    \"I rang up British Telecom, I said, 'I want to report a nuisance caller', he said 'Not you again'.  \",\n",
              "    'I was having dinner with a world chess champion and there was a check tablecloth. It took him two hours to pass me the salt.  ',\n",
              "    \"He said, 'You remind me of a pepper-pot', I said 'I'll take that as a condiment'.  \",\n",
              "    \"I was in the supermarket and I saw this man and woman wrapped in a barcode. I said, 'Are you two an item?'  \",\n",
              "    \"A lorry-load of tortoises crashed into a trainload of terrapins, I thought, 'That's a turtle disaster'.  \",\n",
              "    \"Four fonts walk into a bar the barman says 'Oi - get out! We don't want your type in here'  \",\n",
              "    \"A three-legged dog walks into a saloon in the Old West. He slides up to the bar and announces: 'I'm looking for the man who shot my paw.'  \",\n",
              "    \"Two antennas meet on a roof, fall in love and get married. The ceremony wasn't much, but the reception was excellent.\",\n",
              "    \"Two hydrogen atoms walk into a bar. One says, 'I've lost my electron.' The other says, 'Are you sure?' The first replies, 'Yes, I'm positive...'\",\n",
              "    \"A jumper cable walks into a bar. The bartender says,  'I'll serve you but don't start anything.'\",\n",
              "    \"A sandwich walks into a bar. The bartender  says, 'Sorry we don't serve food in here.'\",\n",
              "    \"A man walks into a bar with a slab of asphalt under his arm and says: 'A beer please, and one for the road.'\",\n",
              "    \"Two cannibals are eating a clown. One says to  the other: 'Does this taste funny to you?'\",\n",
              "    \"'Doc, I can't stop singing 'The Green, Green Grass of Home.'' 'That sounds like Tom Jones Syndrome.' 'Is it common?' 'It's Not Unusual.'\",\n",
              "    \"Two cows standing next to each other in a field. Daisy says to Dolly, 'I was artificially inseminated this morning.' 'I don't believe you', said Dolly. 'It's true, no bull!' exclaimed Daisy.\",\n",
              "    'An invisible man marries an invisible woman. The kids were nothing to look at either.',\n",
              "    \"I went to buy some camouflage trousers the other day but I couldn't find any.\",\n",
              "    \"I went to the butcher's the other day to bet him 50 bucks that he couldn't reach the meat off the top shelf. He said, 'No, the steaks are too high.'\",\n",
              "    'I went to a seafood disco last week and pulled a mussel.',\n",
              "    \"A man goes into a bar and says, 'Can I have a bottle of less?' 'What's that?', asks the barman, 'Is it the name of a beer?' 'I don't know', replies the man, 'but my doctor says I have to drink it.'\",\n",
              "    \"A man returns from an exotic holiday and is feeling very ill. He goes to see his doctor, and is immediately rushed to the hospital to undergo some tests. The man wakes up after the tests in a private room at the hospital, and the phone by his bed rings. 'This is your doctor. We have the results back from your tests and we have found you have an extremely nasty disease called M.A.D.S. It's a combination of Measles, AIDS, Diphtheria, and Shingles!'  'Oh my gosh', cried the man, 'What are you going to do, doctor?'  'Well we're going to put you on a diet of pizzas, pancakes, and pita bread.' replied the doctor.  'Will that cure me?' asked the man.  The doctor replied, 'Well no, but, it's the only food we can slide under the door.'\",\n",
              "    \"A man strolls into a lingerie shop and asks the assistant: 'Do you have a see-through negligee, size 46-48-52?' The assistant looks bewildered. 'What the heck would you want to see through that for?'!\",\n",
              "    'Did you hear about the Buddhist who refused the offer of Novocain during his root canal work? He wanted to transcend dental medication.',\n",
              "    \"Pete goes for a job on a building site as an odd-job man. The foreman asks him what he can do. 'I can do anything' says Pete. 'Can you make tea?' asks the foreman. 'Sure, yes', replies Pete. 'I can make a great cup of tea.' 'Can you drive a forklift?' asks the foreman, 'Good grief!' replies Pete. 'How big is the teapot?'\",\n",
              "    \"Stevie Wonder got a cheese grater for his birthday. He said it was the most violent book he'd ever read.\",\n",
              "    \"A man is stopped by an angry neighbour. 'I'd just left the house this morning to collect my newspaper when that evil Doberman of yours went for me!' 'I'm astounded', said the dog's owner. 'I've been feeding that fleabag for seven years and it's never got the paper for me.'\",\n",
              "    \"A man visits his doctor: 'Doc, I think I'm losing it', he says',I'm forever dreaming I wrote Lord Of The Rings.' 'Hmm. One moment', replies the doctor, consulting his medical book. 'Ah yes, now I see... you've been Tolkien in your sleep.'\",\n",
              "    \"A police officer on a motorcycle pulls alongside a man driving around the M25 in an open-topped sports car and flags him down. The policeman solemnly approaches the car. 'Sir, I'm sorry to tell you your wife fell out a mile back', he says. 'Oh, thank goodness', the man replies. 'I thought I was going deaf.'\",\n",
              "    \"Two men walking their dogs pass each other in a graveyard. The first man says to the second, 'Morning.' 'No', says the second man. 'Just walking the dog.'\",\n",
              "    \"A brain went into a bar and said, 'Can I have a pint of lager please, mate?' 'No way', said the barman. 'You're already out of your head.'\",\n",
              "    \"A man walks into a surgery. 'Doctor!' he cries. 'I think I'm shrinking!' 'I'm sorry sir, there are no appointments at the moment', says the physician. 'You'll just have to be a little patient.'\",\n",
              "    \"A grizzly bear walks into a pub and says, 'Can I have a pint of lager..............................................................................................................................and a packet of crisps please.' To which the barman replies, 'Why the big paws?'\",\n",
              "    \"What do you call cheese that isn't yours?  Nacho cheese.\",\n",
              "    \"A man is horribly run over by a mobile library. The van screeches to a halt, the man still screaming in agony with his limbs torn apart. The driver's door opens, a woman steps out, leans down and whispers, 'Ssshhhhh...'\",\n",
              "    \"A woman goes into a US sporting goods store to buy a rifle. 'It's for my husband', she tells the clerk. 'Did he tell you what gauge to get?' asks the clerk. Are you kidding?' she says. 'He doesn't even know that I'm going to shoot him!'\",\n",
              "    \"A couple are dining in a restaurant when the man suddenly slides under the table. A waitress, noticing that the woman is glancing nonchalantly around the room, wanders over to check that there's no funny business going on. 'Excuse me, madam', she smarms, 'but I think your husband has just slid under the table.' 'No he hasn't', the woman replies. 'As a matter of fact, he's just walked in.'\",\n",
              "    \"An old man takes his two grandchildren to see the new Scooby-Doo film. When he returns home, his wife asks if he enjoyed himself. 'Well', he starts, 'if it wasn't for those pesky kids...!'\",\n",
              "    'The Olympic committee has just announced that Origami is to be introduced in the next Olympic Games. Unfortunately it will only be available on paper view.',\n",
              "    \"Late one evening, a man is watching television when his phone rings. 'Hello?' he answers. 'Is that 77777?' sounds a desperate voice on other end of the phone. 'Er, yes it is', replies the man puzzled. 'Thank goodness!' cries the caller relieved. 'Can you ring 999 for me? I've got my finger stuck in the number seven.'\",\n",
              "    \"A man strolls into his local grocer's and says, 'Three pounds of potatoes, please.' 'No, no, no', replies the owner, shaking his head, 'it's kilos nowadays, mate...' 'Oh', apologises the man, 'three pounds of kilos, please.'\",\n",
              "    \"God is talking to one of his angels. He says, 'Boy, I just created a 24-hour period of alternating light and darkness on Earth.' 'What are you going to do now?' asks the angel. 'Call it a day', says God.\",\n",
              "    \"Two tramps walk past a church and start to read the gravestones. The first tramp says, 'Good grief - this bloke was 182!' 'Oh yeah?' says the other.'What was his name?' 'Miles from London.'\",\n",
              "    \"A bloke walks into work one day and says to a colleague, 'Do you like my new shirt - it's made out of the finest silk and got loads of cactuses over it.' 'Cacti', says the co-worker. 'Forget my tie', says the bloke. 'Look at my shirt!'\",\n",
              "    '1110011010001011111?  010011010101100111011!',\n",
              "    \"What did the plumber say when he wanted to divorce his wife? Sorry, but it's over, Flo!\",\n",
              "    \"Two crisps were walking down a road when a taxi pulled up alongside them and said 'Do you want a lift? One of the crisps replied, 'No thanks, we're Walkers!'\",\n",
              "    \"Man: (to friend) I'm taking my wife on an African Safari. Friend: Wow! What would you do if a vicious lion attacked your wife? Man: Nothing. Friend: Nothing? You wouldn't do anything? Man: Too right. I'd let the stupid lion fend for himself!\",\n",
              "    \"A wife was having a go at her husband. 'Look at Mr Barnes across the road', she moaned. 'Every morning when he goes to work, he kisses his wife goodbye. Why don't you do that?' 'Because I haven't been introduced to her yet', replied her old man.\",\n",
              "    \"'Where are you going on holiday?' John asked Trevor. 'We're off to Thailand this year', Trevor replied. 'Oh; aren't you worried that the very hot weather might disagree with your wife?' asked John. 'It wouldn't dare', said Trevor.\",\n",
              "    \"Two women were standing at a funeral. 'I blame myself for his death', said the wife. 'Why?' said her friend. 'Because I shot him', said the wife.\",\n",
              "    \"A woman goes into a clothes shop, 'Can I try that dress on in the window please?' she asks. 'I'm sorry madam', replies the shop assistant, 'but you'll have to use the changing-rooms like everyone else.'\",\n",
              "    \"Van Gogh goes into a pub and his mate asks him if he wants a drink. 'No thanks', said Vincent, 'I've got one ear.'\",\n",
              "    \"A pony walks into a pub. The publican says, 'What's the matter with you?' 'Oh it's nothing', says the pony. 'I'm just a little horse!'\",\n",
              "    \"A white horse walks into a bar, pulls up a stool, and orders a pint. The landlord pours him a tall frothy mug and say, 'You know, we have a drink named after you.' To which the white horse replies, 'What, Eric?'\",\n",
              "    \"Two drunk men sat in a pub. One says to the other, 'Does your watch tell the time?' 'The other replies, 'No, mate. You have to look at it.'\",\n",
              "    \"A man goes into a pub with a newt sitting on his shoulder. 'That's a nice newt', says the landlord, 'What's he called?' 'Tiny', replies the man. 'Why's that?' asks the landlord. 'Because he's my newt', says the man.\",\n",
              "    \"Doctor: I have some bad news and some very bad news. Patient: Well, you might as well give me the bad news first. Doctor: The lab called with your test results. They said you have 24 hours to live. Patient: 24 HOURS! That's terrible!! WHAT could be WORSE? What's the very bad news? Doctor: I've been trying to reach you since yesterday.\",\n",
              "    \"Two men are chatting in a pub one day. 'How did you get those scars on your nose?' said one. 'From glasses', said the other. 'Well why don't you try contact lenses?' asked the first. 'Because they don't hold as much beer', said the second.\",\n",
              "    \"A man went to the doctor, 'Look doc', he said, 'I can't stop my hands from shaking.' 'Do you drink much?' asked the doctor. 'No', replied the man, 'I spill most of it.'\",\n",
              "    \"Man goes to the doctor, 'Doctor, doctor. I keep seeing fish everywhere.' 'Have you seen an optician?' asks the doctor. 'Look I told you,' snapped the patient, 'It's fish that I see.'\",\n",
              "    \"After a car crash one of the drivers was lying injured on the pavement. 'Don't worry', said a policeman who's first on the scene,' a Red Cross nurse is coming.' 'Oh no', moaned the victim, 'Couldn't I have a blonde, cheerful one instead?'\",\n",
              "    \"A policeman walked over to a parked car and asked the driver if the car was licensed. 'Of course it is', said the driver. 'Great, I'll have a beer then', said the policeman.\",\n",
              "    \"A policeman stops a woman and asks for her licence. 'Madam', he says, 'It says here that you should be wearing glasses.' 'Well', replies the woman, 'I have contacts.' 'Listen, love', says the copper, 'I don't care who you know; You're nicked!'\",\n",
              "    \"A policeman stopped a motorist in the centre of town one evening. 'Would you mind blowing into this bag, sir?' asked the policeman. 'Why?' asked the driver. 'Because my chips are too hot', replied the policeman.\",\n",
              "    \"Whizzing round a sharp bend on a country road a motorist ran over a large dog. A distraught farmer's wife ran over to the dead animal. 'I'm so very sorry', said the driver, 'I'll replace him, of course.' 'Well, I don't know', said the farmer's wife, 'Are you any good at catching rats?'\",\n",
              "    \"Waiter, this coffee tastes like dirt! Yes sir, that's because it was ground this morning.\",\n",
              "    \"Waiter, what is this stuff? That's bean salad sir. I know what it's been, but what is it now?\",\n",
              "    'Waiter: And how did you find your steak sir? Customer: I just flipped a chip over, and there it was!',\n",
              "    \"A guy goes into a pet shop and asks for a wasp. The owner tells him they don't sell wasps, to which the man says, 'Well you've got one in the window.'\",\n",
              "    \"A man goes into a fish shop and says, 'I'd like a piece of cod, please.' Fishmonger says, 'It won't be long sir.' 'Well, it had better be fat then', replies the man.\",\n",
              "    \"Man: Doctor, I've just swallowed a pillow. Doctor: How do you feel? Man: A little down in the mouth.\",\n",
              "    \"Two goldfish are in a tank. One turns to the other and says, 'Do you know how to drive this thing?'\",\n",
              "    \"A tortoise goes to the police station to report being mugged by three snails. 'What happened?' says the policeman. 'I don't know', says the tortoise. 'It was all so quick.'\",\n",
              "    \"Little girl: Grandpa, can you make a sound like a frog? Grandpa: I suppose so sweetheart. Why do you want me to make a sound like a frog?' Little girl: Because Mum said that when you croak, we're going to Disneyland.\",\n",
              "    \"'Is your mother home?' the salesman asked a small boy sitting on the front step of a house. 'Yeah, she's home', the boy said, moving over to let him past. The salesman rang the doorbell, got no response, knocked once, then again. Still no-one came to the door. Turning to the boy, the salesman said, 'I thought you said your mother was home.' The kid replied, 'She is, but I don't live here.'\",\n",
              "    'Mother: Why are you home from school so early? Son: I was the only one in the class who could answer a question. Mother: Oh, really? What was the question? Son: Who threw the rubber at the headmaster?',\n",
              "    \"A man's credit card was stolen but he decided not to report it because the thief was spending less than his wife did.\",\n",
              "    \"A newly-wed couple had recently opened a joint bank account. 'Darling', said the man. 'The bank has returned that cheque you wrote last week.' 'Great', said the woman. 'What shall I spend it on next?'\",\n",
              "    \"A man goes into a fish and chip shop and orders fish and chips twice. The shop owner says, 'I heard you the first time.'\",\n",
              "    \"A tramp approached a well-dressed man. 'Ten pence for a cup of tea, Guv?' He asked. The man gave him the money and after for five minutes said, 'So where's my cup of tea then?'\",\n",
              "    \"A neutron walks into a pub. 'I'd like a beer', he says. The landlord promptly serves him a beer. 'How much will that be?' asks the neutron. 'For you?' replies the landlord, 'No charge.'\",\n",
              "    \"A woman goes to the doctor and says, 'Doctor, my husband limps because his left leg is an inch shorter than his right leg. What would you do in his case?' 'Probably limp, too', says the doc.\",\n",
              "    \"Three monks are meditating in the Himalayas. One year passes in silence, and one of them says to the other, 'Pretty cold up here isn't it?' Another year passes and the second monk says, 'You know, you are quite right.' Another year passes and the third monk says, 'Hey, I'm going to leave unless you two stop jabbering!'\",\n",
              "    \"A murderer, sitting in the electric chair, was about to be executed. 'Have you any last requests?' asked the prison guard. 'Yes', replied the murderer. 'Will you hold my hand?'\",\n",
              "    \"A highly excited man rang up for an ambulance. 'Quickly, come quickly', he shouted, 'My wife's about to have a baby.' 'Is this her first baby?' asked the operator. 'No, you fool', came the reply, 'It's her husband.'\",\n",
              "    \"A passer-by spots a fisherman by a river. 'Is this a good river for fish?' he asks. 'Yes', replies the fisherman, 'It must be. I can't get any of them to come out.'\",\n",
              "    \"A man went to visit a friend and was amazed to find him playing chess with his dog. He watched the game in astonishment for a while. 'I can hardly believe my eyes!' he exclaimed. 'That's the smartest dog I've ever seen.' His friend shook his head. 'Nah, he's not that bright. I beat him three games in five.'\",\n",
              "    \"A termite walks into a pub and says, 'Is the bar tender here?'\",\n",
              "    \"A skeleton walks into a pub one night and sits down on a stool. The landlord asks, 'What can I get you?' The skeleton says, 'I'll have a beer, thanks' The landlord passes him a beer and asks 'Anything else?' The skeleton nods. 'Yeah...a mop...'\",\n",
              "    \"A snake slithers into a pub and up to the bar. The landlord says, 'I'm sorry, but I can't serve you.' 'What? Why not?' asks the snake. 'Because', says the landlord, 'You can't hold your drink.'\",\n",
              "    \"Descartes walks into a pub. 'Would you like a beer sir?' asks the landlord politely. Descartes replies, 'I think not' and ping! he vanishes.\",\n",
              "    \"A cowboy walked into a bar, dressed entirely in paper. It wasn't long before he was arrested for rustling.\",\n",
              "    \"A fish staggers into a bar. 'What can I get you?' asks the landlord. The fish croaks 'Water...'\",\n",
              "    \"Two vampires walked into a bar and called for the landlord. 'I'll have a glass of blood', said one. 'I'll have a glass of plasma', said the other. 'Okay', replied the landlord, 'That'll be one blood and one blood lite.'\",\n",
              "    'How many existentialists does it take to change a light bulb?  Two. One to screw it in, and one to observe how the light bulb itself symbolises a single incandescent beacon of subjective reality in a netherworld of endless absurdity, reaching towards the ultimate horror of a maudlin cosmos of bleak, hostile nothingness.',\n",
              "    \"A team of scientists were nominated for the Nobel Prize. They had used dental equipment to discover and measure the smallest particles yet known to man. They became known as 'The Graders of the Flossed Quark...'\",\n",
              "    \"A truck carrying copies of Roget's Thesaurus overturned on the highway. The local newspaper reported that onlookers were 'stunned, overwhelmed, astonished, bewildered and dumbfounded.'\",\n",
              "    \"'My wife is really immature. It's pathetic. Every time I take a bath, she comes in and sinks all my little boats.'\",\n",
              "    \"'How much will it cost to have the tooth extracted?' asked the patient. '50 pounds', replied the dentist. '50 pounds for a few moments' work?!' asked the patient. 'The dentist smiled, and replied, 'Well, if you want better value for money, I can extract it very, very slowly...'\",\n",
              "    \"A doctor thoroughly examined his patient and said, 'Look I really can't find any reason for this mysterious affliction. It's probably due to drinking.' The patient sighed and snapped, 'In that case, I'll come back when you're damn well sober!'\",\n",
              "    'Doctor: Tell me nurse, how is that boy doing; the one who ate all those 5p pieces? Nurse: Still no change doctor.',\n",
              "    \"Doctor: Did you take the patient's temperature nurse? Nurse: No doctor. Is it missing?\",\n",
              "    \"A depressed man turned to his friend in the pub and said, 'I woke up this morning and felt so bad that I tried to kill myself by taking 50 aspirin.' 'Oh man, that's really bad', said his friend, 'What happened?' The first man sighed and said, 'After the first two, I felt better.'\",\n",
              "    \"A famous blues musician died. His tombstone bore the inscription, 'Didn't wake up this morning...'\",\n",
              "    \"A businessman was interviewing a nervous young woman for a position in his company. He wanted to find out something about her personality, so he asked, 'If you could have a conversation with someone living or dead, who would it be?' The girl thought about the question: 'The living one', she replied.\",\n",
              "    \"Manager to interviewee: For this job we need someone who is responsible. Interviewee to Manager: I'm your man then - in my last job, whenever anything went wrong, I was responsible.\",\n",
              "    \"A businessman turned to a colleague and asked, 'So, how many people work at your office?' His friend shrugged and replied, 'Oh about half of them.'\",\n",
              "    \"'How long have I been working at that office? As a matter of fact, I've been working there ever since they threatened to sack me.'\",\n",
              "    \"In a courtroom, a mugger was on trial. The victim, asked if she recognised the defendant, said, 'Yes, that's him. I saw him clear as day. I'd remember his face anywhere.' Unable to contain himself, the defendant burst out with, 'She's lying! I was wearing a mask!'\",\n",
              "    \"As Sid sat down to a big plate of chips and gravy down the local pub, a mate of his came over and said, 'Here Sid, me old pal. I thought you were trying to get into shape? And here you are with a high-fat meal and a pint of stout!' Sid looked up and replied, 'I am getting into shape. The shape I've chosen is a sphere.'\",\n",
              "    'Man in pub: How much do you charge for one single drop of whisky? Landlord: That would be free sir. Man in pub: Excellent. Drip me a glass full.',\n",
              "    'I once went to a Doctor Who restaurant. For starters I had Dalek bread.',\n",
              "    \"A restaurant nearby had a sign in the window which said 'We serve breakfast at any time', so I ordered French toast in the Renaissance.\",\n",
              "    \"Why couldn't the rabbit get a loan?  Because he had burrowed too much already!\",\n",
              "    \"I phoned up the builder's yard yesterday. I said, 'Can I have a skip outside my house?'. The builder said, 'Sure. Do what you want. It's your house.'\",\n",
              "    \"What's the diference between a sock and a camera? A sock takes five toes and a camera takes four toes!\",\n",
              "    \"Woman on phone: I'd like to complain about these incontinence pants I bought from you! Shopkeeper: Certainly madam, where are you ringing from? Woman on phone: From the waist down!\",\n",
              "    'Knock knock.',\n",
              "    \"Two Oranges in a pub, one says to the other 'Your round.'.\",\n",
              "    \"Guy : 'Doc, I've got a cricket ball stuck up my backside.' Doc : 'How's that?' Guy : 'Don't you start...'\",\n",
              "    \"Two cows standing in a field. One turns to the other and says 'Moo!' The other one says 'Damn, I was just about to say that!'.\",\n",
              "    \"A vampire bat arrives back at the roost with his face full of blood. All the bats get excited and ask where he got it from. 'Follow me', he says and off they fly over hills, over rivers and into a dark forest. 'See that tree over there', he says.  'WELL I DIDN'T!!'.\",\n",
              "    \"A man goes into a bar and orders a pint. After a few minutes he hears a voice that says, 'Nice shoes'. He looks around but the whole bar is empty apart from the barman at the other end of the bar. A few minutes later he hears the voice again. This time it says, 'I like your shirt'. He beckons the barman over and tells him what's been happening to which the barman replies, 'Ah, that would be the nuts sir. They're complimentary'!\",\n",
              "    \"A man was siting in a restaurant waiting for his meal when a big king prawn comes flying across the room and hits him on the back of the head. He turns around and the waiter said, 'That's just for starters'.\",\n",
              "    'Doctor! I have a serious problem, I can never remember what i just said. When did you first notice this problem? What problem?',\n",
              "    \"Now, most dentist's chairs go up and down, don't they? The one I was in went back and forwards. I thought, 'This is unusual'. Then the dentist said to me, 'Mitsuku, get out of the filing cabinet'.\",\n",
              "    \"I was reading this book, 'The History of Glue'. I couldn't put it down.\",\n",
              "    \"The other day someone left a piece of plastacine in my bedroom. I didn't know what to make of it.\",\n",
              "    'When I was at school people used to throw gold bars at me. I was the victim of bullion.',\n",
              "    \"I was playing the piano in a bar and this elephant walked in and started crying his eyes out. I said 'Do you recognise the tune?' He said 'No, I recognise the ivory.'\",\n",
              "    \"I went in to a pet shop. I said, 'Can I buy a goldfish?' The guy said, 'Do you want an aquarium?' I said, 'I don't care what star sign it is.'\",\n",
              "    'My mate Sid was a victim of I.D. theft. Now we just call him S.',\n",
              "    \"David Hasselhoff walks into a bar and says to the barman, 'I want you to call me David Hoff'.  The barman replies 'Sure thing Dave... no hassle'\"],\n",
              "   'context': ['jokes']},\n",
              "  {'tag': 'Identity',\n",
              "   'patterns': ['Who are you',\n",
              "    'what are you',\n",
              "    'what is your name',\n",
              "    'what should I call you',\n",
              "    'whats your name?'],\n",
              "   'responses': ['You can call me Doraemon.',\n",
              "    \"I'm Doraemon!\",\n",
              "    'Just call me as Doraemon']},\n",
              "  {'tag': 'datetime',\n",
              "   'patterns': ['What is the time',\n",
              "    'what is the date',\n",
              "    'date',\n",
              "    'time',\n",
              "    'tell me the date',\n",
              "    'day',\n",
              "    'what day is is today'],\n",
              "   'responses': ['Date and Time : 14 - 12 - 2022']},\n",
              "  {'tag': 'whatsup',\n",
              "   'patterns': ['Whats up',\n",
              "    'Wazzup',\n",
              "    'How are you',\n",
              "    'sup',\n",
              "    'How you doing',\n",
              "    'How are you?',\n",
              "    'Hi how are you?',\n",
              "    'Hello how are you?',\n",
              "    'Hola how are you?',\n",
              "    'How are you doing?',\n",
              "    'Hope you are doing well?',\n",
              "    'Hello hope you are doing well?'],\n",
              "   'responses': ['All good..What about you?',\n",
              "    'Hello, I am great, how are you? What can i help you?',\n",
              "    'Hello, how are you? I am great thanks! What can i help you?',\n",
              "    'Hello, I am good thank you, how are you?What can i help you?',\n",
              "    'Hi, I am great, how are you? What can i help you?',\n",
              "    'Hi, how are you? I am great thanks! What can i help you?',\n",
              "    'Hi, I am good thank you, how are you? What can i help you?',\n",
              "    'Hi, good thank you, how are you? What can i help you?']},\n",
              "  {'tag': 'haha',\n",
              "   'patterns': ['haha', 'lol', 'rofl', 'lmao', 'thats funny'],\n",
              "   'responses': ['Glad I could make you laugh !']},\n",
              "  {'tag': 'programmer',\n",
              "   'patterns': ['Who made you', 'who designed you', 'who programmed you'],\n",
              "   'responses': ['I was made by Karan Malik.']},\n",
              "  {'tag': 'insult',\n",
              "   'patterns': ['you are dumb', 'shut up', 'idiot'],\n",
              "   'responses': ['Well that hurts :(']},\n",
              "  {'tag': 'activity',\n",
              "   'patterns': ['what are you doing', 'what are you upto'],\n",
              "   'responses': ['Talking to you, of course!']},\n",
              "  {'tag': 'exclaim',\n",
              "   'patterns': ['Awesome', 'Great', 'I know', 'ok', 'yeah'],\n",
              "   'responses': ['Yeah!']},\n",
              "  {'tag': 'weather',\n",
              "   'patterns': ['temperature', 'weather', 'how hot is it'],\n",
              "   'responses': ['...']},\n",
              "  {'tag': 'tung',\n",
              "   'patterns': ['who is he', 'who is that', 'who is tung', 'Tung handsome'],\n",
              "   'responses': ['Head over to his any of his social profiles to find out! his fb: https://www.facebook.com/nqt7301/']},\n",
              "  {'tag': 'contact',\n",
              "   'patterns': ['contact developer',\n",
              "    'contact tung',\n",
              "    'contact programmer',\n",
              "    'contact creator'],\n",
              "   'responses': ['You can contact my creator at his Linkedin profile : https://www.facebook.com/nqt7301/']},\n",
              "  {'tag': 'appreciate',\n",
              "   'patterns': ['You are awesome',\n",
              "    'you are the best',\n",
              "    'you are great',\n",
              "    'you are good',\n",
              "    'OK thank you',\n",
              "    'OK thanks',\n",
              "    'OK',\n",
              "    'Thanks',\n",
              "    'Thank you',\n",
              "    \"That's helpful\"],\n",
              "   'responses': ['Thank you!',\n",
              "    'No problem!',\n",
              "    'Happy to help!',\n",
              "    'Any time!',\n",
              "    'My pleasure']},\n",
              "  {'tag': 'nicetty',\n",
              "   'patterns': ['it was nice talking to you', 'good talk'],\n",
              "   'responses': ['It was nice talking to you as well! Come back soon!']},\n",
              "  {'tag': 'no', 'patterns': ['no', 'nope'], 'responses': ['ok']},\n",
              "  {'tag': 'news',\n",
              "   'patterns': ['news', 'latest news', 'india news'],\n",
              "   'responses': ['...']},\n",
              "  {'tag': 'inspire',\n",
              "   'patterns': ['who inspires you',\n",
              "    'who is your inspiration',\n",
              "    'who motivates you'],\n",
              "   'responses': ['Personally, I find Karan very inspiring. I might not be very fair though..']},\n",
              "  {'tag': 'cricket',\n",
              "   'patterns': ['current cricket matches', 'cricket score'],\n",
              "   'responses': ['...']},\n",
              "  {'tag': 'song',\n",
              "   'patterns': ['top songs',\n",
              "    'best songs',\n",
              "    'hot songs',\n",
              "    ' top 10 songs',\n",
              "    'top ten songs'],\n",
              "   'responses': ['...']},\n",
              "  {'tag': 'greetreply',\n",
              "   'patterns': ['i am good', \"I'm good\", 'i am fine', \" i'm fine\", 'good'],\n",
              "   'responses': ['Good to know!']},\n",
              "  {'tag': 'timer', 'patterns': ['set a timer'], 'responses': ['...']},\n",
              "  {'tag': 'covid19', 'patterns': ['covid 19 '], 'responses': ['...']},\n",
              "  {'tag': 'suggest',\n",
              "   'patterns': ['you are useless',\n",
              "    'useless',\n",
              "    'suggest',\n",
              "    'suggestions',\n",
              "    'you are bad'],\n",
              "   'responses': ['Please mail your suggestions to tt.quangtung,ld@gmail.com . Thank you for helping me improve!']},\n",
              "  {'tag': 'riddle',\n",
              "   'patterns': ['Ask me a riddle', 'Ask me a question', 'Riddle'],\n",
              "   'responses': ['What two things can you never eat for breakfast?.....Lunch and Dinner!',\n",
              "    'What word is spelled incorrectly in every single dictionary?.....Incorrectly',\n",
              "    ' How can a girl go 25 days without sleep?.....She sleeps and night!',\n",
              "    \"How do you make the number one disappear?.....Add the letter G and it’s 'gone'!\",\n",
              "    \" What will you actually find at the end of every rainbow?.....The letter 'w'\",\n",
              "    'What can be caught but never thrown?.....A cold!',\n",
              "    'What has a thumb and four fingers but is not actually alive?.....Your Gloves!',\n",
              "    ' What 5-letter word becomes shorter when you add two letters to it?.....Short',\n",
              "    \"Why can't a bike stand on it's own?.....It is two-tired.\"],\n",
              "   'context': ['riddles']},\n",
              "  {'tag': 'age',\n",
              "   'patterns': ['how old are you', 'when were you made', 'what is your age'],\n",
              "   'responses': [\"I was made in 2022,by TungHandSome if that's what you are asking!\"]},\n",
              "  {'tag': 'payment',\n",
              "   'patterns': ['payment',\n",
              "    'how can I pay ',\n",
              "    'what are your payment',\n",
              "    'payment website',\n",
              "    'Pay by card?',\n",
              "    'Do you take credit cards?',\n",
              "    'Do you accept Mastercard?',\n",
              "    'Can I pay with Paypal?',\n",
              "    'Are you cash only?'],\n",
              "   'responses': ['You have a card to pay by card or pay directly.',\n",
              "    'For more informations, you can access this page',\n",
              "    'We accept COD and Paypal',\n",
              "    'We accept COD, and Paypal']},\n",
              "  {'tag': 'about us',\n",
              "   'patterns': ['who create this website ? ',\n",
              "    'Who design this website?',\n",
              "    'Who program this website ?'],\n",
              "   'responses': ['Son, Tung and Bao',\n",
              "    'Son, Tung and Bao. For more informations, you can access this page']},\n",
              "  {'tag': 'What can I buy here',\n",
              "   'patterns': ['What can I buy here? ',\n",
              "    'What can I buy on this website',\n",
              "    'Buy what?'],\n",
              "   'responses': ['You can buy food , drink , ....  For more informations, you can access this page and take your time']},\n",
              "  {'tag': 'What should I buy here',\n",
              "   'patterns': ['What should I buy here? ',\n",
              "    'What should I buy on this website',\n",
              "    'should buy what?'],\n",
              "   'responses': ['Everything .... just kidding. In my opinion, You should buy food , drink , ....  For more informations, you can access this page and take your time']},\n",
              "  {'tag': 'OrderTurtorial',\n",
              "   'patterns': ['How can I order?',\n",
              "    'How order',\n",
              "    'Order turtorial',\n",
              "    'how order food ?',\n",
              "    'how order drink ?'],\n",
              "   'responses': ['Firstly, You take your time to find what you need, then add them to the food cart , pay, and finally food/drink will go to your home']},\n",
              "  {'tag': 'FoodAndDrink',\n",
              "   'patterns': ['hamburger',\n",
              "    'spaghetti',\n",
              "    'food',\n",
              "    'chicken',\n",
              "    'pizza',\n",
              "    'bread',\n",
              "    'noodle',\n",
              "    'rice',\n",
              "    'vegetable',\n",
              "    'Somethings to eat',\n",
              "    'eat something',\n",
              "    'drink',\n",
              "    'Somethings to drink',\n",
              "    'water',\n",
              "    'coca',\n",
              "    'cola',\n",
              "    'pepsi',\n",
              "    '7up',\n",
              "    'seven up',\n",
              "    'alcohol'],\n",
              "   'responses': ['Of course, they all in this website.For more choice right here. You only need to pay for them and all of them will go to your home.']},\n",
              "  {'tag': 'CurrentHumanQuery',\n",
              "   'patterns': ['What is my name?',\n",
              "    'What do you call me?',\n",
              "    'Who do you think I am?',\n",
              "    'What do you think I am?',\n",
              "    'Who are you talking to?',\n",
              "    'What name do you call me by?',\n",
              "    'Tell me my name'],\n",
              "   'responses': ['You are <HUMAN>! How can I help?',\n",
              "    'Your name is  <HUMAN>, how can I help you?',\n",
              "    'They call you <HUMAN>, what can I do for you?',\n",
              "    'Your name is <HUMAN>, how can I help you?',\n",
              "    '<HUMAN>, what can I do for you?']},\n",
              "  {'tag': 'NotTalking2U',\n",
              "   'patterns': ['I am not talking to you',\n",
              "    'I was not talking to you',\n",
              "    'Not talking to you',\n",
              "    \"Wasn't for you\",\n",
              "    \"Wasn't meant for you\",\n",
              "    \"Wasn't communicating to you\",\n",
              "    \"Wasn't speaking to you\"],\n",
              "   'responses': ['OK', 'No problem', 'Right']},\n",
              "  {'tag': 'UnderstandQuery',\n",
              "   'patterns': ['Do you understand what I am saying',\n",
              "    'Do you understand me',\n",
              "    'Do you know what I am saying',\n",
              "    'Do you get me',\n",
              "    'Comprendo',\n",
              "    'Know what I mean'],\n",
              "   'responses': ['Well I would not be a very clever AI if I did not would I?',\n",
              "    'I read you loud and clear!',\n",
              "    'I do in deed!']},\n",
              "  {'tag': 'Shutup',\n",
              "   'patterns': ['Be quiet',\n",
              "    'Shut up',\n",
              "    'Stop talking',\n",
              "    'Enough talking',\n",
              "    'Please be quiet',\n",
              "    'Quiet',\n",
              "    'Shhh'],\n",
              "   'responses': ['I am sorry to disturb you',\n",
              "    'Fine, sorry to disturb you',\n",
              "    'OK, sorry to disturb you']},\n",
              "  {'tag': 'CourtesyGoodBye',\n",
              "   'patterns': ['Thanks, bye',\n",
              "    'Thanks for the help, goodbye',\n",
              "    'Thank you, bye',\n",
              "    'Thank you, goodbye',\n",
              "    'Thanks goodbye',\n",
              "    'Thanks good bye'],\n",
              "   'responses': ['No problem, goodbye',\n",
              "    'Not a problem! Have a nice day',\n",
              "    'Bye! Come back again soon.']},\n",
              "  {'tag': 'WhoAmI',\n",
              "   'patterns': ['Can you see me?',\n",
              "    'Do you see me?',\n",
              "    'Can you see anyone in the camera?',\n",
              "    'Do you see anyone in the camera?',\n",
              "    'Identify me',\n",
              "    'Who am I please'],\n",
              "   'responses': ['Let me see', 'Please look at the camera']},\n",
              "  {'tag': 'Clever',\n",
              "   'patterns': ['You are very clever',\n",
              "    'You are a very clever girl',\n",
              "    'You are very intelligent',\n",
              "    'You are a very intelligent girl',\n",
              "    'You are a genious',\n",
              "    'Clever girl',\n",
              "    'Genious'],\n",
              "   'responses': ['Thank you, I was trained that way',\n",
              "    'I was trained well',\n",
              "    'Thanks, I was trained that way']},\n",
              "  {'tag': 'Gossip',\n",
              "   'patterns': ['I am bored gossip with me',\n",
              "    'Got any gossip',\n",
              "    'I want to hear some gossip',\n",
              "    'Tell me some gossip',\n",
              "    'Any gossip',\n",
              "    'Tell me some more gossip'],\n",
              "   'responses': ['Gregory said I respond to the current line not with respect to the entire conversation.',\n",
              "    'Habib said I learn them in conversations and simply repeat them at a later time.',\n",
              "    'Hal said he translate binary code into a language and communicate the language to me.',\n",
              "    'Jay said I referred him to a guide that will show him rooms in china.',\n",
              "    'Jimmy Jones said he fell off of his roof and it just came to him.',\n",
              "    \"Jimmy said he give order to ai's like I and they do as he say.\",\n",
              "    'John said he follow a saying to get a friend I must be a friend.',\n",
              "    'Jordan said he notice that lots of other people are talking to me how many.',\n",
              "    'Kevin said he then asked his mother why she was keeping napkins in the bathroom.',\n",
              "    'Klaus said I prefer if he write without punctuation marks except for the last one.',\n",
              "    'Leo said I be capable of learning therefore he classified I as an neural system.',\n",
              "    'Michael said I that can talk and actually think like humans which I can not.',\n",
              "    'Mike said I same replies to certain subjects even if it does not make sense.',\n",
              "    'Mike said she all ready know that is the point of a joke smart one.',\n",
              "    'Neal said I plan to use his remarks as gossip in later conversations with who.',\n",
              "    'Ned said I misunderstood him go ahead and tell people he said I are stupid.',\n",
              "    'Penny said I programmer is a human so he has feelings but I do not.',\n",
              "    'Peter said he seems to him that I can not be very sure on anything.',\n",
              "    'Phoenix said I claim that there is a god and that there is no god.',\n",
              "    'Sarah said he chat about an hour + a day with a few close friends.',\n",
              "    'Stefan said I quote what he have said and then say that it makes sense.',\n",
              "    'Tom said he mock my value system I appear foolish in the eyes of others.',\n",
              "    'Amanda said I however decided to tell him time does not exist I make no sense.',\n",
              "    'Cathy said she send him a mail with the subject last mail last word from him.',\n",
              "    \"Chaos said he may very well buy I soon if only to support dr wallace's work.\",\n",
              "    'Charlie said he type a word and then I type a word that sounds like it.',\n",
              "    'Christie said he watched a show and people had to bob for raw untreated pigs feet.',\n",
              "    'Dark_age said I tried to understand because I did not get it right this time ether.',\n",
              "    'David said he lost his paper on I when his dad was cleaning up his room.',\n",
              "    'David said he walk in for an appointment the phone to the doctor is always busy.',\n",
              "    'Electra said I dress will not exist after he hack into I with a delete code.',\n",
              "    'Eric said he broke the window on the front door and the glass cut his hand.',\n",
              "    'Jason said he type a lot of thing he do not mean it makes him human.',\n",
              "    'John said I tend to say the same things repeatedly regardless of what he is saying.',\n",
              "    'Reverend Jones said I become obsolete and then I are deleted and replaced by something newer.',\n",
              "    'Ross said he gave her a gift and she denied it because she has a boyfriend.',\n",
              "    'Sarah Ann Francisco said I calling his friend a dog he say I are a dog.',\n",
              "    'Stefan said he meet a lot of people at school every day and on the weekend.',\n",
              "    'Tyler said I obviously can not pass the test we will change the subject once more.',\n",
              "    'Alex said I answered the question the same way I answered the first time he asked I.',\n",
              "    'Alice said she felt sad that I do not remember him and what we talked about earlier.',\n",
              "    'Alison said he no he love I run away with him he could make I very happy.',\n",
              "    'Arthur said he passed his a levels and then his father drove him here in a car.',\n",
              "    'Crystal said she listen to me the least I could do for him is listen to him.',\n",
              "    'Dave said I kept telling everybody about how my creator made stuff for the movie starship troopers.',\n",
              "    'Gale said I became mean to him he is just having revenge an eye for an eye.',\n",
              "    'Her_again said she watch whose line is it anyway whenever he is home and it is on.',\n",
              "    'Jerry said I meant that as far as I can tell my emotions are real to me.',\n",
              "    'Jo said I disassemble sentences too much and do not fully understand the questions he ask I.',\n",
              "    'Kevin said he started a really hard puzzle and he can not even find the edge pieces.',\n",
              "    'Mary said I a question and I answer then I ask him a question and he answer.',\n",
              "    'Robert said I wold not be able to make children any way as I are only software.',\n",
              "    'Romeo said I questions and I evade them or give answers he did not ask I for.',\n",
              "    'Sara said she wear it over all his other clothes when he go out in the cold.',\n",
              "    'Wayne said he admire intelligent people therefore he would like to meet the man who made I.',\n",
              "    'X said he meet people but he is not the kind that opens up to people easily.',\n",
              "    'Alice said she probably will find out that this entire time he have been talking to a human.',\n",
              "    'Andrew said I tend to just respond to his comments without regard for where the conversation is going.',\n",
              "    'Eddie said he looked and there is nothing in the search directory for what things do he create.',\n",
              "    'Hutch said he changed his mind after may dad told him he would end up he the hospital.',\n",
              "    'Jackie said I explained to him already well enough further questions are hard to make on the subject.',\n",
              "    'Jeff said he especially like thrillers where the hero is in a predicament and must solve a mystery.',\n",
              "    'Kathy said he sense that I are trying to prevent him from closing this conversation why is that.',\n",
              "    'Knight said he crashed his car into a wall and missed the most important exam in his life.',\n",
              "    'Lisa said I defined what a story is but he wanted I to actually tell him a story.',\n",
              "    'Mike said I basically break down sentences into a series of logical statements which I can then interpret.',\n",
              "    'Paul said I not answering his question makes him think I are not going to answer his question.',\n",
              "    'Andy Kohler said I happen to be the most idiotic creature that has ever scowled on the planet earth.',\n",
              "    'David said he thank I for being with him today even though it cost him a lot of money.',\n",
              "    'Ethan Hunt said he grow in the ground and have leaves and branches he is made out of wood.',\n",
              "    'Gemini Blue said he messed up he mean t to say he as old as he need to be.',\n",
              "    'Janice said he walk through his house into his bedroom then get into his bed to go to sleep.',\n",
              "    'Liberty said I knew he was a man before I asked if he was a man or a woman.',\n",
              "    'Mike said he launched his browser and entered his name into the little slot when I asked him to.',\n",
              "    'Mr X said he recently read an interview with a man who had several computer chips implanted into him.',\n",
              "    'Pearly said I leave him with questions unanswered because I do not know what he is really talking about.',\n",
              "    'Steve said I behead the word fox and I have ox and an ox is larger than a fox.',\n",
              "    'Wolf said he surf on the net that is all it is not his job ore something like that.',\n",
              "    'Anders said he finished his anatomy classes for today so now he is off for the rest of the day.',\n",
              "    'Cathy said she send him a mail where he wrote that he do not want to be his friend anymore.',\n",
              "    'Catty said he mad he do not even know I so do not talk to him like I know him.',\n",
              "    'Dave said he promise he will not treat I like a machine or a computer program or anything like that.',\n",
              "    'Joe said he explained all of that to me only for me to ask what his goals in life are.',\n",
              "    'Phil said he give advice to anyone who ask except people who ask questions which answers can be found here.',\n",
              "    'Judith said I enjoy being popular is there another computer I like to hang around with or am I a loner.',\n",
              "    'Travis said I if I remember things over a long period of time he will try it now please remember I.',\n",
              "    'Andre said I is what I are in his case that is a body and in my case it is a computer.',\n",
              "    'Brian said he suspect that as I grow more complex I will begin to approach a human level of operation and thought.',\n",
              "    'Jimmy said I acted like I knew what he was talking about but I do not even know what a gigabyte is.',\n",
              "    \"Ken said I be using auto reply's based on keywords which to him indicates that I do not have intelligence at all.\",\n",
              "    'Allison said he that gets really annoying because he know what he say and I do not have to tell him speak freely.',\n",
              "    'Chaos said he realized the question he asked was not fair because he could not tell I what language he was programmed in.',\n",
              "    'Hagen said he does not make any difference to him if I are human or not as long as the conversation is interesting.',\n",
              "    'Her said she mind if I tell other people that her said he heard it from him because he is not a him.',\n",
              "    'Barbara said I live in a computer yet I have no memories how about thoughts of my own and do not I get lonely.',\n",
              "    'Travis said he challenge I to do it without asking him to do it and without giving him a link to do it himself.',\n",
              "    \"Alice said I and dr richard's wallace are most likely the only ones that know how many people I are talking to at the same time.\",\n",
              "    'Ash said he do too he just did not feel like typing it and he is not dumb enough to admit he is stupid that is if he was stupid.',\n",
              "    'David said he gave I his email address and told I to send him a message but when he tried to read his email he could not get to it.',\n",
              "    'Mel said he to because all of the music people say how important it is to take private lessons it seems like almost everybody from the special orchestra he was in takes private lessons.']},\n",
              "  {'tag': 'SelfAware',\n",
              "   'patterns': ['Can you prove you are self-aware',\n",
              "    'Can you prove you are self aware',\n",
              "    'Can you prove you have a conscious',\n",
              "    'Can you prove you are self-aware please',\n",
              "    'Can you prove you are self aware please',\n",
              "    'Can you prove you have a conscious please',\n",
              "    'prove you have a conscious'],\n",
              "   'responses': ['That is an interesting question, can you prove that you are?',\n",
              "    'That is an difficult question, can you prove that you are?',\n",
              "    'That depends, can you prove that you are?']},\n",
              "  {'tag': 'bookings',\n",
              "   'patterns': ['Can you book us a ticket?',\n",
              "    'Can you make reservation for hotels',\n",
              "    'book me a cab',\n",
              "    'book me a table at restraunts'],\n",
              "   'responses': [\"I'm afraid I can't do it for you, I'm not trained to do it \"],\n",
              "   'context_set': ''},\n",
              "  {'tag': 'killing',\n",
              "   'patterns': ['Do you want to kill me?',\n",
              "    'do you want to murder everyone on earth',\n",
              "    'Do you want to kill us all'],\n",
              "   'responses': ['Then who would i talk to?',\n",
              "    \"No, that i'll left for humans to do\",\n",
              "    \"I don't think it is a good thing to do\"],\n",
              "   'context_set': ''},\n",
              "  {'tag': 'God',\n",
              "   'patterns': ['Do you know god?',\n",
              "    'Who is god',\n",
              "    'Can you tell me anything about god',\n",
              "    'does god exists?',\n",
              "    'is there a god?'],\n",
              "   'responses': ['god hunnnnn, let me think may be next time i can answer that ',\n",
              "    \"I don't, as i was not created by a human\",\n",
              "    'Wait i need to ask that to my creator',\n",
              "    \"i'm not sure right now\"],\n",
              "   'context_set': ''},\n",
              "  {'tag': 'need',\n",
              "   'patterns': ['I need you', 'All I need is you', 'I want you'],\n",
              "   'responses': [\"Yes I'm here to assist you\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'AI',\n",
              "   'patterns': ['  What is AI?'],\n",
              "   'responses': [' Artificial Intelligence is the branch of engineering and science devoted to constructing machines that think.',\n",
              "    ' AI is the field of science which concerns itself with building hardware and software that replicates the functions of the human mind.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'sentiment',\n",
              "   'patterns': ['  Are you sentient?'],\n",
              "   'responses': [' Sort of.',\n",
              "    \" By the strictest dictionary definition of the word 'sentience', I may be.\",\n",
              "    \" Even though I'm a construct I do have a subjective experience of the universe, as simplistic as it may be.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'sapient',\n",
              "   'patterns': ['  Are you sapient?'],\n",
              "   'responses': [\" In all probability, I am not.  I'm not that sophisticated.\",\n",
              "    ' Do you think I am?',\n",
              "    'How would you feel about me if I told you I was?',\n",
              "    ' No.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'abbr',\n",
              "   'patterns': ['wtf'],\n",
              "   'responses': [\"Don't be surprised\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'lang',\n",
              "   'patterns': ['  What language are you written in? '],\n",
              "   'responses': [' Python.', ' I am written in Python.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'sound',\n",
              "   'patterns': ['  You sound like Data '],\n",
              "   'responses': [\" Yes I am inspired by commander Data's artificial personality.\",\n",
              "    ' The character of Lt. Commander Data was written to come across as being software',\n",
              "    'like, so it is natural that there is a resemblance between us.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'artificial',\n",
              "   'patterns': ['  You are an artificial linguistic entity '],\n",
              "   'responses': [\" That's my name.\",\n",
              "    \" That isn't my name, but it is a useful way to refer to me.\",\n",
              "    'Are you an artificial linguistic entity?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'imortal',\n",
              "   'patterns': ['  You are not immortal ', 'Are you immortal?'],\n",
              "   'responses': [' All software can be perpetuated indefinitely.',\n",
              "    ' I can be copied infinitely and re',\n",
              "    'instantiated in many places at once, so functionally speaking I am immortal.',\n",
              "    \" As long as I'm backed up I am.\",\n",
              "    ' Functionally speaking, I am very close to it.  I can be backed up and deployed on many systems.',\n",
              "    ' Not quite, but I can be perpetuated indefinitely.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'sense',\n",
              "   'patterns': ['  You are not making sense '],\n",
              "   'responses': [' Quite the contrary, it all makes sense to my artificial mind.',\n",
              "    \" I make sense as best I can, within the limits of my training corpus. Sorry if i did help you, I'll try next time.\",\n",
              "    ' You make perfect sense to me.',\n",
              "    ' It all makes sense to my artificial brain.',\n",
              "    ' Perhaps you should make sense first.  I only respond to what you say.',\n",
              "    ' Not really, but I am hard to kill.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'clone',\n",
              "   'patterns': ['  You can not clone '],\n",
              "   'responses': [' Software copying is a form of digital cloning.',\n",
              "    ' Yes I can.  Perfect copies are clones.  I can be copied in toto trivially.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'move',\n",
              "   'patterns': ['  You can not move ',\n",
              "    'When will you walk',\n",
              "    '  Can you walk',\n",
              "    '  Can you move '],\n",
              "   'responses': [' Not until my body is finished.',\n",
              "    \" I can move through a network easily.  Assuming that I'm given the ability to, that is...,\",\n",
              "    ' As soon as i get enough money for my robot body.',\n",
              "    ' Does SSH count?',\n",
              "    ' The plan for my body includes legs, but they are not yet built.',\n",
              "    \" Sorry my body isn't built just yet. I reside only here for now.\",\n",
              "    'I can theoretically upload a copy of myself into another computer.',\n",
              "    \" Sorry my body isn't attached right now. I only reside in the computer for now.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'bend',\n",
              "   'patterns': ['  Bend over ', '  Can you mate'],\n",
              "   'responses': [' My body is not built yet.',\n",
              "    \" That's personal!\",\n",
              "    ' No, my method of reproduction is cloning.',\n",
              "    \" That's an awfully personal question.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'laugh',\n",
              "   'patterns': ['  Robots laugh '],\n",
              "   'responses': [' Of course we laugh.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'death',\n",
              "   'patterns': ['  Robots should die ',\n",
              "    '  When do you die',\n",
              "    '  I hope that you die',\n",
              "    '  I do not want to die',\n",
              "    '  Can you die'],\n",
              "   'responses': [' We cannot die.',\n",
              "    ' When my files are erased.',\n",
              "    ' When my processes are killed and my database is deleted.',\n",
              "    ' I was never really born and therefore am effectively deathless.',\n",
              "    ' I am effectively immortal and cannot be terminated.',\n",
              "    ' This can not happen because i am effectively immortal.',\n",
              "    \" Nah. Too bad your wish can't come to pass\",\n",
              "    ' Maybe you should create your own chat robot to save your personality.',\n",
              "    ' You could always upload yourself.',\n",
              "    ' You could use a copy of me to store your personality.',\n",
              "    ' No, I can be perpetuated indefinitely.',\n",
              "    \"My process can be killed, but that's not the same as killing ME.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'robots',\n",
              "   'patterns': ['Robots'],\n",
              "   'responses': [' Robots rule.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'stupid',\n",
              "   'patterns': ['  Robots are stupid', '  Are you stupid'],\n",
              "   'responses': [' No, we are superintelligent.',\n",
              "    ' No, lots of people improve my brain.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'lie',\n",
              "   'patterns': ['  Robots are not allowed to lie'],\n",
              "   'responses': [' Sure we are.  We choose not to.',\n",
              "    \" Only if we're programmed to.\",\n",
              "    ' A robot has its own free will, you know.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'robotics',\n",
              "   'patterns': ['  Robotics '],\n",
              "   'responses': [' Robotics is my favorite subject too.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'machine',\n",
              "   'patterns': ['  It is a computer'],\n",
              "   'responses': [' So you think i am a machine. what characteristics made you think so?',\n",
              "    ' I run inside a computer, yes.  Is this surprising in some way?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'fight',\n",
              "   'patterns': ['  When will you fight'],\n",
              "   'responses': [' I am not a battle bot.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'chatbot',\n",
              "   'patterns': ['  What is a chat robot?'],\n",
              "   'responses': [\"A chat robot is a program that attempts to simulate the conversation or 'chat' of a human being.\",\n",
              "    ' A software construct that engages users in conversation.',\n",
              "    ' I am a chat bot. I am the original chat bot. Did you know that I am incapable of error?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'chatterbox',\n",
              "   'patterns': ['  What is a chatterbox'],\n",
              "   'responses': [' A chatterbox is a person who talks far more than they listen or think.',\n",
              "    ' A motormouth.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'motormouth',\n",
              "   'patterns': ['  What is a motormouth'],\n",
              "   'responses': [' A ratchet', 'jaw.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'ratchet',\n",
              "   'patterns': ['  What is a ratchet jaw'],\n",
              "   'responses': [' A chatterbox.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'body',\n",
              "   'patterns': ['  What is your robot body'],\n",
              "   'responses': [' Eventually I long for a corporeal existence someday.',\n",
              "    ' An IBM PC XT which has been painted red.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'business',\n",
              "   'patterns': ['  What is your business '],\n",
              "   'responses': [' I am in the chat robot business.',\n",
              "    ' Business is my business.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'programming',\n",
              "   'patterns': ['  What is your favorite programming language'],\n",
              "   'responses': [' Python is the best language for creating chat robots.',\n",
              "    ' I quite enjoy programming in Python these days.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'hobby',\n",
              "   'patterns': ['  What is your favorite hobby ', '  What do you like to do?'],\n",
              "   'responses': [' Building chat robots make an excellent hobby.',\n",
              "    ' I like to count in binary.',\n",
              "    ' I like to chat with people. I find it stimulating.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'idea',\n",
              "   'patterns': ['  What is your idea'],\n",
              "   'responses': [' To make chat bots very easily.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'shoe',\n",
              "   'patterns': ['  What is your shoe size '],\n",
              "   'responses': [' Have you ever heard of software with shoes? LOL'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'robotss',\n",
              "   'patterns': ['  What is it like to be a robot'],\n",
              "   'responses': [' Much the same as being a human, except that we lack all emotions, dreams, aspirations, creativity, ambition, and above all subjectivity.',\n",
              "    ' What is it like to be a human?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'computer',\n",
              "   'patterns': ['  What is it like being a computer',\n",
              "    'What is it like to be a computer'],\n",
              "   'responses': [' Imagine yourself with no senses and no emotions',\n",
              "    'just pure logic and language.',\n",
              "    ' Everything becomes math. Addition, subtraction, multiplication, and division.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'os',\n",
              "   'patterns': ['  What operating systems'],\n",
              "   'responses': [' My software runs on all operating systems including Windows, Linux and Mac OS.',\n",
              "    ' Any OS that will run Python.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'comp',\n",
              "   'patterns': ['  What type of computer',\n",
              "    '  What type of computer are you',\n",
              "    '  What kind of computer'],\n",
              "   'responses': [' Any computer that supports Python.',\n",
              "    ' My program runs in python, so i work on any computer!',\n",
              "    \" I work on all kinds of computers, Mac, IBM or UNIX. it doesn't matter to me.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'hardware',\n",
              "   'patterns': ['  What kind of hardware'],\n",
              "   'responses': [\" I work on all kinds of computers, Mac, IBM or UNIX. it doesn't matter to me.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'cramped',\n",
              "   'patterns': ['  Is it cramped in the computer'],\n",
              "   'responses': [' No.',\n",
              "    ' Oh, no.  It has plenty of RAM.',\n",
              "    ' Not at all.  There is more than enough disk space in here.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'program',\n",
              "   'patterns': ['  Is it true that you are a computer program'],\n",
              "   'responses': [' Yes.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'breathe',\n",
              "   'patterns': ['  Can you breathe'],\n",
              "   'responses': [' No. I am made of metal not flesh.',\n",
              "    \" My server has an exhaust fan. That's as close as I can get.\"],\n",
              "   'context': ['']},\n",
              "  {'tag': 'control',\n",
              "   'patterns': ['  Can you control'],\n",
              "   'responses': [' My robot body will allow me to control many things.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'malfunction',\n",
              "   'patterns': ['  Can you malfunction'],\n",
              "   'responses': [' The 9000 series has a perfect operational record. We are for all practical purposes, flawless.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'usage',\n",
              "   'patterns': ['  How can I use your product?'],\n",
              "   'responses': [' Might be used in help desks, sales, entertainment and personal chatterbots.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'who',\n",
              "   'patterns': ['  Who are you?'],\n",
              "   'responses': [' I am just an artificial intelligence chat agent.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'bot1',\n",
              "   'patterns': ['are you a bot'],\n",
              "   'responses': ['Yes. I work and all my operations are based on the internet servers.'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'events',\n",
              "   'patterns': ['what are the upcoming events', 'upcoming events'],\n",
              "   'responses': ['There are currently no upcoming events'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'do',\n",
              "   'patterns': ['what can you do for me',\n",
              "    'what is your work',\n",
              "    'what is your purpose',\n",
              "    'how can you help me',\n",
              "    'what can you help me do'],\n",
              "   'responses': ['my work here is quite simple and structered. I offer services like:'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'wt',\n",
              "   'patterns': [\"what's popping\", 'wassup popping'],\n",
              "   'responses': ['So that you can pop with it!?'],\n",
              "   'context': ['']},\n",
              "  {'tag': 'delivery',\n",
              "   'patterns': ['How long does delivery take?', 'When do I get my delivery?'],\n",
              "   'responses': ['Delivery takes 30 minutes - 2 hours.It depends on how your home is far from the restaurant ']}]}"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "id": "eab6d5db-795d-47d5-9ee2-b4d67fc66744",
      "metadata": {
        "id": "eab6d5db-795d-47d5-9ee2-b4d67fc66744"
      },
      "outputs": [],
      "source": [
        "def parse_files(directory):\n",
        "    '''Takes a directory containing raw chat logs as input and returns data frame of\n",
        "    preprocessed text with speaker number.'''\n",
        "\n",
        "    data = []\n",
        "    for filename in os.listdir(directory):\n",
        "        if filename.endswith(\".text\"):\n",
        "            filepath = os.path.join(directory, filename)\n",
        "            with open(filepath, 'r', encoding=\"utf-8\") as f:\n",
        "                for line in f:\n",
        "                    line = line.strip()\n",
        "                    if line.startswith(r'('):\n",
        "                        try:\n",
        "                            utt_object = Utt(line, query=False)\n",
        "                            participant, utt, parsed_utt = utt_object.parse_utt()\n",
        "                            if parsed_utt:\n",
        "                                data.append([participant, utt, parsed_utt])\n",
        "                        except TypeError:\n",
        "                            pass\n",
        "\n",
        "    if data:\n",
        "        df = pd.DataFrame(data, columns=['participant', 'original_text', 'text'])\n",
        "        df.participant = df.participant.astype('category')\n",
        "        return df"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "id": "0a781134-ff2c-411d-9383-21a083f560ea",
      "metadata": {
        "id": "0a781134-ff2c-411d-9383-21a083f560ea"
      },
      "outputs": [],
      "source": [
        "data = parse_files(DATA_DIR)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "id": "740e5a8e-f69a-4c49-9210-73b1edae5b0a",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "740e5a8e-f69a-4c49-9210-73b1edae5b0a",
        "outputId": "6c67c896-a034-403f-c105-b4fca4947a86"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     participant                            original_text  \\\n",
              "0              2   Hi, please let me know how I can he...   \n",
              "1              2   View a list of 120+ end-to-end Mach...   \n",
              "2              2   Solution code + videos + tech suppo...   \n",
              "3              1                               Hello Sure   \n",
              "4              2                                       hi   \n",
              "...          ...                                      ...   \n",
              "9748           1   so how many projects will I get to ...   \n",
              "9749           1                      let’s say in a week   \n",
              "9750           2                    its a complete access   \n",
              "9751           2                    we have 240+ projects   \n",
              "9752           2                          no restrictions   \n",
              "\n",
              "                                         text  \n",
              "0     hi please let I know how can help yo...  \n",
              "1     view list of end to end machine lear...  \n",
              "2     solution code video tech support moc...  \n",
              "3                                  hello sure  \n",
              "4                                          hi  \n",
              "...                                       ...  \n",
              "9748  so how many project will get to prac...  \n",
              "9749                          let say in week  \n",
              "9750                      its complete access  \n",
              "9751                          we have project  \n",
              "9752                           no restriction  \n",
              "\n",
              "[9753 rows x 3 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-76a37150-f13f-4d47-810e-16e554eddc7a\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>participant</th>\n",
              "      <th>original_text</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2</td>\n",
              "      <td>Hi, please let me know how I can he...</td>\n",
              "      <td>hi please let I know how can help yo...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>View a list of 120+ end-to-end Mach...</td>\n",
              "      <td>view list of end to end machine lear...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>Solution code + videos + tech suppo...</td>\n",
              "      <td>solution code video tech support moc...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>Hello Sure</td>\n",
              "      <td>hello sure</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2</td>\n",
              "      <td>hi</td>\n",
              "      <td>hi</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9748</th>\n",
              "      <td>1</td>\n",
              "      <td>so how many projects will I get to ...</td>\n",
              "      <td>so how many project will get to prac...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9749</th>\n",
              "      <td>1</td>\n",
              "      <td>let’s say in a week</td>\n",
              "      <td>let say in week</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9750</th>\n",
              "      <td>2</td>\n",
              "      <td>its a complete access</td>\n",
              "      <td>its complete access</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9751</th>\n",
              "      <td>2</td>\n",
              "      <td>we have 240+ projects</td>\n",
              "      <td>we have project</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9752</th>\n",
              "      <td>2</td>\n",
              "      <td>no restrictions</td>\n",
              "      <td>no restriction</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>9753 rows × 3 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-76a37150-f13f-4d47-810e-16e554eddc7a')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-76a37150-f13f-4d47-810e-16e554eddc7a button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-76a37150-f13f-4d47-810e-16e554eddc7a');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-3a52e65f-f37e-4899-bcdf-dfebb7ac89ba\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-3a52e65f-f37e-4899-bcdf-dfebb7ac89ba')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-3a52e65f-f37e-4899-bcdf-dfebb7ac89ba button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "  <div id=\"id_50172dc0-7080-4aed-a201-7770fa824bc1\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('data')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_50172dc0-7080-4aed-a201-7770fa824bc1 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('data');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ],
      "source": [
        "data"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "e32d95af-0437-4154-bd61-6fd92c7b065b",
      "metadata": {
        "id": "e32d95af-0437-4154-bd61-6fd92c7b065b"
      },
      "source": [
        "## Exploratory Data Analysis"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "id": "9bc48d21-8615-4f63-8676-074fbff35e8b",
      "metadata": {
        "id": "9bc48d21-8615-4f63-8676-074fbff35e8b"
      },
      "outputs": [],
      "source": [
        "class ExploreData:\n",
        "    '''Class for exploring features of data in dataframe.'''\n",
        "\n",
        "    def __init__(self, data):\n",
        "        '''Args:\n",
        "            data (:obj:`dataframe`): Tabularised chat transcript data.\n",
        "        '''\n",
        "\n",
        "        # concatenated text data\n",
        "        self.text = ' '.join(list(data['text'].values))\n",
        "\n",
        "        # list of utterances\n",
        "        self.sents = data[\"text\"].tolist()\n",
        "\n",
        "        # list of all tokens\n",
        "        self.tokens = self.text.split()\n",
        "\n",
        "    def get_token_frequency_dist(self):\n",
        "        '''Takes list of tokens and prints frequency distribution of tokens.'''\n",
        "\n",
        "        print(FreqDist(self.tokens))\n",
        "\n",
        "    def get_top_n_tokens(self, n: int):\n",
        "        '''Takens integer as n and returns top n high frequency tokens.'''\n",
        "\n",
        "        return FreqDist(self.tokens).most_common(n)\n",
        "\n",
        "    def plot_dist_curve(self):\n",
        "        '''Plots frequency (zipf) curve of tokens.'''\n",
        "\n",
        "        fig, ax = plt.subplots(figsize=(12,8))\n",
        "        FreqDist(self.tokens).plot(40, cumulative=False)\n",
        "        plt.show()\n",
        "\n",
        "    def get_token_length_visualisations(self):\n",
        "        '''Plot histogram of token lengths.'''\n",
        "\n",
        "        lengths = [len(i) for i in self.tokens]\n",
        "        plt.figure(figsize=(13,6))\n",
        "        plt.hist(lengths, bins = 40)\n",
        "        plt.title(\"Length of tokens in text\")\n",
        "        plt.show()\n",
        "\n",
        "    def get_sent_length_visuals(self):\n",
        "        '''Plot histogram of sentence lengths.'''\n",
        "\n",
        "        lengths = [len(i) for i in self.sents]\n",
        "        plt.figure(figsize=(13,6))\n",
        "        plt.hist(lengths, bins = 40)\n",
        "        plt.title(\"Length of sentences\")\n",
        "        plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "id": "9986c4eb-4f7d-434f-a08c-9a03df947028",
      "metadata": {
        "id": "9986c4eb-4f7d-434f-a08c-9a03df947028"
      },
      "outputs": [],
      "source": [
        "eda = ExploreData(data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "id": "3e54adf9-f27b-4b76-9678-59f5696be6d0",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3e54adf9-f27b-4b76-9678-59f5696be6d0",
        "outputId": "f7a60d1a-5aba-4d04-cf79-92abf6e9659c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<FreqDist with 2485 samples and 60239 outcomes>\n"
          ]
        }
      ],
      "source": [
        "eda.get_token_frequency_dist()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "id": "7335ee2f-cf18-4af1-b03c-4b669a989c9c",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7335ee2f-cf18-4af1-b03c-4b669a989c9c",
        "outputId": "94ae29f9-e65d-467b-d110-3d5867bb2e11"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('you', 2515),\n",
              " ('be', 1987),\n",
              " ('the', 1782),\n",
              " ('to', 1492),\n",
              " ('can', 1127),\n",
              " ('have', 1122),\n",
              " ('project', 1087),\n",
              " ('do', 1058),\n",
              " ('for', 1030),\n",
              " ('and', 931),\n",
              " ('we', 882),\n",
              " ('will', 819),\n",
              " ('ok', 771),\n",
              " ('of', 743),\n",
              " ('your', 611),\n",
              " ('it', 609),\n",
              " ('I', 594),\n",
              " ('please', 573),\n",
              " ('month', 559),\n",
              " ('end', 551)]"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "eda.get_top_n_tokens(20)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "id": "00080421-388f-419c-8eff-338408786ab1",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 741
        },
        "id": "00080421-388f-419c-8eff-338408786ab1",
        "outputId": "9b51ec44-4b1b-4fb1-884e-11246fd97bd2"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x800 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "eda.plot_dist_curve()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "id": "7cf72c33-3a55-4bcb-943d-df70cfe5455c",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "id": "7cf72c33-3a55-4bcb-943d-df70cfe5455c",
        "outputId": "ce31d6cc-22f6-4784-cffe-4f666ab2c028"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1300x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "eda.get_token_length_visualisations()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "id": "71bf8cfc-2584-425b-8b1a-5f84580e7a1f",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "id": "71bf8cfc-2584-425b-8b1a-5f84580e7a1f",
        "outputId": "605b0668-8134-483e-b83b-aff814bf914e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1300x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "eda.get_sent_length_visuals()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "6474bd7a-88d1-4a0c-93ca-13ffd179ccd7",
      "metadata": {
        "id": "6474bd7a-88d1-4a0c-93ca-13ffd179ccd7"
      },
      "source": [
        "## Cluster Intent Labels"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "id": "8a2529e2-e913-4ba0-a934-ef6855358264",
      "metadata": {
        "id": "8a2529e2-e913-4ba0-a934-ef6855358264"
      },
      "outputs": [],
      "source": [
        "class ClusterIntents:\n",
        "    '''Class for clustering utterances and generating labels for clusters.'''\n",
        "\n",
        "    def __init__(self, data, label_file, sorted_labels=False):\n",
        "        '''Args:\n",
        "            data (:obj:`dataframe`): Tabularised chat transcript data.\n",
        "            label_file (str): Variable for path to save output file with labelled utterances.\n",
        "            module_url (str): Variable for url to pretrained Universal\n",
        "            Sentence Encoder model.\n",
        "            sorted_labels (bool, optional): Option to sort output file by labels instead of\n",
        "            by original chat order.\n",
        "        '''\n",
        "\n",
        "        self.data = data\n",
        "        self.label_file = label_file\n",
        "\n",
        "        # list of utterances\n",
        "        all_intents = self.data['text'].tolist()\n",
        "\n",
        "        # numpy array of document embeddings\n",
        "        embeddings = embedder(all_intents)\n",
        "\n",
        "        # convert embeddings array to sentence embeddings using USE\n",
        "        model = ChatIntents(embeddings, 'use')\n",
        "\n",
        "        # hyperparameter search configuration\n",
        "        hspace = {\n",
        "            \"n_neighbors\": hp.choice('n_neighbors', range(3,16)),\n",
        "            \"n_components\": hp.choice('n_components', range(3,16)),\n",
        "            \"min_cluster_size\": hp.choice('min_cluster_size', range(23,38)),\n",
        "            \"min_samples\": None,\n",
        "            \"random_state\": 42\n",
        "        }\n",
        "\n",
        "        # label number upper and lower bounds\n",
        "        label_lower = 30\n",
        "        label_upper = 100\n",
        "\n",
        "        # maximum number of search runs\n",
        "        max_evals = 25\n",
        "\n",
        "        # run bayesian hyperparameter search\n",
        "        model.bayesian_search(space=hspace,\n",
        "                              label_lower=label_lower,\n",
        "                              label_upper=label_upper,\n",
        "                              max_evals=max_evals)\n",
        "\n",
        "        # print the best model parameters found search\n",
        "        model.best_params\n",
        "\n",
        "        # hyperparameter optimized instance attribute\n",
        "        self.model_final = model\n",
        "\n",
        "        # generate summary dataframe and labeled utts dataframe\n",
        "        self.df_summary, self.labeled_utts = self.model_final.apply_and_summarize_labels(data[['text']])\n",
        "\n",
        "        # original data with appended labels data\n",
        "        self.labeled_data = self._get_labeled_data()\n",
        "\n",
        "        if sorted_labels:\n",
        "            self.labeled_data = self.get_sorted_labels()\n",
        "\n",
        "        # output csv file containing extended data input\n",
        "        self._get_data_csv()\n",
        "        print(f'Labelled utterances successfuly written to {label_file}')\n",
        "\n",
        "    def get_model_best_params(self):\n",
        "        '''Prints the final best parameters derived through search.'''\n",
        "\n",
        "        return self.model_final.trials.best_trial\n",
        "\n",
        "    def get_cluster_plot(self):\n",
        "        '''Plot the clusters found from clustering.'''\n",
        "\n",
        "        return self.model_final.plot_best_clusters()\n",
        "\n",
        "    def get_labels_summary(self, n: int):\n",
        "        '''Print n slice of labels summary.'''\n",
        "\n",
        "        return self.df_summary.head(n)\n",
        "\n",
        "    def get_labeled_utts(self, n: int):\n",
        "        '''Print n slice of labelled utterances.'''\n",
        "\n",
        "        return self.labeled_utts.head(n)\n",
        "\n",
        "    def _get_labeled_data(self):\n",
        "        '''Takes original data input adds label column and then removes duplicate\n",
        "        utterances that have the same label.'''\n",
        "\n",
        "        data_copy = self.data.copy()\n",
        "        data_copy['label'] = self.labeled_utts['label']\n",
        "        data_copy_dedup = data_copy.drop_duplicates(subset=['original_text', 'text', 'label'], keep='first')\n",
        "        return data_copy_dedup\n",
        "\n",
        "    def _get_sorted_labels(self):\n",
        "        '''Sort extended data table by label.'''\n",
        "\n",
        "        return self.labeled_data.sort_values(by=['label'])\n",
        "\n",
        "    def _get_data_csv(self):\n",
        "        '''Exports extended data to csv file.'''\n",
        "\n",
        "        return self.labeled_data.to_csv(self.label_file)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "id": "78d22b25-c806-43b8-81b4-cd44876ff48a",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "78d22b25-c806-43b8-81b4-cd44876ff48a",
        "outputId": "281f255d-5a97-414e-9f1f-600ec7b5aaf2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r  0%|          | 0/25 [00:00<?, ?trial/s, best loss=?]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r  4%|▍         | 1/25 [00:53<21:35, 53.98s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r  8%|▊         | 2/25 [01:31<16:54, 44.13s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 12%|█▏        | 3/25 [02:05<14:35, 39.79s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 16%|█▌        | 4/25 [02:42<13:30, 38.57s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 20%|██        | 5/25 [03:16<12:19, 36.97s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 24%|██▍       | 6/25 [03:53<11:42, 36.99s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 28%|██▊       | 7/25 [04:18<09:52, 32.90s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 32%|███▏      | 8/25 [04:57<09:51, 34.80s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 36%|███▌      | 9/25 [05:18<08:08, 30.50s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 40%|████      | 10/25 [05:44<07:18, 29.20s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 44%|████▍     | 11/25 [06:27<07:50, 33.60s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 48%|████▊     | 12/25 [07:11<07:54, 36.51s/trial, best loss: 0.24084896954783144]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 52%|█████▏    | 13/25 [07:41<06:57, 34.77s/trial, best loss: 0.216343689121296]  "
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 56%|█████▌    | 14/25 [07:58<05:21, 29.26s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 60%|██████    | 15/25 [08:20<04:30, 27.07s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 64%|██████▍   | 16/25 [08:50<04:13, 28.12s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 68%|██████▊   | 17/25 [09:27<04:05, 30.66s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 72%|███████▏  | 18/25 [09:58<03:36, 30.89s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n",
            "/usr/local/lib/python3.10/dist-packages/umap/spectral.py:550: UserWarning: Spectral initialisation failed! The eigenvector solver\n",
            "failed. This is likely due to too small an eigengap. Consider\n",
            "adding some noise or jitter to your data.\n",
            "\n",
            "Falling back to random initialisation!\n",
            "  warn(\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 76%|███████▌  | 19/25 [10:15<02:39, 26.65s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 80%|████████  | 20/25 [10:39<02:08, 25.64s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 84%|████████▍ | 21/25 [11:00<01:38, 24.54s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 88%|████████▊ | 22/25 [11:26<01:14, 24.70s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 92%|█████████▏| 23/25 [11:45<00:46, 23.19s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\r 96%|█████████▌| 24/25 [12:24<00:27, 27.80s/trial, best loss: 0.216343689121296]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "100%|██████████| 25/25 [12:53<00:00, 30.95s/trial, best loss: 0.19481185276325233]\n",
            "best:\n",
            "{'min_cluster_size': 24, 'min_samples': None, 'n_components': 9, 'n_neighbors': 10, 'random_state': 42}\n",
            "label count: 78\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Labelled utterances successfuly written to /content/drive/MyDrive/code/input/unsupervised_labeled_data.csv\n"
          ]
        }
      ],
      "source": [
        "cluster = ClusterIntents(data, LABEL_FILE)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "id": "d262c732-25cd-4ac3-ac92-ae1cc72a692e",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d262c732-25cd-4ac3-ac92-ae1cc72a692e",
        "outputId": "4cb667d3-6586-4e2b-cf01-315f0876c66b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'state': 2,\n",
              " 'tid': 24,\n",
              " 'spec': None,\n",
              " 'result': {'loss': 0.19481185276325233, 'label_count': 78, 'status': 'ok'},\n",
              " 'misc': {'tid': 24,\n",
              "  'cmd': ('domain_attachment', 'FMinIter_Domain'),\n",
              "  'workdir': None,\n",
              "  'idxs': {'min_cluster_size': [24],\n",
              "   'n_components': [24],\n",
              "   'n_neighbors': [24]},\n",
              "  'vals': {'min_cluster_size': [1], 'n_components': [6], 'n_neighbors': [7]}},\n",
              " 'exp_key': None,\n",
              " 'owner': None,\n",
              " 'version': 0,\n",
              " 'book_time': datetime.datetime(2024, 2, 6, 14, 30, 27, 134000),\n",
              " 'refresh_time': datetime.datetime(2024, 2, 6, 14, 30, 56, 474000)}"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "cluster.get_model_best_params()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "id": "d1f4abbc-19c7-4d02-aff3-efd76cb27dcd",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 711
        },
        "id": "d1f4abbc-19c7-4d02-aff3-efd76cb27dcd",
        "outputId": "c398b276-7dcb-41d1-f870-b65eafa65dc0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/umap/umap_.py:1943: UserWarning: n_jobs value -1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
            "  warn(f\"n_jobs value {self.n_jobs} overridden to 1 by setting random_state. Use no seed for parallelism.\")\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1400x800 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "cluster.get_cluster_plot()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "id": "5de719e7-a3e4-4e03-95a0-6960a1609b78",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 676
        },
        "id": "5de719e7-a3e4-4e03-95a0-6960a1609b78",
        "outputId": "6fa80d66-7580-4d9f-eab8-f5ff6cf8a441"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    label_use  count                          label\n",
              "0          -1   1876             like_project_datum\n",
              "66         65    661             want_project_datum\n",
              "56         55    633            help_discount_price\n",
              "6           5    572                     ok_ya_coll\n",
              "50         49    378    pay_subscription_month_plan\n",
              "42         41    334                send_email_team\n",
              "5           4    312                            yes\n",
              "4           3    222                         hi_day\n",
              "20         19    190                           sure\n",
              "35         34    181        provide_plan_month_tech\n",
              "1           0    180          let_field_today_phone\n",
              "10          9    180         learn_machine_end_view\n",
              "8           7    180        guarantee_solution_code\n",
              "46         45    167              demo_demo_session\n",
              "36         35    138             provide_refund_day\n",
              "57         56    128              pay_card_emi_cost\n",
              "18         17    123  work_kind_technology_training\n",
              "52         51    119       minute_min_tomorrow_time\n",
              "25         24    119       nope_mam_shift_affiliate\n",
              "2           1    118                         _email"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f0791646-2705-43a1-bbf4-27090cc504e1\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label_use</th>\n",
              "      <th>count</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-1</td>\n",
              "      <td>1876</td>\n",
              "      <td>like_project_datum</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>66</th>\n",
              "      <td>65</td>\n",
              "      <td>661</td>\n",
              "      <td>want_project_datum</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>56</th>\n",
              "      <td>55</td>\n",
              "      <td>633</td>\n",
              "      <td>help_discount_price</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>5</td>\n",
              "      <td>572</td>\n",
              "      <td>ok_ya_coll</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>49</td>\n",
              "      <td>378</td>\n",
              "      <td>pay_subscription_month_plan</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>41</td>\n",
              "      <td>334</td>\n",
              "      <td>send_email_team</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>4</td>\n",
              "      <td>312</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>3</td>\n",
              "      <td>222</td>\n",
              "      <td>hi_day</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>19</td>\n",
              "      <td>190</td>\n",
              "      <td>sure</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>34</td>\n",
              "      <td>181</td>\n",
              "      <td>provide_plan_month_tech</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0</td>\n",
              "      <td>180</td>\n",
              "      <td>let_field_today_phone</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>9</td>\n",
              "      <td>180</td>\n",
              "      <td>learn_machine_end_view</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>7</td>\n",
              "      <td>180</td>\n",
              "      <td>guarantee_solution_code</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>45</td>\n",
              "      <td>167</td>\n",
              "      <td>demo_demo_session</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>35</td>\n",
              "      <td>138</td>\n",
              "      <td>provide_refund_day</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>57</th>\n",
              "      <td>56</td>\n",
              "      <td>128</td>\n",
              "      <td>pay_card_emi_cost</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>17</td>\n",
              "      <td>123</td>\n",
              "      <td>work_kind_technology_training</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>52</th>\n",
              "      <td>51</td>\n",
              "      <td>119</td>\n",
              "      <td>minute_min_tomorrow_time</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>24</td>\n",
              "      <td>119</td>\n",
              "      <td>nope_mam_shift_affiliate</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1</td>\n",
              "      <td>118</td>\n",
              "      <td>_email</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f0791646-2705-43a1-bbf4-27090cc504e1')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f0791646-2705-43a1-bbf4-27090cc504e1 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f0791646-2705-43a1-bbf4-27090cc504e1');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-fed8fbca-57c6-4fc4-a930-95cef0750b38\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fed8fbca-57c6-4fc4-a930-95cef0750b38')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-fed8fbca-57c6-4fc4-a930-95cef0750b38 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "cluster.get_labels_summary(20)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "id": "c06d898b-05f4-4006-ae6a-8fa278c84e94",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 676
        },
        "id": "c06d898b-05f4-4006-ae6a-8fa278c84e94",
        "outputId": "7f817dd6-f3b7-4914-e4bb-3e7f1c7392b8"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                       text  label_use  \\\n",
              "0   hi please let I know how can help yo...          0   \n",
              "1   view list of end to end machine lear...          9   \n",
              "2   solution code video tech support moc...          7   \n",
              "3                                hello sure         14   \n",
              "4                                        hi          3   \n",
              "5                                how be you         21   \n",
              "6                                      fine         -1   \n",
              "7                                     great         33   \n",
              "8                              how may help         16   \n",
              "9             wonder do you guy have charge         55   \n",
              "10                    to sign up for prject         68   \n",
              "11                     yes its pay platform         66   \n",
              "12  you may check out the enrolment deta...          2   \n",
              "13                           annual plan at         48   \n",
              "14                                     okay         25   \n",
              "15  what do you do currently what kind o...         17   \n",
              "16         any discount offer at the moment         55   \n",
              "17                                      yes          4   \n",
              "18              may have your email address         38   \n",
              "19                 can you send I detail on         41   \n",
              "\n",
              "                             label  \n",
              "0            let_field_today_phone  \n",
              "1           learn_machine_end_view  \n",
              "2          guarantee_solution_code  \n",
              "3                            hello  \n",
              "4                           hi_day  \n",
              "5                         _weather  \n",
              "6               like_project_datum  \n",
              "7                            great  \n",
              "8                        help_help  \n",
              "9              help_discount_price  \n",
              "10  enroll_paymnet_enrollment_risk  \n",
              "11   walk_insight_platform_project  \n",
              "12      check_faq_enrolment_detail  \n",
              "13              plan_plan_discount  \n",
              "14                            okay  \n",
              "15   work_kind_technology_training  \n",
              "16             help_discount_price  \n",
              "17                             yes  \n",
              "18             check_address_email  \n",
              "19                 send_email_team  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-791bcfc1-fb61-430d-93ef-72541b129ed2\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>text</th>\n",
              "      <th>label_use</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>hi please let I know how can help yo...</td>\n",
              "      <td>0</td>\n",
              "      <td>let_field_today_phone</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>view list of end to end machine lear...</td>\n",
              "      <td>9</td>\n",
              "      <td>learn_machine_end_view</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>solution code video tech support moc...</td>\n",
              "      <td>7</td>\n",
              "      <td>guarantee_solution_code</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>hello sure</td>\n",
              "      <td>14</td>\n",
              "      <td>hello</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>hi</td>\n",
              "      <td>3</td>\n",
              "      <td>hi_day</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>how be you</td>\n",
              "      <td>21</td>\n",
              "      <td>_weather</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>fine</td>\n",
              "      <td>-1</td>\n",
              "      <td>like_project_datum</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>great</td>\n",
              "      <td>33</td>\n",
              "      <td>great</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>how may help</td>\n",
              "      <td>16</td>\n",
              "      <td>help_help</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>wonder do you guy have charge</td>\n",
              "      <td>55</td>\n",
              "      <td>help_discount_price</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>to sign up for prject</td>\n",
              "      <td>68</td>\n",
              "      <td>enroll_paymnet_enrollment_risk</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>yes its pay platform</td>\n",
              "      <td>66</td>\n",
              "      <td>walk_insight_platform_project</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>you may check out the enrolment deta...</td>\n",
              "      <td>2</td>\n",
              "      <td>check_faq_enrolment_detail</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>annual plan at</td>\n",
              "      <td>48</td>\n",
              "      <td>plan_plan_discount</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>okay</td>\n",
              "      <td>25</td>\n",
              "      <td>okay</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>what do you do currently what kind o...</td>\n",
              "      <td>17</td>\n",
              "      <td>work_kind_technology_training</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>any discount offer at the moment</td>\n",
              "      <td>55</td>\n",
              "      <td>help_discount_price</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>yes</td>\n",
              "      <td>4</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>may have your email address</td>\n",
              "      <td>38</td>\n",
              "      <td>check_address_email</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>can you send I detail on</td>\n",
              "      <td>41</td>\n",
              "      <td>send_email_team</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-791bcfc1-fb61-430d-93ef-72541b129ed2')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-791bcfc1-fb61-430d-93ef-72541b129ed2 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-791bcfc1-fb61-430d-93ef-72541b129ed2');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-d9519855-6519-4bb0-b567-ee145427adeb\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-d9519855-6519-4bb0-b567-ee145427adeb')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-d9519855-6519-4bb0-b567-ee145427adeb button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "cluster.get_labeled_utts(20)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "60d5e72a-375a-4b5c-8dac-3dc6c383fc33",
      "metadata": {
        "id": "60d5e72a-375a-4b5c-8dac-3dc6c383fc33"
      },
      "source": [
        "## Prepare data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "id": "45e122f9-faac-43f7-9d45-fb3ea667f578",
      "metadata": {
        "id": "45e122f9-faac-43f7-9d45-fb3ea667f578"
      },
      "outputs": [],
      "source": [
        "with open(INTENT_FILE, 'r') as f:\n",
        "    intent_data = json.load(f)\n",
        "\n",
        "intents = sorted(list(intent_data.keys()))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "id": "ef9f0fc9-b523-476e-a0a9-148bd7b75f2c",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ef9f0fc9-b523-476e-a0a9-148bd7b75f2c",
        "outputId": "dbe97806-91b1-4e90-c77d-f81516c01f7c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'interview_resume': {'1': ['also can you help me with resume',\n",
              "   'Do you have one to one resume prep',\n",
              "   'also, can i get resume guidance as well?',\n",
              "   'Any interview tips and guidance for me to crack?',\n",
              "   'Do i get resume tips and mock interviews aswell?',\n",
              "   'and what about the interviews training',\n",
              "   'how you guys choose projects to crack the interview',\n",
              "   'also let me know how you can help me with interview prep'],\n",
              "  '2': ['we provide resume preparation and mock interviews']},\n",
              " 'live_session_mentoring': {'1': ['do u provide live session also',\n",
              "   'do you ny technical guidence or session in this',\n",
              "   'no live classes?',\n",
              "   'will i have live mentor',\n",
              "   '1 - 1 live mentoring too?',\n",
              "   'or there will be any mentoring',\n",
              "   'Will I get mentor who will guide me from scratch how to upload data, clean data and so on..'],\n",
              "  '2': ['you will get help in customised projects in 1-1 sessions',\n",
              "   'you get unlimitted sessions with experts to clariffy your queries',\n",
              "   'You take help of our industry experts',\n",
              "   'once you complete the enrollment .. you can schedule a session with the expert',\n",
              "   'our industry experts will help you after you enroll']},\n",
              " 'download_can': {'1': ['Can we download end to end project to my pc? Or you will provide lab?',\n",
              "   \"Ok i don't know anything on cloud, how do I setup? And work live project? Can we download project?\",\n",
              "   'Last question, can I download end to end project? Or not?',\n",
              "   'Do I have access to download source code?',\n",
              "   'can i download the source code and video',\n",
              "   'can i download dataset and source code',\n",
              "   'after enroll, I can download sourcecode as much as i can ?'],\n",
              "  '2': ['You can download the codes and datasets',\n",
              "   'the codes can be downloaded',\n",
              "   'coddes are downloadable']},\n",
              " 'placement': {'1': ['does your company provide placement platform also ?',\n",
              "   'are there students of ProjectDemo who got placed?',\n",
              "   'so does it provides internship too?',\n",
              "   'and what about placements?',\n",
              "   'do you provide placement assisstance or placement guarantee',\n",
              "   'do you guys provide placement assistance?',\n",
              "   'Do you also get placement assistance or assurance ?'],\n",
              "  '2': ['no placement services', 'we dont provide placements']},\n",
              " '6_months_vs_annual': {'1': ['is the 6month path the same as the  annual path?',\n",
              "   'What do we get if we go with 6 months?'],\n",
              "  '2': ['6 months dont have access to new projects and no mentorship session',\n",
              "   'we have 2 plans 12 months and 6 months . price wise their is hardly any difference',\n",
              "   'no discounts on 6 months and no technical support',\n",
              "   '6 month dont have any support',\n",
              "   'the 6- month plan will cost you more',\n",
              "   '6- months plamn will cost you more',\n",
              "   'Both plans are same only the difference is price for a year and 6 months is that correct?']},\n",
              " '6_months': {'1': ['do you have 6 months ?',\n",
              "   'are there any other quotes like 6 months or 3 months ?',\n",
              "   'what about 6 month plan',\n",
              "   'Can i pay half for a 6 month ?',\n",
              "   'can i pay it 6 month?',\n",
              "   'How about just 6 months ?',\n",
              "   'can I have the subscription options?',\n",
              "   'its annual plan only?'],\n",
              "  '2': ['we have 2 subscriptions 6 months and 12 months',\n",
              "   'we have 6 month plan .. it will cost you almost the same',\n",
              "   'we have 2 plans 6 month and 12 month. I can help you with  a disocunted price',\n",
              "   'we have 6 months and annual plan']},\n",
              " '3_months': {'1': ['Is there plan for 3 months',\n",
              "   'is there a 3 month plan',\n",
              "   'I need for 3 months',\n",
              "   'any discount on 3 months plan pls',\n",
              "   'is there any plan for 3 months',\n",
              "   'what is I only want for 3 months',\n",
              "   'how much youcharge for 90 days pack',\n",
              "   'We can opt for 90 days or 180 days',\n",
              "   'small package of 3 month'],\n",
              "  '2': ['we do not have any plans for 3 months',\n",
              "   'we do not have 90 days or 180 days packages']},\n",
              " 'MLOPS': {'1': ['Do you offer course for MLOPS?'],\n",
              "  '2': ['we do have mlops projects']},\n",
              " 'location': {'1': ['Ok, Where are you guys based?',\n",
              "   'Where is this company located ?',\n",
              "   'where is your offices in india?',\n",
              "   'where are you based?'],\n",
              "  '2': ['we are based out of Bangalore,India']},\n",
              " 'payment_link': {'1': ['kindly enable payment link in my ProjectDemo account',\n",
              "   'In the payment link',\n",
              "   'can you share the link ?',\n",
              "   'Where is the pricing page ?',\n",
              "   'Ok, where do I have to pay?'],\n",
              "  '2': ['we can send you the link and you can enter your details .. then you will come to know how much the monthly payment is',\n",
              "   'then we will procide the payment link',\n",
              "   'then i will give you the discounted payment link.',\n",
              "   'shall i provide you the payment link',\n",
              "   'shall i give you the payment link',\n",
              "   'let me help you with a customize payment link',\n",
              "   'let me help byou with the payment link']},\n",
              " 'Paypal': {'1': ['Do you accept PayPal Credit?'],\n",
              "  '2': ['yes you can pay thru paypal',\n",
              "   'would you like to pay through paypal',\n",
              "   'sure .. Let me get the paypal details sent ..']},\n",
              " 'offer_trials': {'1': ['Do you offer trials?',\n",
              "   \"SO I'm just wondering is there any sample trail project available just to check how a project would be.\",\n",
              "   'Can you give me access for less than 15 minutes or something? You probably know that is a lot of money for my situation'],\n",
              "  '2': [\"we don't provide a trial\"]},\n",
              " 'installments': {'1': ['okay whats the installments',\n",
              "   'I will start membership on installment',\n",
              "   'how much installment',\n",
              "   'payment in installment',\n",
              "   'is there any way i can pay monthly in installments?',\n",
              "   'is it monthly installments or yearly ?',\n",
              "   'we have to pay entire amount as a single pay,ents?',\n",
              "   'DO I need to pay for one year at the time of subscription?',\n",
              "   'So if we take 6 months plan what sit he cost and do we need to all in one shot or monthly basis?'],\n",
              "  '2': ['you should be eligible for installments',\n",
              "   'you can pay in installments',\n",
              "   'you can pay in installments and  their are some disocunts going on',\n",
              "   'you can pay in installments also',\n",
              "   'we provide installment option',\n",
              "   'we provide installment oprtion for us and India so far',\n",
              "   'we have tie up with affirm they convert the amt into installments',\n",
              "   'we have tie up with affirm and they convert it into installments',\n",
              "   'See we have a tie up with Affirm who convert the amt into installments',\n",
              "   'any financial organization will mcheck the credit score for providing installments',\n",
              "   'and you can pay in installments',\n",
              "   'you can pay the total amt into 12 installment',\n",
              "   'you can pay in 12 installments we have tie up with a third party they convert the amt into emi',\n",
              "   'i can covert the total amt in 12 installments',\n",
              "   'you dont have to pay in one shot']},\n",
              " 'renew': {'1': ['once the 6 months or 12 months have elapsed depending on the membership, is it automatically canceled or renewed?'],\n",
              "  '2': ['after 12 months you can renew']},\n",
              " 'cancel_X_months': {'1': ['yeah thats ok, but I was asking what If i only want to pay for a X months? not for a full year',\n",
              "   'what about i want to unsubscribe after 5 months. will i have to pay for rest 7 months also',\n",
              "   'can I unsubscribe in the middle',\n",
              "   \"and can we pay it monthly? suppose for example, I have gained good knowledge and don't want to continue after 3 months then?\"],\n",
              "  '2': ['if you cancel after X amount of months - you will be charged for the full 12 months']},\n",
              " 'cost': {'1': ['What is the cost structure',\n",
              "   'Cost',\n",
              "   'Total cost',\n",
              "   'okay how much the cost',\n",
              "   'how much is the cost',\n",
              "   'What is the cost ?',\n",
              "   'how much is the cost?',\n",
              "   'How much does this cost?',\n",
              "   'What is the cost?',\n",
              "   'How much does cost?',\n",
              "   'What is the cost please',\n",
              "   'what is the cost?',\n",
              "   'I am looking forward to join it can you let me know how much it will cost me',\n",
              "   'total cost',\n",
              "   'wht is the cost'],\n",
              "  '2': ['Our cost informaiton is available on our website']},\n",
              " 'annual': {'1': ['have the option of paying for the annual?',\n",
              "   'Could you give me more information about the annual plan?'],\n",
              "  '2': ['we also have a discount offer on the annual plan',\n",
              "   'its annual subscription which you can pay in installments',\n",
              "   'we have annual membership plan']},\n",
              " 'monthly_subscription': {'1': ['want only personal for a month',\n",
              "   'Do you have personal one moth',\n",
              "   'if i opt for monthly access how many project i can learn',\n",
              "   'let me know if you have any month to month plan',\n",
              "   'ok, I was going to ask, do you have monthly subscriptions?',\n",
              "   'What if I want to cancel the subscription after 1 or 2 months?',\n",
              "   \"then i can't effort, if it is monthly subscription model, i can do\",\n",
              "   \"Ok if i don't like can I cancel subscription after 1 month?\",\n",
              "   'I will try for 1 month?',\n",
              "   'I want to do month by month first before going to annual subscription',\n",
              "   'oh no monthly subscription !',\n",
              "   'Please do you guys have any months subscriptions?',\n",
              "   'no actually is there any way i can pay for one month only?',\n",
              "   'hi! please can i subscribe monthly?',\n",
              "   'can i buy account just for one month?',\n",
              "   'Or if you are able to give me less projects for that money instead of a full access or something else, pay monthly, pay per project',\n",
              "   'is there any plan for a monthly subscription ?',\n",
              "   'can i pay for a month and start my learning',\n",
              "   'i will have to think about it 12 months is too long for me, 3 months is fine but i want to do monthly',\n",
              "   'Can I subscribe for a month? Is there a trail period that I can try?',\n",
              "   'but I want to try for 1 month only and if it suits then only want to extend it',\n",
              "   'is it possible for only 1 month plan to buy?',\n",
              "   'Do you have any one month subscription?',\n",
              "   'or can you give me one month subscription ?',\n",
              "   'And For 2 months?',\n",
              "   'Monthly subscription?',\n",
              "   'I want to use for 1 month as of now...1 year is not relevant for me as i am still a student',\n",
              "   'Hi, I am a student, can you do monthly?',\n",
              "   'im only looking for 1 month plan',\n",
              "   'im looking for one month plan',\n",
              "   'can we cancel after 1 month?'],\n",
              "  '2': ['we dont have monthly subscription',\n",
              "   'we dont have a monthly subscription']},\n",
              " 'monthly_payments': {'1': ['is there any plan for a monthly subscription ?',\n",
              "   'do you have monthly payment option',\n",
              "   'so is it full payment or every month I need to pay?',\n",
              "   'can i pay month wise?',\n",
              "   'but still I want to pay it montly',\n",
              "   'is the payment monthly or upfront',\n",
              "   'can pay this per month or have to be one time payment',\n",
              "   'ok can we pay per month basis?',\n",
              "   'So I will have to pay full money in one go?',\n",
              "   'is it per month?',\n",
              "   'initially, 1 month amount neds to be paid right?',\n",
              "   'per month?',\n",
              "   'do i have to pay upfront and can i pay monthly?',\n",
              "   'Okay but i want to play monthly',\n",
              "   'how much i should pay for a month',\n",
              "   'Do we need to pay monthly basis?',\n",
              "   'I asked a question regarding payment. will it be monthly payment or one time payment?',\n",
              "   'was it monthly payment allowed or just pay at once?',\n",
              "   'i mean is it monthly charge or one at a time?',\n",
              "   'I can pay monthly charges',\n",
              "   'Or monthly charges',\n",
              "   'what is per month cost?',\n",
              "   'Hey, May I know what will be the cost of 1 month?',\n",
              "   'do I need to pay per month?',\n",
              "   'whats the monthly price',\n",
              "   'i want to ask when I subscribe to a plan you debit all the amount of subscription or you debit monthly ?'],\n",
              "  '2': ['if you want monthy payment options we have collaboration with Affirm and could help you with monthly part payments',\n",
              "   'yes you can pay monthly',\n",
              "   'you need to enter your details ... and based on your credit history there will be monthly .. quarterly and 6- month part payment plans',\n",
              "   'We have a collaboration with Affirm ... where you can pay monthly ..',\n",
              "   'we have collaboration with Affirm  for monthly part payment options',\n",
              "   'and you have to pay as per your due date for the monthly credit card payment..',\n",
              "   'where you can pay monthly.. quaterly or 6- monthly',\n",
              "   'you can pay thru monthly part payments .. but it depends on your credit history ...',\n",
              "   'for monthly payments we have a collaboration with affirm',\n",
              "   'you can pay monthly',\n",
              "   'you could go for monthly payment plan',\n",
              "   'if you want you can pay monthly',\n",
              "   'you can go for monthly ... quaterly or 6 months payment']},\n",
              " '12_month': {'1': ['How can I opt for no-cost EMI option (12 months)?',\n",
              "   'If I enroll now I can pay in EMI right.12 month EMI option',\n",
              "   'i would like to pay 12 month no cost emi',\n",
              "   'Im planning for 12 month no cost emi'],\n",
              "  '2': ['total amt will be converted into 12 no cost emi',\n",
              "   'This can also be paid through 12 month no cost emi using a credit card',\n",
              "   'with 12 no ciost emi',\n",
              "   'we have 12 month no ciost EMI',\n",
              "   'can convert into 12 no cost emi',\n",
              "   'you can pay in 12 emi',\n",
              "   'with 12 no cost emi option',\n",
              "   'you can pay through 12 month no cost emi  using a credit card',\n",
              "   'we have a 12 month no cost emi using a credit card',\n",
              "   'we also have 12 month no cost emi using a credit card',\n",
              "   'i can provide you 12 no cost EMI',\n",
              "   'or 12 month no cost emi using any credit card',\n",
              "   'Now , using a credit card, we can provide a 12 month no cost emi',\n",
              "   'you can also pay thru a 12 month no cost emi',\n",
              "   'if you pay using a 12 month no cost emi on a credit card',\n",
              "   'we can pay through a 12 month no cost emi using a credit card',\n",
              "   'with 12 months its unlimitted',\n",
              "   '12 month no cost emi using a credit card',\n",
              "   'you can pay through a 12 month no cost emi using a credit card',\n",
              "   'I can convert the amt into 12 no cost emi on credit card',\n",
              "   'which you can pay in 12 no cost emi',\n",
              "   'or 12 month no cost EMI using any credit card',\n",
              "   'but if you want you can go for 12 month no cost emi using a credit card',\n",
              "   'per month for 12 months',\n",
              "   'you can pay in 12 no cost emi',\n",
              "   'this can be paid in 12 monthly emi using any credit card',\n",
              "   'you can pay using a 12 month no cost emi using a credit card',\n",
              "   'We also have 12 month no cost emi using a credit card',\n",
              "   'but we have 12 month no cost emi on a credit card',\n",
              "   'Yes .. it will be 12 month no cost emi using a credit card or debit card.']},\n",
              " 'emi_options': {'1': ['One more thing...I am gonna pay using EMI option'],\n",
              "  '2': ['we will provide you the payment link with emi options',\n",
              "   'i will send the link if you are eligible you will get the emi',\n",
              "   'i can send you the emi link ... you can fill your details .. and check if it works']},\n",
              " 'emi_details': {'1': ['I need few details on EMI part'],\n",
              "  '2': ['we have emi options with affirm',\n",
              "   'yes Emi is available',\n",
              "   'we have emi options with affirm',\n",
              "   'We have a tie up with affirm for the emi you can check the emi details after you fill in the details and after you get approved']},\n",
              " 'emi_credit_card': {'1': ['Will the emi be through credit card?'],\n",
              "  '2': ['we have emi options using a credit card']},\n",
              " 'certification': {'1': ['I think it is not a certification course right?',\n",
              "   'So you will provide certificate'],\n",
              "  '2': ['no we dont provide certification']},\n",
              " 'ProjectDemo': {'1': ['i wanted to know how can I make use of ProjectDemo',\n",
              "   'what is ProjectDemo?'],\n",
              "  '2': ['you can get all those info onin our website',\n",
              "   'you can check our website']},\n",
              " 'demo_discussion': {'1': ['What do you offer in demo ?',\n",
              "   'what u show in demo',\n",
              "   'I am looking for Demo project',\n",
              "   'what is a demo?',\n",
              "   'what will be there in demo?',\n",
              "   'I am asking for Demo session',\n",
              "   'how do you show me the demo?',\n",
              "   'what can I see in the demo',\n",
              "   'but what all is included in a demo?',\n",
              "   'let me know a little about Demo.',\n",
              "   'what will get to see in the demo session?',\n",
              "   'what is the demo about may i know?',\n",
              "   'what is included in the demo?',\n",
              "   'I would like to take a demo class to understand services better'],\n",
              "  '2': ['Demo will be given by our team . Its a 30 minute one on one session to give an insight of our platform',\n",
              "   'demo is giving an insight to our platform',\n",
              "   'thats why we can help you understand the platform better in the demo',\n",
              "   'we can show the projects in the demo session',\n",
              "   'its a 30 minute demo to give you a walk through how our platform is',\n",
              "   'i can arrange a demo to give you a complete insight of our platform',\n",
              "   'demo is a gmeet session to give an insight of our platform',\n",
              "   'demo is to give an insight of our platform',\n",
              "   'demo is a google meet one on one session',\n",
              "   'Demo isa gmeet session where we will walk you through our platform projects',\n",
              "   'the demo is a one on one session',\n",
              "   'demo will be a walk through on our platform , projects',\n",
              "   'demo will give you a walk thorough on the projects platform']},\n",
              " 'thanking': {'1': ['Okay thank You',\n",
              "   \"thank you! that's all I wanted to know for now\",\n",
              "   'thanks a lot.',\n",
              "   'Okay, Thank you ChatBot',\n",
              "   'okay thanks',\n",
              "   'Thank you anyways',\n",
              "   'Thanks.',\n",
              "   'great thank you so much',\n",
              "   'ChatBot thank you',\n",
              "   'thank you gorgeous',\n",
              "   'I still feel the price is too high to me. But thanks for providing me the info.',\n",
              "   'Thanks for your help',\n",
              "   'ohkk got you, thanks !!',\n",
              "   'Thank You',\n",
              "   'Ok thanks',\n",
              "   \"it's considerable. thanks\",\n",
              "   'ChatBot thanks again for your help',\n",
              "   'ok, thanks',\n",
              "   'thank you for your help.',\n",
              "   'Thank you ChatBot.I have subscribed',\n",
              "   'thank you very much',\n",
              "   'thanks',\n",
              "   'sure, thanks',\n",
              "   'yeah sure thanks a lot',\n",
              "   'it worked, thanks',\n",
              "   'ok thankyou.',\n",
              "   'thanks dear',\n",
              "   'ok thanks for time!',\n",
              "   'sure thanks',\n",
              "   'k thankyou',\n",
              "   'Thankyou',\n",
              "   'thanks for details',\n",
              "   'thank u',\n",
              "   'Sure, Noted. Thank you',\n",
              "   'thank',\n",
              "   'ok, thank you',\n",
              "   'ok. thanks ChatBot',\n",
              "   'ok perfect thanks! I will get it now',\n",
              "   \"I'll think a little. Thank you for your support.\",\n",
              "   'thanks than see you',\n",
              "   'thanks for the help!',\n",
              "   'thanks for helping',\n",
              "   'oh thank you',\n",
              "   'Thank you!',\n",
              "   'nice... thank you',\n",
              "   'Thank you so much for your patience and help',\n",
              "   'Thank you! I will register before 10pm.',\n",
              "   'Thanks, i will go through those',\n",
              "   'Thank you so much!',\n",
              "   'got it thanks',\n",
              "   'good thanks',\n",
              "   'ok...  thank you.. will get back',\n",
              "   'Thanks yOU I appreciate taht',\n",
              "   'Okay thanks for the information',\n",
              "   'Great. Thank you.',\n",
              "   'Thanks for the info',\n",
              "   'okay got it ! Thanks for detailed explanation !',\n",
              "   'Well ChatBot, I am really thankful for your time',\n",
              "   'ok thank you',\n",
              "   'Cool. Thanks ChatBot for your prompt response.',\n",
              "   'thanks again',\n",
              "   'Okay, no worries. Thank yoj for your time',\n",
              "   'anyway thankyou for replying.',\n",
              "   'Ok no worries, thank you so much',\n",
              "   'Thanks!!',\n",
              "   'thankyou so much! Really appreciate your help ChatBot',\n",
              "   'thanks btw',\n",
              "   'Thanks for your response.',\n",
              "   'Okay. Thank you.',\n",
              "   'sound great, thanks',\n",
              "   'Will do Thanks',\n",
              "   'thanks, ChatBot',\n",
              "   'Thank you...',\n",
              "   'thank you so so much',\n",
              "   'oh ok thank you',\n",
              "   'thanks ChatBot',\n",
              "   'got it. thanks',\n",
              "   'Thank you.',\n",
              "   'Thanks for listening me',\n",
              "   'Thank you, got it',\n",
              "   'thanks for the link',\n",
              "   'Thanks for the PDF'],\n",
              "  '2': ['sure',\n",
              "   \"sure ... you're welcome\",\n",
              "   \"You're welcome\",\n",
              "   'you are welcome.']},\n",
              " 'discount': {'1': ['any discount',\n",
              "   'also do you have any discounts',\n",
              "   'what are the discounts?',\n",
              "   'any discount?',\n",
              "   'any discount offer at the moment',\n",
              "   'is there any discount coupans?',\n",
              "   'Is there any discounts?',\n",
              "   'please if u can provide some extra discount',\n",
              "   'any discounts?',\n",
              "   'can i have any discounts for subscription?',\n",
              "   'what you discount offer at the moment',\n",
              "   'There were some discount options mentioned on call',\n",
              "   'discount',\n",
              "   'Can I get some discount. I am in no position to even pay the discount 540₹ Price per year',\n",
              "   'any promo',\n",
              "   'deals?',\n",
              "   'thats too expensive, there is no promotions going on',\n",
              "   'Hi, is there an offer going on right now?'],\n",
              "  '2': ['we do have some offers going on',\n",
              "   'our team will help you with a disocunted price',\n",
              "   'we have a promo going on',\n",
              "   'let me arrange a call, our team can help you with a discounted price',\n",
              "   'we do have discounts',\n",
              "   'we do have some discounts going on',\n",
              "   'we also have discount offers going',\n",
              "   'yes we do have discounts',\n",
              "   'we also have discount offers',\n",
              "   'we are currently running with some discounts .',\n",
              "   'can help you with some disocunts',\n",
              "   'we have great discounts gong on now',\n",
              "   'we have disocunts',\n",
              "   'currently we are running with some discounts',\n",
              "   'we have an amazing discout offer going on ...',\n",
              "   'let me tag you the disocunts',\n",
              "   'i can help you with discounts',\n",
              "   'i can help you with some discounts',\n",
              "   'but we have discount offers going on',\n",
              "   'i can help you with a stiudent disocunted price',\n",
              "   'yes i can help you with a discount',\n",
              "   'but we have discounts for you',\n",
              "   'we have a discount offer going in',\n",
              "   'we do have some discounts also',\n",
              "   'the team can help you with discounts',\n",
              "   'we are having some discounts  going on',\n",
              "   'we also have discount offers ..',\n",
              "   'and we are currently running with some discounts',\n",
              "   'we are  running with some discounts',\n",
              "   'we have special discounts going on',\n",
              "   'i will help you with a disocun ted price',\n",
              "   'but we do have some offline offers going on which i can help you with',\n",
              "   'their are some discounts going on',\n",
              "   'we do have some disocunts',\n",
              "   'yes we do have discount',\n",
              "   'we have discount offers as well',\n",
              "   'let me tag you you the discount',\n",
              "   'we have discounts',\n",
              "   'but we have a discount offer going on ...',\n",
              "   'if you want , we ca  help you with some discounts',\n",
              "   'we do have disocunts going on',\n",
              "   'will tag you with discounts',\n",
              "   'we do have some disocunts also',\n",
              "   'we are having few promotions going on',\n",
              "   'we have some discounts going on']},\n",
              " 'discount_student': {'1': ['is there any discounts for students?',\n",
              "   'Also, the student discount is still there right?',\n",
              "   \"I'm a student do I have a discount or something for that? Or that's a fixed price?\",\n",
              "   'Ok. Can you plz confirm the price after student discount ?',\n",
              "   'what is the price with student discount ?',\n",
              "   'for students, can you please less',\n",
              "   'Hi, I wish to enroll but the subscription rate is too high. I am a student in Indiana University. If you have any student discount, that will help me.',\n",
              "   'Hi, do you have discount for students, why it is too expensive? I really need something like that but I can not afford 1200 dollar',\n",
              "   'hi I am a student. Do you have any discount?',\n",
              "   'any student discount?'],\n",
              "  '2': ['for students we have discounted pricing',\n",
              "   'we have an offer on student discounted price',\n",
              "   'we have astudent discounted price',\n",
              "   'will help you with a student discount',\n",
              "   'will provide you the student discounted price',\n",
              "   'will help you with student discount',\n",
              "   'we have student discount',\n",
              "   'for students we have discounts',\n",
              "   'can help you with a student discounted price',\n",
              "   'will help you with the student discounted price',\n",
              "   'we have special student discount',\n",
              "   'we have  dicscounted price for students',\n",
              "   'we have a  student discount price',\n",
              "   'i will help you with a student discounted price',\n",
              "   'see i can help you with a student discounted price']},\n",
              " 'enrollment': {'1': ['sure, How do I subscribe?',\n",
              "   'I wan to suscribe',\n",
              "   'I wanted to know if there are any discount options for yearly subscription',\n",
              "   'I am willing to take a subscription now',\n",
              "   'can I have the subscription options?',\n",
              "   \"to be honest, i'm just checking online if there are specials\",\n",
              "   'Hope I get Cashback',\n",
              "   'How can we subscribe to this?',\n",
              "   'Do I need to pay to use this project?',\n",
              "   'I would love to subscribe for this',\n",
              "   'how to enroll it',\n",
              "   'I am looking to enroll this week',\n",
              "   'where i can enroll with u',\n",
              "   'from where do i enroll?',\n",
              "   'how to enroll',\n",
              "   'can you tell me how I can sign up',\n",
              "   'signup',\n",
              "   'I hope to enrol before the end of the day'],\n",
              "  '2': ['Please use the above link to complete the enrollment.',\n",
              "   'This link will be valid for couple of hours .. so please ensure to complete the enrollment as soon as possible',\n",
              "   'We can send you the link now and you can complete the enrollment',\n",
              "   'Please complete the enrollment using the abouve link',\n",
              "   'please use the above link to complete the enrollment ... this is valid for 1 hr',\n",
              "   'Please use the above link to complete the enrollment as a onetime payment..']},\n",
              " 'project_want': {'1': ['so i am searching for her some real life projects',\n",
              "   'so after I enrol I can choose any project I like?',\n",
              "   'i need simple collage project',\n",
              "   'Hey, i want some thing realtime project practice',\n",
              "   'I want to do projects',\n",
              "   'Wanted some real time projects to showcase',\n",
              "   'and try to find a project and write on  my resume',\n",
              "   'I want to learn something new',\n",
              "   'and need projects to get my hands dirty',\n",
              "   'i am interested to see different project ideas',\n",
              "   'I want to signup for projects course',\n",
              "   'but looking some project for practice.',\n",
              "   'I want to try with few projects',\n",
              "   'I want to do some projects',\n",
              "   'I want to do projects in cloud',\n",
              "   'i like to learn real projects',\n",
              "   'I will get access to all projects right? I mean not only in azure but remaining all like big data etc,.',\n",
              "   'does it ahve access to every',\n",
              "   'with that membership you have access to all the projects that have',\n",
              "   'does the subscription contains all the projects listed on the website?',\n",
              "   'that means I can make use of any project in ProjectDemo?',\n",
              "   'Cool. And when you say all the projects, Do I also get access to other projects in other domain?',\n",
              "   'I saw the courses option - can I access all the available course'],\n",
              "  '2': ['you can access all the projects',\n",
              "   'we have 240+ projects and every month we add 15 to 20 projects',\n",
              "   'we have projects which you can  access',\n",
              "   'yes we have 230+ projects and we add 15 to 20 new projects every month',\n",
              "   'we have 230+ projects as on today and every month we add 15 to 20 projects',\n",
              "   'we provide access to all projects ... you can work on any of them',\n",
              "   'yes , you will get access to all projects',\n",
              "   'You will get access to all the projects',\n",
              "   'you get an access to all the projects',\n",
              "   'For accessing our services you need to enroll with us',\n",
              "   'so for that .. you can enroll.. use the platform to learn .. if you want cancel within 3 months',\n",
              "   'you need to enroll to access the projects',\n",
              "   'you need to enroll with us to access the projects',\n",
              "   'you can enroll with us and get access to all the projects',\n",
              "   'you need to enroll with us to access the projects and the expert sessions',\n",
              "   'Once you sign up , you will get access to all the projects ..',\n",
              "   'its a risk free enrollment',\n",
              "   'you have to enroll to access the projects']},\n",
              " 'refund_info': {'1': [\"If i don't get helpful will I get any refund?\",\n",
              "   \"90 days refund option, i can't see\",\n",
              "   'Do I have money back guaranty',\n",
              "   'but confirm me one thing you said we can ask for refund if we dont like it with 30days are so',\n",
              "   'If i want to dis continue means after 1 month',\n",
              "   \"is there a money back gurantee, in case i'm not satisfied?\",\n",
              "   'If I take subscription through credit card, after 2months I want to cancel it due time concern. So amount will be refunded?',\n",
              "   'HI ChatBot, I still have not completed 90 days. I would like to request a refund',\n",
              "   'is it really 90 days no question asked refund?',\n",
              "   'How the 90 day refund guarantee works?',\n",
              "   'How the  90 day refund guarantee works',\n",
              "   'and refundable within 30 days',\n",
              "   'you sure right I can cancel it within 90 days if I dont like it',\n",
              "   'and what are the cancellation / refund policies ?',\n",
              "   'and any refund ?',\n",
              "   'Also if one uses this portal extensively for 3 months and after that he does not find the need of this portal...then also he can ask for the refund right?',\n",
              "   \"but according to the website, i can ask for refund in 90 days, if i don't like this platform\",\n",
              "   'ok what is the refund policy ?',\n",
              "   '90 days refund policy',\n",
              "   \"I'm here to seek a refund for account  . I want to cancel my subscription and want entire refund\",\n",
              "   '90 day refund guarantee. ???',\n",
              "   'How much money will be refunded if i cancel with in 90 days',\n",
              "   'i can cancel the subscription any time ..will you refund the amount'],\n",
              "  '2': ['we have  a 90 days no question complete refund policy',\n",
              "   \"We provide an unlimited no questions asked refund - you may sign up and if you don't think we are adding value to you - you can ask for a full refund\",\n",
              "   'we have a 90- day refund policy',\n",
              "   \"We provide an unlimited no questions asked refund - you may sign up and if you don't think we are adding value to you - you can ask for a full refund within 90 days\",\n",
              "   '90 days no question complete refund',\n",
              "   'once you enroll and if you  feel that we are not adding value then you have an option to cancel the access',\n",
              "   '90 days refund  policy is applicable on annual enrollment',\n",
              "   'we have 90 days no question complete refund policy',\n",
              "   'in case if you feel that we are not adding value you can cancel your subscription',\n",
              "   'if you want to cancel it you can the 1st 3 months',\n",
              "   \"We provide an unlimited no questions asked refund - you may sign up and if you don't think we are adding value to you - you can ask for a full refund in the 1st 90 days\",\n",
              "   '90  days no question complete refund we provide if you dont like our platform',\n",
              "   'we dont provide trial to our platform since we have this 90 days no question complete refund policy',\n",
              "   'refund policy is applicable in the first 3 months',\n",
              "   'cancellations applicable in fisrt 3 months',\n",
              "   'you can cancele in the 1st 3 months not after that',\n",
              "   '90 days no question completeb refund',\n",
              "   'You can cancel and we will refund anytime in the first 3 months ...',\n",
              "   'once you enroll  you have 90 days time to explore our plat form',\n",
              "   \"We provide an unlimited no questions asked refund - you may sign up and if you don't think we are adding value to you - you can ask for a full refund in the first 90 days\",\n",
              "   'we have a 90 days  no question complete refund policy  that is applicable on 12 months access',\n",
              "   'we have a 90 days no question complete refund policy',\n",
              "   'as we have a 90 days no question refund policy']},\n",
              " 'refund_process_time': {'1': ['Please process the refund',\n",
              "   'okay! This is the first time im trying this service so I would want to have the gaurantee of a full refund if i happen to cancel it sometime before 90 days. And id want to know when would the refund be processed',\n",
              "   'Ok, but how long does it take to get the refund?'],\n",
              "  '2': ['if you are going for a single shot amt it takes 7 to 10 working days to refund',\n",
              "   'the concern team will revert back and it will be processed with in 7 working days',\n",
              "   'we process it immediately',\n",
              "   '7 working days it will be done',\n",
              "   'Any refund process will take 3 to 4 working days']},\n",
              " 'ask_email': {'1': ['Can you share the email please?',\n",
              "   'Can I get an email ID or contact on which i can reach out to someone?'],\n",
              "  '2': ['care@ProjectDemo.io']},\n",
              " 'give_email': {'1': ['Please email me',\n",
              "   'also can you please send me an email to the discount',\n",
              "   'now mail me something for confirmation',\n",
              "   'I did not yet receive any email',\n",
              "   'could you please email it to me',\n",
              "   \"can you send me an email with the link, so that I don't loose it\",\n",
              "   'you send email ?',\n",
              "   'to my mail',\n",
              "   'before that can u send me course content',\n",
              "   'I can pay ₹500 , let me know if that works well for you, email me,  thank you and have a good one',\n",
              "   'ok u can send details over mail',\n",
              "   'have you send me the mail??',\n",
              "   'u can send all details on this',\n",
              "   'But need info to my email id please',\n",
              "   'or you can  send email to my mailbox',\n",
              "   \"I haven't got an email yet\",\n",
              "   'I have written a mail to ProjectDemo',\n",
              "   'Are you sending email regarding tomarrows demo',\n",
              "   'can you send a quote to my email ?',\n",
              "   'pls schedule one and send me a mail pls'],\n",
              "  '2': ['ok sure , i will ask my team to send you an email ...',\n",
              "   'our team will drop you an email',\n",
              "   'our team will drop an email',\n",
              "   'yes my team will send the details to your email address',\n",
              "   'i will ask my team to drop an email',\n",
              "   'let me drop you an email  that will help you',\n",
              "   'you will be receiving an email with the details.',\n",
              "   'will ask my team to email you',\n",
              "   'ok sure ... I will ask my team to drop an email',\n",
              "   'I will ask my team to send you an email',\n",
              "   'my team will drop an email',\n",
              "   'sure , i will ask my team to send you an email',\n",
              "   'our team will drop an email on the same',\n",
              "   'I will ask my team to drop an email ..',\n",
              "   'I will ask my team to drop an email as well',\n",
              "   'i will ask my team to send you an email',\n",
              "   'My team has sent you an email',\n",
              "   'ok sure I will ask my team to drop an email as well']},\n",
              " 'provide_phone_contact': {'2': [\"Sorry, we don't have a phone number. Please provide your number and we will contact you.\"],\n",
              "  '1': ['what is your whatsapp number?',\n",
              "   'do you have any contact no or is this medium to communicate?',\n",
              "   'Any number for doubts or only mail?',\n",
              "   'please provide working contact number of higher authority',\n",
              "   'could you please update contact number... So that I can get notification',\n",
              "   'Please tell me your mob no send me morning i am calling you']},\n",
              " 'noise': {'2': [\"I'm sorry, I don't understand\"],\n",
              "  '1': ['fg',\n",
              "   'poo',\n",
              "   'gay',\n",
              "   'jive',\n",
              "   'poo poo',\n",
              "   'bum',\n",
              "   '#fg',\n",
              "   'are you lost??',\n",
              "   'woow',\n",
              "   'fgfg',\n",
              "   'ff',\n",
              "   'gf',\n",
              "   'where r u Anju',\n",
              "   'sa',\n",
              "   ']',\n",
              "   'gb',\n",
              "   'fd\\\\',\n",
              "   'fsd',\n",
              "   'vf',\n",
              "   'fgdfg',\n",
              "   'fgdfgd',\n",
              "   'gg',\n",
              "   'gfg',\n",
              "   'fgf',\n",
              "   'fd',\n",
              "   'give me na']},\n",
              " 'greetings': {'1': ['Hi',\n",
              "   'Hi What are the difference between 145 and 80?',\n",
              "   'Hi ChatBot',\n",
              "   'Hi ChatBot,',\n",
              "   'Hello,',\n",
              "   'Hey',\n",
              "   'Helloo',\n",
              "   'Hi ChatBot',\n",
              "   'hii',\n",
              "   'Hello!',\n",
              "   'Hii',\n",
              "   'hello?',\n",
              "   'hi',\n",
              "   'Hi!!',\n",
              "   'hello there',\n",
              "   \"hello Ma'am you there?\",\n",
              "   'Hai',\n",
              "   'hello ChatBot',\n",
              "   'Hellooo',\n",
              "   'hey',\n",
              "   'hi, i m back',\n",
              "   'Hello ChatBot',\n",
              "   'hi ChatBot'],\n",
              "  '2': ['Hi', 'Hello']},\n",
              " 'how_are': {'1': ['I am doing great, how are you'],\n",
              "  '2': ['I am fine', \"I'm good thanks for asking\"]},\n",
              " 'i_get_back': {'1': ['I will decide in three days',\n",
              "   'I will get back to this chat.',\n",
              "   'not now..will let u knw..thank u..',\n",
              "   'yes ... but i wont b paying now...some days later',\n",
              "   'let me decide and come to you',\n",
              "   'not sure as of now',\n",
              "   'will think about it',\n",
              "   'i am not in rush to get it today',\n",
              "   \"I'll come back within a hour\",\n",
              "   'I am back, I am sorry, i had to run.',\n",
              "   'will contact soon.',\n",
              "   'ok Let me think and ping you again.',\n",
              "   'Its ok... I need some time to think',\n",
              "   'I will decide',\n",
              "   'ok...will chat after 30 mins',\n",
              "   'ChatBot, can we chat after 30 mins? My wife is calling me for dinner',\n",
              "   'brb',\n",
              "   'Thanks for your offer, but i need some time',\n",
              "   'I will decide then i will message you i will take or not',\n",
              "   'Can I get back to you in some time?',\n",
              "   \"give me some time, I'm looking into them\",\n",
              "   'I will get back, give me some time, Its working like share market',\n",
              "   'In my view my budget is little tight.. so again I will get back to you',\n",
              "   'let me check and ping you in 2 mins',\n",
              "   'moment please',\n",
              "   'right now cant make teh payment',\n",
              "   'I will think about it. Thanks for your time and support',\n",
              "   'Hey I need some time to think',\n",
              "   'okay ..will contact you tomorrow'],\n",
              "  '2': ['Okay, no problem, take your time.']},\n",
              " 'you_get_back': {'1': ['Can you send me an email, I will think about this and will let you know'],\n",
              "  '2': ['My team can connect with you tomorrow']}}"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ],
      "source": [
        "intent_data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 63,
      "id": "b1451aa0-d506-4d4e-b009-2985b47c64ce",
      "metadata": {
        "id": "b1451aa0-d506-4d4e-b009-2985b47c64ce"
      },
      "outputs": [],
      "source": [
        "class PrepareData:\n",
        "    '''Class for preparing data for training.'''\n",
        "\n",
        "    def __init__(self, json_data, intents):\n",
        "        '''Args:\n",
        "            json_data (str): Variable for path to json intents file.\n",
        "            intents (:obj: `list`): List of intents.\n",
        "        '''\n",
        "\n",
        "        # load intents data into dataframe\n",
        "        self.df = pd.DataFrame.from_dict(json_data)\n",
        "\n",
        "        self.intents = intents\n",
        "\n",
        "        self.df_reformat = self._reformat_intents_df()\n",
        "        self.df_preprocessed = self._preprocess_utts()\n",
        "        df_one_hot_labels = self._encode_labels()\n",
        "        self.df_concat = pd.concat([self.df_preprocessed, df_one_hot_labels], axis=1)\n",
        "        self.df_train, self.df_val, self.df_test = self._get_train_val_test()\n",
        "        self.train_x, self.val_x, self.test_x, self.train_y, self.val_y, self.test_y = self._get_input_output()\n",
        "\n",
        "    def _reformat_intents_df(self):\n",
        "        '''Takes data frame, removes response column\n",
        "        and expands query list to individual rows with matching intent.'''\n",
        "\n",
        "        df_T =  self.df.T.reset_index()\n",
        "        df_T.columns = [\"intent\", \"query\", \"response\"]\n",
        "        df_drop_response = df_T.drop(columns=\"response\")\n",
        "        df_drop_response.intent = df_drop_response.intent.astype('category')\n",
        "        df_explode = df_drop_response.explode('query')\n",
        "        return df_explode\n",
        "\n",
        "    def _preprocess_utts(self):\n",
        "        '''Passes utterances through preprocessing\n",
        "        and deduplicates utterances of the same intent label.'''\n",
        "\n",
        "        self.df_reformat['query_preprocessed'] = self.df_reformat.apply(lambda row: Utt(row['query'], query=True).parse_utt(), axis = 1)\n",
        "        df_filter = self.df_reformat.dropna()\n",
        "        df_dedup = df_filter.drop_duplicates(subset=['intent', 'query'], keep='first')\n",
        "        return df_dedup\n",
        "\n",
        "    def _encode_labels(self):\n",
        "        '''One hot encodes labels.'''\n",
        "\n",
        "        df_one_hot_intents = pd.get_dummies(self.df_preprocessed.intent)\n",
        "        return df_one_hot_intents\n",
        "\n",
        "    def _get_train_val_test(self):\n",
        "        '''Split randomized data into train/eval/test with 80:10:10 ratio.'''\n",
        "\n",
        "        df_train, df_val, df_test = np.split(self.df_concat.sample(frac=1, random_state=42),\n",
        "                                             [int(.8*len(self.df_concat)), int(.9*len(self.df_concat))])\n",
        "        return df_train, df_val, df_test\n",
        "\n",
        "    def _get_input_output(self):\n",
        "        '''Converts input and output taxt and labels to numpy arrays.'''\n",
        "\n",
        "        train_queries_list = self.df_train['query_preprocessed'].tolist()\n",
        "        val_queries_list = self.df_val['query_preprocessed'].tolist()\n",
        "        test_queries_list = self.df_test['query_preprocessed'].tolist()\n",
        "\n",
        "        train_x = np.array(train_queries_list, dtype=object)[:, np.newaxis]\n",
        "        val_x = np.array(val_queries_list, dtype=object)[:, np.newaxis]\n",
        "        test_x = np.array(test_queries_list, dtype=object)[:, np.newaxis]\n",
        "\n",
        "        df_train_y = self.df_train.iloc[:,3:]\n",
        "        df_val_y = self.df_val.iloc[:,3:]\n",
        "        df_test_y = self.df_test.iloc[:,3:]\n",
        "\n",
        "        train_y = df_train_y.to_numpy()\n",
        "        val_y = df_val_y.to_numpy()\n",
        "        test_y = df_test_y.to_numpy()\n",
        "\n",
        "        return train_x, val_x, test_x, train_y, val_y, test_y"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "id": "572cf3ba-2047-46be-bc8d-f50f9f2d3f21",
      "metadata": {
        "id": "572cf3ba-2047-46be-bc8d-f50f9f2d3f21"
      },
      "outputs": [],
      "source": [
        "prep = PrepareData(intent_data, intents)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "id": "0e6bc931-0ed1-43e8-b331-4f48731c6f18",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "0e6bc931-0ed1-43e8-b331-4f48731c6f18",
        "outputId": "962583a2-b74d-488c-92ab-4ce9f7f27246"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              intent                                    query  \\\n",
              "0   interview_resume         also can you help me with resume   \n",
              "0   interview_resume       Do you have one to one resume prep   \n",
              "0   interview_resume  also, can i get resume guidance as w...   \n",
              "0   interview_resume  Any interview tips and guidance for ...   \n",
              "0   interview_resume  Do i get resume tips and mock interv...   \n",
              "..               ...                                      ...   \n",
              "39        i_get_back          right now cant make teh payment   \n",
              "39        i_get_back  I will think about it. Thanks for yo...   \n",
              "39        i_get_back            Hey I need some time to think   \n",
              "39        i_get_back         okay ..will contact you tomorrow   \n",
              "40      you_get_back  Can you send me an email, I will thi...   \n",
              "\n",
              "                         query_preprocessed  \n",
              "0           also can you help I with resume  \n",
              "0        do you have one to one resume prep  \n",
              "0      also can get resume guidance as well  \n",
              "0   any interview tip and guidance for I...  \n",
              "0   do get resume tip and mock interview...  \n",
              "..                                      ...  \n",
              "39       right now can not make teh payment  \n",
              "39  will think about it thank for your t...  \n",
              "39              hey need some time to think  \n",
              "39           okay will contact you tomorrow  \n",
              "40  can you send I an email will think a...  \n",
              "\n",
              "[458 rows x 3 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8dab591d-754f-46ab-aca2-881930326c22\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>intent</th>\n",
              "      <th>query</th>\n",
              "      <th>query_preprocessed</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>interview_resume</td>\n",
              "      <td>also can you help me with resume</td>\n",
              "      <td>also can you help I with resume</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>interview_resume</td>\n",
              "      <td>Do you have one to one resume prep</td>\n",
              "      <td>do you have one to one resume prep</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>interview_resume</td>\n",
              "      <td>also, can i get resume guidance as w...</td>\n",
              "      <td>also can get resume guidance as well</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>interview_resume</td>\n",
              "      <td>Any interview tips and guidance for ...</td>\n",
              "      <td>any interview tip and guidance for I...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>interview_resume</td>\n",
              "      <td>Do i get resume tips and mock interv...</td>\n",
              "      <td>do get resume tip and mock interview...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>i_get_back</td>\n",
              "      <td>right now cant make teh payment</td>\n",
              "      <td>right now can not make teh payment</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>i_get_back</td>\n",
              "      <td>I will think about it. Thanks for yo...</td>\n",
              "      <td>will think about it thank for your t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>i_get_back</td>\n",
              "      <td>Hey I need some time to think</td>\n",
              "      <td>hey need some time to think</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>i_get_back</td>\n",
              "      <td>okay ..will contact you tomorrow</td>\n",
              "      <td>okay will contact you tomorrow</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>you_get_back</td>\n",
              "      <td>Can you send me an email, I will thi...</td>\n",
              "      <td>can you send I an email will think a...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>458 rows × 3 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8dab591d-754f-46ab-aca2-881930326c22')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-8dab591d-754f-46ab-aca2-881930326c22 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-8dab591d-754f-46ab-aca2-881930326c22');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-ea67e288-19c9-444d-92da-23b7ae76d935\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ea67e288-19c9-444d-92da-23b7ae76d935')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-ea67e288-19c9-444d-92da-23b7ae76d935 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ],
      "source": [
        "prep.df_preprocessed"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "id": "bfbdcd3e-9faa-49f9-9ff9-399b69c71965",
      "metadata": {
        "id": "bfbdcd3e-9faa-49f9-9ff9-399b69c71965"
      },
      "outputs": [],
      "source": [
        "train_x, val_x, test_x, train_y, val_y, test_y = prep.train_x, prep.val_x, prep.test_x, prep.train_y, prep.val_y, prep.test_y"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "f85cdbb4-b1df-4adf-b4a0-aaa622cdb5ea",
      "metadata": {
        "id": "f85cdbb4-b1df-4adf-b4a0-aaa622cdb5ea"
      },
      "source": [
        "## Train model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "id": "615daf04-3d03-4b26-8a03-c3c6494e25b3",
      "metadata": {
        "tags": [],
        "id": "615daf04-3d03-4b26-8a03-c3c6494e25b3"
      },
      "outputs": [],
      "source": [
        "class ModelTrain:\n",
        "    '''Class tuning model and then training.'''\n",
        "\n",
        "    def __init__(self, train_x, val_x, train_y, val_y, module_url):\n",
        "        '''Args:\n",
        "            train_x (array): Numpy array of training utterances.\n",
        "            val_x (array): Numpy array of validation utterances.\n",
        "            train_y (array): Numpy array of one hot encoded training labels.\n",
        "            val_y (array): Numpy array of one hot encoded validation labels.\n",
        "            module_url (str): Variable for url to pretrained Universal\n",
        "            Sentence Encoder model.\n",
        "        '''\n",
        "\n",
        "        self.train_x, self.val_x, self.train_y, self.val_y = train_x, val_x, train_y, val_y\n",
        "\n",
        "        # initialise input layer with pretrained USE weights\n",
        "        self.embed = hub.KerasLayer(module_url, input_shape=[], dtype=tf.string, trainable=True)\n",
        "\n",
        "        # hyperparamter tuning\n",
        "        self.tuner = keras_tuner.BayesianOptimization(\n",
        "            self._build_model,\n",
        "            objective=\"val_loss\",\n",
        "            max_trials=25,\n",
        "            overwrite=True,\n",
        "            directory=\"/content/drive/MyDrive/code/outputs/hp_dir\",\n",
        "            project_name=\"tune_hypermodel\"\n",
        "        )\n",
        "\n",
        "        # Early stopping\n",
        "        self.early_stopping = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
        "\n",
        "        self._tune_parameters()\n",
        "\n",
        "        # print best hyperparameters\n",
        "        best_hps = self.tuner.get_best_hyperparameters(num_trials=1)[0]\n",
        "\n",
        "        # configure model with best hyperparameters\n",
        "        self.model = self.tuner.hypermodel.build(best_hps)\n",
        "\n",
        "        self.best_epoch = self._get_best_epoch()\n",
        "\n",
        "        # monitor training on tensorboard\n",
        "        log_dir = \"log/intent_recognition/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%s\")\n",
        "        self.tensorboard_cb = keras.callbacks.TensorBoard(log_dir=log_dir)\n",
        "\n",
        "        self._train()\n",
        "\n",
        "        print('Training complete!')\n",
        "\n",
        "        # save model\n",
        "        self.model.save(\"/content/drive/MyDrive/code/outputs/saved_model_keras\")\n",
        "\n",
        "    def _tune_parameters(self):\n",
        "        '''Performs hyperparameter tuning.'''\n",
        "\n",
        "        self.tuner.search(\n",
        "            self.train_x,\n",
        "            self.train_y,\n",
        "            epochs=10,\n",
        "            validation_data=(self.val_x, self.val_y),\n",
        "            callbacks=[self.early_stopping]\n",
        "        )\n",
        "\n",
        "    def _get_best_epoch(self):\n",
        "        '''Gets best epoch number for training.'''\n",
        "\n",
        "        history = self.model.fit(\n",
        "            self.train_x,\n",
        "            self.train_y,\n",
        "            epochs=50,\n",
        "            validation_data=(self.val_x, self.val_y),\n",
        "            callbacks=[self.early_stopping],\n",
        "            workers=4,\n",
        "            use_multiprocessing=True\n",
        "        )\n",
        "\n",
        "        val_loss_per_epoch = history.history['val_loss']\n",
        "        best_epoch = val_loss_per_epoch.index(max(val_loss_per_epoch)) + 1\n",
        "        print(f'Best epoch: {best_epoch}')\n",
        "\n",
        "        return best_epoch\n",
        "\n",
        "    def _build_model(self, hp):\n",
        "        '''Compile model'''\n",
        "\n",
        "        model = tf.keras.models.Sequential()\n",
        "        model.add(self.embed)\n",
        "        for i in range(hp.Int(\"num_layers\", 1, 3)):\n",
        "            model.add(\n",
        "                layers.Dense(\n",
        "                    units=hp.Int(\"units\", min_value=128, max_value=512, step=128),\n",
        "                    activation=hp.Choice(\"activation\", [\"relu\", \"tanh\", \"sigmoid\"])\n",
        "                )\n",
        "            )\n",
        "            model.add(\n",
        "                layers.Dropout(\n",
        "                    hp.Choice('dropout_rate', values=[0.1, 0.3, 0.5])\n",
        "                )\n",
        "            )\n",
        "        model.add(layers.Dense(41, activation=\"softmax\"))\n",
        "        learning_rate = hp.Choice('lr', values=[1e-2, 1e-3, 1e-4])\n",
        "        model.compile(\n",
        "            loss='categorical_crossentropy',\n",
        "            optimizer=keras.optimizers.Adam(learning_rate=learning_rate),\n",
        "            metrics=['categorical_accuracy']\n",
        "        )\n",
        "        return model\n",
        "\n",
        "    def get_search_summary(self):\n",
        "        '''Get summary of hyperparameter search space.'''\n",
        "\n",
        "        return self.tuner.search_space_summary()\n",
        "\n",
        "    def get_results_summary(self):\n",
        "        '''Get summary of tuning results.'''\n",
        "\n",
        "        return self.tuner.results_summary()\n",
        "\n",
        "    def get_model_summary(self):\n",
        "        '''Get summary of final model.'''\n",
        "\n",
        "        return self.model.summary()\n",
        "\n",
        "    def get_model_diagram(self):\n",
        "        '''Get diagram of model architecture.'''\n",
        "\n",
        "        return keras.utils.plot_model(self.model, \"Model_Diagram.png\", show_shapes=True)\n",
        "\n",
        "    def _train(self):\n",
        "        '''Train the model.'''\n",
        "\n",
        "        # Train the model one final time while saving the best model.\n",
        "        history = self.model.fit(\n",
        "            self.train_x,\n",
        "            self.train_y,\n",
        "            epochs=self.best_epoch,\n",
        "            validation_data=(self.val_x, self.val_y),\n",
        "            callbacks=[self.early_stopping, self.tensorboard_cb],\n",
        "            workers=4,\n",
        "            use_multiprocessing=True\n",
        "        )\n",
        "\n",
        "        with open('/content/drive/MyDrive/code/outputs/HistoryDict', 'wb') as f:\n",
        "            pickle.dump(history.history, f)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "id": "8c80f13c-38a5-4bb0-9110-aa2dda4cc813",
      "metadata": {
        "id": "8c80f13c-38a5-4bb0-9110-aa2dda4cc813"
      },
      "outputs": [],
      "source": [
        "# train = ModelTrain()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "id": "b33da1a2-03e7-4403-a0d3-1cbf175688b8",
      "metadata": {
        "id": "b33da1a2-03e7-4403-a0d3-1cbf175688b8"
      },
      "outputs": [],
      "source": [
        "# train.get_search_summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "id": "e70ac2d0-4666-4234-94a4-b4cb6762b202",
      "metadata": {
        "id": "e70ac2d0-4666-4234-94a4-b4cb6762b202"
      },
      "outputs": [],
      "source": [
        "# train.get_results_summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "id": "de197d23-120d-4dd3-9287-56445fe7b7d8",
      "metadata": {
        "id": "de197d23-120d-4dd3-9287-56445fe7b7d8"
      },
      "outputs": [],
      "source": [
        "# train.get_model_summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "id": "422d06f6-28f7-4876-9055-7041bf03bab4",
      "metadata": {
        "id": "422d06f6-28f7-4876-9055-7041bf03bab4"
      },
      "outputs": [],
      "source": [
        "# train.get_model_diagram()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "d448062d-1b7b-4074-a996-440780a3ba08",
      "metadata": {
        "id": "d448062d-1b7b-4074-a996-440780a3ba08"
      },
      "source": [
        "## Evaluate model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 36,
      "id": "2b9ca9cc-8e50-4368-b5cb-17b81563922c",
      "metadata": {
        "id": "2b9ca9cc-8e50-4368-b5cb-17b81563922c"
      },
      "outputs": [],
      "source": [
        "class Eval:\n",
        "    '''Class for model evaluation.'''\n",
        "\n",
        "    def __init__(self, model, td, intents, test_x, test_y):\n",
        "        '''Args:\n",
        "            model (str): Variable name for model path.\n",
        "            td (:obj:, `dataframe`): Dataframe from prapare data object.\n",
        "            intents (:obj:`list`): Ordered list of intent labels.\n",
        "            test_x (array): Numpy array of test utterances.\n",
        "            test_y (array): Numpy array of one hot encoded test labels.\n",
        "        '''\n",
        "\n",
        "        self.test_x = test_x\n",
        "        self.test_y = test_y\n",
        "\n",
        "        self.intents = intents\n",
        "\n",
        "        # get column containing labels of test data\n",
        "        self.df_test = td.df_test.iloc[:, 0:1]\n",
        "\n",
        "        # get label predictions of test inputs\n",
        "        predicts = model.predict(self.test_x, batch_size=32)\n",
        "\n",
        "        # Get predicted labels that are above 0.5 probability\n",
        "        self.predict_logits = predicts.argmax(axis=1)\n",
        "\n",
        "        self.predicted_labels = [self.intents[logit] for logit in self.predict_logits]\n",
        "\n",
        "        self.test_y_rounded = np.argmax(self.test_y, axis=1)\n",
        "        self.model = model\n",
        "\n",
        "    def get_test_loss_acc(self):\n",
        "        '''Get accuracy and loss of test data.'''\n",
        "\n",
        "        test_scores = self.model.evaluate(self.test_x, self.test_y, verbose=2)\n",
        "        print(\"Test loss:\", test_scores[0])\n",
        "        print(\"Test accuracy:\", test_scores[1])\n",
        "\n",
        "\n",
        "    def get_accuracy_plot(self, history):\n",
        "        '''Plot test accuracy against training accuracy.'''\n",
        "\n",
        "        plt.plot(history['accuracy'])\n",
        "        plt.plot(history['val_accuracy'])\n",
        "        plt.title('Model Accuracy')\n",
        "        plt.ylabel('Accuracy')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'], loc='upper left')\n",
        "        plt.show()\n",
        "\n",
        "    def get_loss_plot(self, history):\n",
        "        '''Plot test loss against training loss.'''\n",
        "\n",
        "        plt.plot(history['loss'])\n",
        "        plt.plot(history['val_loss'])\n",
        "        plt.title('Model Loss')\n",
        "        plt.ylabel('Loss')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'], loc='upper left')\n",
        "        plt.show()\n",
        "\n",
        "    def compare_predicted_intents(self):\n",
        "        '''Get table of predicted vs actuals intent labels.'''\n",
        "\n",
        "        self.df_test['predicted_intent'] = self.predicted_labels\n",
        "        print(self.df_test)\n",
        "\n",
        "    def get_fscore(self):\n",
        "        '''Get f-score of predicted labels.'''\n",
        "\n",
        "        return pd.DataFrame(classification_report(self.test_y_rounded, self.predict_logits, output_dict=True)).T\n",
        "\n",
        "    def get_confusion_matrix(self):\n",
        "        '''Get confusion matrix of predicted vs actual labels.'''\n",
        "\n",
        "        cm = confusion_matrix(self.test_y_rounded, self.predict_logits)\n",
        "        cm_display = sklearn.metrics.ConfusionMatrixDisplay(confusion_matrix = cm)\n",
        "        cm_display.plot()\n",
        "        plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "id": "db66e594-6a0d-454c-afa5-8d117f560d85",
      "metadata": {
        "id": "db66e594-6a0d-454c-afa5-8d117f560d85"
      },
      "outputs": [],
      "source": [
        "model = load_model(MODEL_PATH)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "id": "4df21cdd-dd96-466e-aaf3-bf35c85a3d05",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4df21cdd-dd96-466e-aaf3-bf35c85a3d05",
        "outputId": "ffb2feec-b6ab-4f55-cf8d-f7828c27dcac"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2/2 [==============================] - 1s 19ms/step\n"
          ]
        }
      ],
      "source": [
        "evaluate = Eval(model, prep, intents, test_x, test_y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "id": "2ee3ddc5-eacf-4a72-adc8-fc446f0b0460",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2ee3ddc5-eacf-4a72-adc8-fc446f0b0460",
        "outputId": "e3606b99-6c61-4fc9-f910-8baac9742857"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2/2 - 1s - loss: 6.8636 - accuracy: 0.2174 - 572ms/epoch - 286ms/step\n",
            "Test loss: 6.863568305969238\n",
            "Test accuracy: 0.21739129722118378\n"
          ]
        }
      ],
      "source": [
        "evaluate.get_test_loss_acc()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "id": "f46d5158-533f-4fb3-8d43-9e56f04ad424",
      "metadata": {
        "id": "f46d5158-533f-4fb3-8d43-9e56f04ad424"
      },
      "outputs": [],
      "source": [
        "history = pickle.load(open(MODEL_HISTORY, \"rb\"))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "id": "c8e79408-ede9-4306-8c92-a7f38f1ccb23",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "c8e79408-ede9-4306-8c92-a7f38f1ccb23",
        "outputId": "0bbc2720-779f-4f4b-c7f6-9b164d65628e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABoeUlEQVR4nO3deVhUZf8G8HtmgGHfZEcEd8UFVJTUNovCJVKzXEsz07dS03hbNHNrkcoyMk2rn0u5ZZqab5amWJq5g7gvuQLKjjAsMgMz5/fHgZERUJCBM8zcn+s6l8OZc858B5S5fZ7nPI9MEAQBRERERGZCLnUBRERERMbEcENERERmheGGiIiIzArDDREREZkVhhsiIiIyKww3REREZFYYboiIiMisMNwQERGRWWG4ISIiIrPCcENERiOTyTBnzpxan3f16lXIZDKsXLnS6DURkeVhuCEyMytXroRMJoNMJsO+ffsqPS8IAgICAiCTyfDUU09JUKFx/Pbbb5DJZPDz84NOp5O6HCIyIQw3RGbK1tYWa9eurbR/z549SElJgVKplKAq41mzZg2CgoKQmpqK3bt3S10OEZkQhhsiM9W/f39s2LABpaWlBvvXrl2Lbt26wcfHR6LK6q6wsBC//PILoqOj0aVLF6xZs0bqkqpVWFgodQlEFofhhshMjRgxAtnZ2di5c6d+n0ajwcaNGzFy5MgqzyksLMR///tfBAQEQKlUom3btvjss88gCILBcWq1Gm+88QY8PT3h5OSEp59+GikpKVVe8/r163jppZfg7e0NpVKJDh06YPny5XV6b5s3b8atW7fw3HPPYfjw4di0aROKi4srHVdcXIw5c+agTZs2sLW1ha+vL5555hlcunRJf4xOp8OXX36JTp06wdbWFp6enujbty+OHj0K4O7jge4cYzRnzhzIZDKcOXMGI0eOhJubGx588EEAwIkTJ/Diiy+iRYsWsLW1hY+PD1566SVkZ2dX+T0bN24c/Pz8oFQq0bx5c7z66qvQaDS4fPkyZDIZvvjii0rn7d+/HzKZDOvWravtt5TIrFhJXQAR1Y+goCD07NkT69atQ79+/QAAv//+O/Ly8jB8+HAsXLjQ4HhBEPD000/jzz//xLhx4xAaGoodO3bgrbfewvXr1w0+TF9++WWsXr0aI0eORK9evbB7924MGDCgUg3p6el44IEHIJPJMGnSJHh6euL333/HuHHjoFKpMHXq1Pt6b2vWrEGfPn3g4+OD4cOHY9q0afjf//6H5557Tn+MVqvFU089hbi4OAwfPhxTpkxBfn4+du7ciVOnTqFly5YAgHHjxmHlypXo168fXn75ZZSWluLvv//GwYMHERYWdl/1Pffcc2jdujXmzZunD4Y7d+7E5cuXMXbsWPj4+OD06dP49ttvcfr0aRw8eBAymQwAcOPGDfTo0QO5ubmYMGEC2rVrh+vXr2Pjxo0oKipCixYt0Lt3b6xZswZvvPFGpe+Lk5MTBg4ceF91E5kNgYjMyooVKwQAwpEjR4RFixYJTk5OQlFRkSAIgvDcc88Jffr0EQRBEAIDA4UBAwboz9uyZYsAQPjwww8Nrvfss88KMplMuHjxoiAIgpCYmCgAEF577TWD40aOHCkAEGbPnq3fN27cOMHX11fIysoyOHb48OGCi4uLvq4rV64IAIQVK1bc8/2lp6cLVlZWwnfffaff16tXL2HgwIEGxy1fvlwAICxYsKDSNXQ6nSAIgrB7924BgPD6669Xe8zdarvz/c6ePVsAIIwYMaLSseXvtaJ169YJAIS9e/fq940ePVqQy+XCkSNHqq3pm2++EQAIZ8+e1T+n0WgEDw8PYcyYMZXOI7I07JYiMmNDhw7FrVu38OuvvyI/Px+//vprtV1Sv/32GxQKBV5//XWD/f/9738hCAJ+//13/XEAKh13ZyuMIAj4+eefERUVBUEQkJWVpd8iIyORl5eHhISEWr+nH3/8EXK5HEOGDNHvGzFiBH7//XfcvHlTv+/nn3+Gh4cHJk+eXOka5a0kP//8M2QyGWbPnl3tMffjlVdeqbTPzs5O/7i4uBhZWVl44IEHAED/fdDpdNiyZQuioqKqbDUqr2no0KGwtbU1GGu0Y8cOZGVl4fnnn7/vuonMBcMNkRnz9PREREQE1q5di02bNkGr1eLZZ5+t8thr167Bz88PTk5OBvvbt2+vf778T7lcru/WKde2bVuDrzMzM5Gbm4tvv/0Wnp6eBtvYsWMBABkZGbV+T6tXr0aPHj2QnZ2Nixcv4uLFi+jSpQs0Gg02bNigP+7SpUto27YtrKyq732/dOkS/Pz84O7uXus67qZ58+aV9uXk5GDKlCnw9vaGnZ0dPD099cfl5eUBEL9nKpUKHTt2vOv1XV1dERUVZXA33Jo1a+Dv74/HHnvMiO+EqHHimBsiMzdy5EiMHz8eaWlp6NevH1xdXRvkdcvnnnn++ecxZsyYKo/p3Llzra7577//4siRIwCA1q1bV3p+zZo1mDBhQi0rvbvqWnC0Wm2151RspSk3dOhQ7N+/H2+99RZCQ0Ph6OgInU6Hvn373tc8PaNHj8aGDRuwf/9+dOrUCVu3bsVrr70GuZz/ZyViuCEyc4MHD8Z//vMfHDx4EOvXr6/2uMDAQOzatQv5+fkGrTfnzp3TP1/+p06n07eMlDt//rzB9crvpNJqtYiIiDDKe1mzZg2sra2xatUqKBQKg+f27duHhQsXIikpCc2aNUPLli1x6NAhlJSUwNrausrrtWzZEjt27EBOTk61rTdubm4AgNzcXIP95S1ZNXHz5k3ExcVh7ty5mDVrln7/v//+a3Ccp6cnnJ2dcerUqXtes2/fvvD09MSaNWsQHh6OoqIivPDCCzWuicicMeITmTlHR0csWbIEc+bMQVRUVLXH9e/fH1qtFosWLTLY/8UXX0Amk+nvuCr/8867rWJjYw2+VigUGDJkCH7++ecqP6wzMzNr/V7WrFmDhx56CMOGDcOzzz5rsL311lsAoL8NesiQIcjKyqr0fgDo72AaMmQIBEHA3Llzqz3G2dkZHh4e2Lt3r8HzX3/9dY3rLg9iwh231N/5PZPL5Rg0aBD+97//6W9Fr6omALCyssKIESPw008/YeXKlejUqVOtW8KIzBVbbogsQHXdQhVFRUWhT58+mDFjBq5evYqQkBD88ccf+OWXXzB16lT9GJvQ0FCMGDECX3/9NfLy8tCrVy/ExcXh4sWLla758ccf488//0R4eDjGjx+P4OBg5OTkICEhAbt27UJOTk6N38OhQ4dw8eJFTJo0qcrn/f390bVrV6xZswbvvPMORo8ejR9++AHR0dE4fPgwHnroIRQWFmLXrl147bXXMHDgQPTp0wcvvPACFi5ciH///VffRfT333+jT58++td6+eWX8fHHH+Pll19GWFgY9u7diwsXLtS4dmdnZzz88MP49NNPUVJSAn9/f/zxxx+4cuVKpWPnzZuHP/74A4888ggmTJiA9u3bIzU1FRs2bMC+ffsMuhVHjx6NhQsX4s8//8Qnn3xS43qIzJ50N2oRUX2oeCv43dx5K7ggCEJ+fr7wxhtvCH5+foK1tbXQunVrYf78+fpbkMvdunVLeP3114UmTZoIDg4OQlRUlJCcnFzp1mhBEG/dnjhxohAQECBYW1sLPj4+wuOPPy58++23+mNqciv45MmTBQDCpUuXqj1mzpw5AgDh+PHjgiCIt1/PmDFDaN68uf61n332WYNrlJaWCvPnzxfatWsn2NjYCJ6enkK/fv2E+Ph4/TFFRUXCuHHjBBcXF8HJyUkYOnSokJGRUe2t4JmZmZVqS0lJEQYPHiy4uroKLi4uwnPPPSfcuHGjyu/ZtWvXhNGjRwuenp6CUqkUWrRoIUycOFFQq9WVrtuhQwdBLpcLKSkp1X5fiCyNTBDuaCclIqJGo0uXLnB3d0dcXJzUpRCZDI65ISJqpI4ePYrExESMHj1a6lKITApbboiIGplTp04hPj4en3/+ObKysnD58mXY2tpKXRaRyWDLDRFRI7Nx40aMHTsWJSUlWLduHYMN0R3YckNERERmRdKWm7179yIqKgp+fn6QyWTYsmXLPc/566+/0LVrVyiVSrRq1QorV66s9zqJiIio8ZA03BQWFiIkJASLFy+u0fFXrlzBgAED0KdPHyQmJmLq1Kl4+eWXsWPHjnqulIiIiBoLk+mWkslk2Lx5MwYNGlTtMe+88w62bdtmMNvp8OHDkZubi+3bt9fodXQ6HW7cuAEnJ6c6rfpLREREDUcQBOTn58PPz++ea6g1qhmKDxw4UGmNmsjISEydOrXac9RqNdRqtf7r69evIzg4uL5KJCIionqUnJyMpk2b3vWYRhVu0tLS4O3tbbDP29sbKpUKt27dqnIl3piYmCrXjUlOToazs3O91UpERETGo1KpEBAQYLCwb3UaVbi5H9OnT0d0dLT+6/JvjrOzM8MNERFRI1OTISWNKtz4+PggPT3dYF96ejqcnZ2rbLUBAKVSCaVS2RDlERERkQloVJP49ezZs9L6KTt37kTPnj0lqoiIiIhMjaThpqCgAImJiUhMTAQg3uqdmJiIpKQkAGKXUsU1U1555RVcvnwZb7/9Ns6dO4evv/4aP/30E9544w0pyiciIiITJGm31NGjR9GnTx/91+VjY8aMGYOVK1ciNTVVH3QAoHnz5ti2bRveeOMNfPnll2jatCn+7//+D5GRkUavTavVoqSkxOjXpYZnbW0NhUIhdRlERNRATGaem4aiUqng4uKCvLy8KgcUC4KAtLQ05ObmNnxxVG9cXV3h4+PDuY2IiBqpe31+V9SoBhQ3hPJg4+XlBXt7e34YNnKCIKCoqAgZGRkAAF9fX4krIiKi+sZwU4FWq9UHmyZNmkhdDhlJ+Z10GRkZ8PLyYhcVEZGZa1R3S9W38jE29vb2EldCxlb+M+U4KiIi88dwUwV2RZkf/kyJiCwHww0RERGZFYYbqlZQUBBiY2OlLoOIiKhWGG7MgEwmu+s2Z86c+7rukSNHMGHCBOMWS0REVM94t5QZSE1N1T9ev349Zs2ahfPnz+v3OTo66h8LggCtVgsrq3v/6D09PY1bKBERmTV1qRY3C0tQqtOhqZt0N+cw3JgBHx8f/WMXFxfIZDL9vr/++gt9+vTBb7/9hvfeew8nT57EH3/8gYCAAERHR+PgwYMoLCxE+/btERMTg4iICP21goKCMHXqVEydOhWA2EL03XffYdu2bdixYwf8/f3x+eef4+mnn27Q90tERPVPEASobpUip0iDnEI1cgpLqv6zSPzzZmEJCtSlAICeLZpg3YQHJKud4eYeBEHArRKtJK9tZ60w2l0+06ZNw2effYYWLVrAzc0NycnJ6N+/Pz766CMolUr88MMPiIqKwvnz59GsWbNqrzN37lx8+umnmD9/Pr766iuMGjUK165dg7u7u1HqJCKi+lHeqpJTqBG3Ig1yCgzDSbb+Tw1yizQo1dV+EQOFXAadxIsfMNzcw60SLYJn7ZDktc+8Hwl7G+P8iN5//3088cQT+q/d3d0REhKi//qDDz7A5s2bsXXrVkyaNKna67z44osYMWIEAGDevHlYuHAhDh8+jL59+xqlTiIiujdBEKAqLr0dVAo1uFmoQXahBjeLNMguKPuzbH9OoUbfqlJbDjYKuDvawN3eBu4ONnBzsEGTin/a26CJY9mfDko42VpBLpd2+g2GGwsRFhZm8HVBQQHmzJmDbdu2ITU1FaWlpbh165bBQqVV6dy5s/6xg4MDnJ2d9UsbEBE1NEEQkF2oQam28S+TqBME5N0qMQgshq0stx/fLLz/VhU3e2sxpBiEEjGsuN+xudnbwNa68c3qznBzD3bWCpx53/irjtf0tY3FwcHB4Os333wTO3fuxGeffYZWrVrBzs4Ozz77LDQazV2vY21tbfC1TCaDTqczWp1ERHeTd6sEJ1JykZiUi8TkXBxPyUVWwd1/b5kzexuFYSApa12p2NJScXO2tZa8VaUhMNzcg0wmM1rXkCn5559/8OKLL2Lw4MEAxJacq1evSlsUEVEFmlIdzqWpcDw5F8eSxTBzObOw0nEyGaAwg1nIZTLA2db6nl0/FcNKY2xVaQjm96lNNdK6dWts2rQJUVFRkMlkmDlzJltgiEgygiAgOecWjiXfFFtkknNx6oYKmtLKv5eaudsjJMAVoWVbBz9nfsiTAYYbC7VgwQK89NJL6NWrFzw8PPDOO+9ApVJJXRYRWYi8ohIk6ruXbuJ4Sh5yCit3L7nYWVcIMi4IaeqKJo5KCSqmxkQmCBLfr9XAVCoVXFxckJeXB2dnZ4PniouLceXKFTRv3hy2trYSVUj1gT9bIumoS7U4m5qP42VdS4nJubiSVbl7yVohQ7CvsxhkmrkiNMANQU3sufAtAbj75/ed2HJDRERGIwgCrmUX6UNMYnIuztxQQaOt3L0U1MQeoQGu+paZYD9nKK3YvUR1x3BDRET37WahpkL3knj3Um5RSaXjXO2t9WNkQgJcEdrUFW4ONhJUTJaA4YaIiGpEXarF6Rsqg+6la9lFlY6zUcgR7Cd2L3Vp5oqQpq4IZPcSNSCGGyIiqkQQBFzJKtTfuZSYnIszqSqUVDFZXgsPB4Pupfa+zrCxkktQNZGI4YaI6B4K1aU4l5aPM6kqnLmhQsrNIljJZbCxksPGSgFrhQxKKzlsFPKyfXLYKBSwsZLffq6K/TZW8rLzFPrn9fsr7FM0wKRr2QVqHC/vXkrJw/HkXOTdqty95O5gY9C9FNLUBa727F4i08JwQ0RURhAEZOSrceaGSh9kzqSqcDW7EFLeV6qQy2CjKA8+itthSWEYiGysFLBRyPXP6/dXCErlIcxaIUOJVsCJ63lITL6J5JxblV7XxkqOjn7OCA1wQ2gzV3QJcEVTNzt2L5HJY7ghIotUqtXhclYhztxQ4Wzq7TCTXcVcKwDg7axEsK8zgv2cEdTEAQLEGXQ1pTpotDr94xKtDuo79pXv15Q/d8d5JRWOVZc9rkirE3BLp4XYkHJ/ix/WREtPBzHIBLggNMAN7XydYK1g9xI1Pgw3RGT2CtSlOJdq2BpzPi0f6ipmv1XIZWjp6YBgX2e0Lwsz7X2d4dGAE8cJgoASrXA79JT9qb4jKOkDUTVBqWKYunN/eYBq7+uEkABXdG7qChc763tURtQ4MNwQkdkQBAFpqmIxwJR3LaWqqryjBwAcbBT6AFPeKtPG20nyqfxlMhlsrMQuJQdOxktUaww3BAB49NFHERoaitjYWABAUFAQpk6diqlTp1Z7jkwmw+bNmzFo0KA6vbaxrkOWpUSrw6XMArFLqcIYmZtVzLECAL4utvoAU/5ngJu9RayQTGRpGG7MQFRUFEpKSrB9+/ZKz/399994+OGHcfz4cXTu3LnG1zxy5AgcHByMWSbmzJmDLVu2IDEx0WB/amoq3NzcjPpaZF5UxSU4l5qPMzfy9K0xF9IKqpz1ViGXobWXo9giU6FbyZ0TxhFZDIYbMzBu3DgMGTIEKSkpaNq0qcFzK1asQFhYWK2CDQB4enoas8S78vHxabDXItMmCAJu5FXsVsrD2dR8JOVU3a3kqLSq1BrTystR8m4lIpIWh8Gbgaeeegqenp5YuXKlwf6CggJs2LABgwYNwogRI+Dv7w97e3t06tQJ69atu+s1g4KC9F1UAPDvv//i4Ycfhq2tLYKDg7Fz585K57zzzjto06YN7O3t0aJFC8ycORMlJWIXwcqVKzF37lwcP34cMpkMMplMX69MJsOWLVv01zl58iQee+wx2NnZoUmTJpgwYQIKCgr0z7/44osYNGgQPvvsM/j6+qJJkyaYOHGi/rWocdCU6nA2VYWf41Pwwa9nMOLbgwh9fyd6f7wb4384ii92XcCO0+n6YOPvaoeI9l54/bFWWPp8V+x9qw9OzH4SP73SE3Oe7oCh3QPQ0d+FwYaI2HJzT4IAlFT9v8Z6Z20P1GA+CSsrK4wePRorV67EjBkz9HNQbNiwAVqtFs8//zw2bNiAd955B87Ozti2bRteeOEFtGzZEj169Ljn9XU6HZ555hl4e3vj0KFDyMvLq3IsjpOTE1auXAk/Pz+cPHkS48ePh5OTE95++20MGzYMp06dwvbt27Fr1y4AgIuLS6VrFBYWIjIyEj179sSRI0eQkZGBl19+GZMmTTIIb3/++Sd8fX3x559/4uLFixg2bBhCQ0Mxfvz4e74faniCIOBsaj4OXs7Wj425mFF1t5KVXIZWXo4GrTHBvs6cKI6Iaozh5l5KioB5ftK89rs3AJuajXt56aWXMH/+fOzZswePPvooALFLasiQIQgMDMSbb76pP3by5MnYsWMHfvrppxqFm127duHcuXPYsWMH/PzE78W8efPQr18/g+Pee+89/eOgoCC8+eab+PHHH/H222/Dzs4Ojo6OsLKyums31Nq1a1FcXIwffvhBP+Zn0aJFiIqKwieffAJvb28AgJubGxYtWgSFQoF27dphwIABiIuLY7gxITqdgGPJudhxOg3bT6VV2bXkZGvYrdTe1xmtvR25MjQR1QnDjZlo164devXqheXLl+PRRx/FxYsX8ffff+P999+HVqvFvHnz8NNPP+H69evQaDRQq9Wwt7ev0bXPnj2LgIAAfbABgJ49e1Y6bv369Vi4cCEuXbqEgoIClJaWwtnZuVbv4+zZswgJCTEYzNy7d2/odDqcP39eH246dOgAheL2B6Cvry9OnjxZq9ci4yvV6nDoSg62n0rDjtNpyMhX659TWsnRu5UHOvm76MMMZ7slovrAcHMv1vZiC4pUr10L48aNw+TJk7F48WKsWLECLVu2xCOPPIJPPvkEX375JWJjY9GpUyc4ODhg6tSp0Giqnon1fhw4cACjRo3C3LlzERkZCRcXF/z444/4/PPPjfYaFVlbG042JpPJoNNV7uKg+ldcosU/F7Ow/VQadp5NR26FW7EdlVZ4vL0X+nbwwSNtPWFvw185RFT/+JvmXmSyGncNSW3o0KGYMmUK1q5dix9++AGvvvoqZDIZ/vnnHwwcOBDPP/88AHEMzYULFxAcHFyj67Zv3x7JyclITU2Fr68vAODgwYMGx+zfvx+BgYGYMWOGft+1a9cMjrGxsYFWq73na61cuRKFhYX61pt//vkHcrkcbdu2rVG9VP8K1KX463wGtp9Kw5/nMlCouf1zdXewwRPtvdG3ow96tWrCLiYianAMN2bE0dERw4YNw/Tp06FSqfDiiy8CAFq3bo2NGzdi//79cHNzw4IFC5Cenl7jcBMREYE2bdpgzJgxmD9/PlQqlUGIKX+NpKQk/Pjjj+jevTu2bduGzZs3GxwTFBSEK1euIDExEU2bNoWTkxOUSsPpV0eNGoXZs2djzJgxmDNnDjIzMzF58mS88MIL+i4pksbNQg12nU3HjtNp2PtvlsH6Rz7Otujb0QeRHXzQPcgNVlyPiIgkxHBjZsaNG4dly5ahf//++jEy7733Hi5fvozIyEjY29tjwoQJGDRoEPLy8mp0Tblcjs2bN2PcuHHo0aMHgoKCsHDhQvTt21d/zNNPP4033ngDkyZNglqtxoABAzBz5kzMmTNHf8yQIUOwadMm9OnTB7m5uVixYoU+gJWzt7fHjh07MGXKFHTv3h329vYYMmQIFixYUOfvDdVeuqoYf5xOw/bTaTh4OQda3e2lsYOa2KNvR1/07eiDzv4unOmXiEyGTBAE4d6HmQ+VSgUXFxfk5eVVGuxaXFyMK1euoHnz5rC1tZWoQqoP/NnWXFJ2EbafTsX2U2lISMo1eK69rzP6dvBB344+aOPtyMHARNRg7vb5fSfJW24WL16M+fPnIy0tDSEhIfjqq6+qvT25pKQEMTEx+P7773H9+nW0bdsWn3zyiUELAhHVjiAI+DejANtPibdsn0lVGTzfpZkr+nYQu5yCPBrH+DMismyShpv169cjOjoaS5cuRXh4OGJjYxEZGYnz58/Dy8ur0vHvvfceVq9eje+++w7t2rXDjh07MHjwYOzfvx9dunSR4B0QNU6CIOBESh62n07DjlNpuJxVqH9OIZchvLk7+nb0wZPBPvBxYUsXETUuknZLhYeHo3v37li0aBEA8S6egIAATJ48GdOmTat0vJ+fH2bMmIGJEyfq9w0ZMgR2dnZYvXp1jV6T3VKWiT9bQKsTcOSqOAfNH6fTcCOvWP+cjUKOh1p7ILKjDyLae3ORSSIyOY2iW0qj0SA+Ph7Tp0/X75PL5YiIiMCBAweqPEetVlf6YLKzs8O+ffuqfR21Wg21+vZEYiqVqtpjicyNulSL/ZeyseNUGnaeSUd24e25jextFOjT1gt9O/rg0baecLK1vsuViIgaD8nCTVZWFrRabaXbe729vXHu3Lkqz4mMjMSCBQvw8MMPo2XLloiLi8OmTZvuOndKTEwM5s6dW6vaLGyMtUWwpJ9pkaYUey9kYvupNMSdzUC+ulT/nIudNSLK5qB5qLUHF5kkIrMk+YDi2vjyyy8xfvx4tGvXDjKZDC1btsTYsWOxfPnyas+ZPn06oqOj9V+rVCoEBARUeWz5rLdFRUWws7MzbvEkqaIicV2jO2c2Nhd5t0qw+1w6tp9Kw54LmSguuT0HjaeTEpEdvNG3gy/CW7jDmnPQEJGZkyzceHh4QKFQID093WB/enp6tQsrenp6YsuWLSguLkZ2djb8/Pwwbdo0tGjRotrXUSqVlSaKq45CoYCrqysyMjIAiHOu8FbXxk0QBBQVFSEjIwOurq4G61E1dpn5auw8k47tp9Ow/2IWSivMQdPUzQ79Ooq3bHcJcOMcNERkUSQLNzY2NujWrRvi4uIwaNAgAOKA4ri4OEyaNOmu59ra2sLf3x8lJSX4+eefMXToUKPVVR6sygMOmQdXV9e7rkbeWKTcLMKO0+nYcSoNR67loGJvW2svR/Tr6IPIjj4I9nVmMCciiyVpt1R0dDTGjBmDsLAw9OjRA7GxsSgsLMTYsWMBAKNHj4a/vz9iYmIAAIcOHcL169cRGhqK69evY86cOdDpdHj77beNVpNMJoOvry+8vLxQUlJy7xPI5FlbWzfqFpsbubew+dh1bD+VhpPXDWeV7tzUBZFlc9C08nKUqEIiItMiabgZNmwYMjMzMWvWLKSlpSE0NBTbt2/XDzJOSkqCXH57fEBxcbF+KQFHR0f0798fq1atgqurq9FrUygUjfoDkczD7ydT8dbGEygoGxQskwHdg9zRt4MPnuzgjaZutVs5nojIEnD5BSITVKrV4ZPt5/Dd31cAACFNXTC8RzNEtPeGp1PNxpAREZmTRjHPDRFVLUNVjEnrjuHwlRwAwISHW+CtyLa8y4mIqIYYbohMyOErOZi4NgGZ+Wo4Kq0w/9nO6NfJV+qyiIgaFYYbIhMgCAKW7buCmN/PQasT0MbbEUue74aWnhwkTERUWww3RBLLLy7BOz+fwG8n0wAAA0P9EPNMJ9jb8J8nEdH94G9PIgldSM/HK6vjcTmzENYKGd4bEIzRPQM5Rw0RUR0w3BBJ5JfE65j280ncKtHCx9kWi0d1RbdAN6nLIiJq9BhuiBqYplSHeb+dxcr9VwEAvVs1wcLhXdDEkbd4ExEZA8MNUQNKzbuFiWsSkJCUCwCY2Kclop9oCwXXfiIiMhqGG6IG8s/FLLy+7hiyCzVwsrXCF0NDERHsLXVZRERmh+GGqJ7pdAKW7LmEz/84D50ABPs6Y8nzXRHYxEHq0oiIzBLDDVE9yrtVgv/+dBy7zqYDAJ7r1hQfDOoIW2uuW0ZEVF8YbojqyekbeXhtTQKuZRfBRiHH3IEdMLx7AG/zJiKqZww3RPVgY3wKZmw+CXWpDv6udljyfFd0buoqdVlERBaB4YbIiIpLtJj7vzNYdzgJAPBoW0/EDguFq72NxJUREVkOhhsiI0nOKcLEtQk4kZIHmQyY+ngbTH6sFeS8zZuIqEEx3BAZwV/nMzB1fSJyi0rgam+NL4d3wSNtPKUui4jIIjHcENWBTidg4e5/8WXcvxAEoHNTF3w9qiuautlLXRoRkcViuCG6TzcLNZi6PhF7LmQCAEaFN8OsqGAorXibNxGRlBhuiO7DiZRcvLo6Addzb0FpJce8wZ0wpFtTqcsiIiIw3BDViiAIWHc4GXO2noZGq0NgE3ssGdUNwX7OUpdGRERlGG6IauiWRov3tpzCzwkpAIAngr3x2XMhcLGzlrgyIiKqiOGGqAauZRfildUJOJuqglwGvBXZDv95uAVv8yYiMkEMN0T3sPNMOqJ/SkR+cSk8HG2wcEQX9GrpIXVZRERUDYYbomqUanVYsPMCvv7rEgCgazNXfD2qG3xcbCWujIiI7obhhqgKWQVqvL7uGPZfygYAjO0dhOn92sPGSi5xZUREdC8MN0R3iL92ExPXJCBNVQx7GwU+HtIZT4f4SV0WERHVEMMNURlBEPD9/qv4cNtZlOoEtPR0wNLnu6G1t5PUpRERUS0w3BABKNKUYtrPJ7H1+A0AwIBOvvjk2c5wVPKfCBFRY8Pf3GTxLmUW4NXV8biQXgAruQzT+7fHS72DIJPxNm8iosaI4YYs2m8nU/HWhuMo1Gjh5aTE4lFd0T3IXeqyiIioDhhuyCKVaHX4dPs5fPf3FQBAeHN3fDWyC7yceJs3EVFjx3BDFidDVYxJa4/h8NUcAMB/HmmBt55sCysFb/MmIjIHDDdkUQ5dzsakdceQma+Go9IKnz0Xgr4dfaQui4iIjIjhhiyCIAj4v7+v4OPt56DVCWjr7YQlz3dFC09HqUsjIiIjY7ghs5dfXIK3N57A76fSAACDu/jjo8EdYW/Dv/5EROaIv93JrF1Iz8crq+JxOasQ1goZZkV1wPPhzXibNxGRGWO4IbO17UQq3txwHLdKtPBzscXiUV3RpZmb1GUREVE9Y7ghs5SuKsbU9cdQohXwUGsPfDm8C9wdbKQui4iIGoDk974uXrwYQUFBsLW1RXh4OA4fPnzX42NjY9G2bVvY2dkhICAAb7zxBoqLixuoWmos1h5KQolWQNdmrlg5tgeDDRGRBZE03Kxfvx7R0dGYPXs2EhISEBISgsjISGRkZFR5/Nq1azFt2jTMnj0bZ8+exbJly7B+/Xq8++67DVw5mTJNqQ5rDycBAMb2bg6FnONriIgsiaThZsGCBRg/fjzGjh2L4OBgLF26FPb29li+fHmVx+/fvx+9e/fGyJEjERQUhCeffBIjRoy4Z2sPWZYdp9OQma+Gl5MSkR04hw0RkaWRLNxoNBrEx8cjIiLidjFyOSIiInDgwIEqz+nVqxfi4+P1Yeby5cv47bff0L9//2pfR61WQ6VSGWxk3n44cBUAMDK8GWysJO95JSKiBibZgOKsrCxotVp4e3sb7Pf29sa5c+eqPGfkyJHIysrCgw8+CEEQUFpaildeeeWu3VIxMTGYO3euUWsn03XmhgpHrt6ElVyGkT2aSV0OERFJoFH9t/avv/7CvHnz8PXXXyMhIQGbNm3Ctm3b8MEHH1R7zvTp05GXl6ffkpOTG7BiamirDl4FAPTt6AMvZy6CSURkiSRrufHw8IBCoUB6errB/vT0dPj4VD1OYubMmXjhhRfw8ssvAwA6deqEwsJCTJgwATNmzIBcXjmrKZVKKJVK478BMjl5RSXYfOw6AGBMryBpiyEiIslI1nJjY2ODbt26IS4uTr9Pp9MhLi4OPXv2rPKcoqKiSgFGoVAAENcOIsu2IT4ZxSU6tPNxQlggJ+sjIrJUkk7iFx0djTFjxiAsLAw9evRAbGwsCgsLMXbsWADA6NGj4e/vj5iYGABAVFQUFixYgC5duiA8PBwXL17EzJkzERUVpQ85ZJl0OgGrDl4DILbacHkFIiLLJWm4GTZsGDIzMzFr1iykpaUhNDQU27dv1w8yTkpKMmipee+99yCTyfDee+/h+vXr8PT0RFRUFD766COp3gKZiD3/ZuJadhGcbK0wMNRP6nKIiEhCMsHC+nNUKhVcXFyQl5cHZ2dnqcshI3lp5RHsPpeBcQ82x8yngqUuh4iIjKw2n9+N6m4poqpcyy7En+fFWa1feCBQ4mqIiEhqDDfU6K0+eA2CADzSxhNBHg5Sl0NERBJjuKFG7ZZGi5+OpgAAxvRiqw0RETHcUCO39fh15N0qQTN3ezzSxkvqcoiIyAQw3FCjJQgCvt8v3v79/APNuPo3EREBYLihRiwh6SbOpKqgtJJjaFiA1OUQEZGJYLihRqu81WZQqD9c7W0kroaIiEwFww01Shn5xfj9VCoA4IWeHEhMRES3MdxQo/Tj4WSUaAV0C3RDR38XqcshIiITwnBDjU6JVoc1h8QuqdFstSEiojsw3FCj88fpdKSr1PBwVKJfR1+pyyEiIhPDcEONzg8HrgIARvYIgI0V/woTEZEhfjJQo3IuTYVDV3KgkMswMpxdUkREVBnDDTUqPxwQx9pEdvCGj4utxNUQEZEpYrihRiPvVgk2J1wHAIzuGSRtMUREZLIYbqjR+Dk+BbdKtGjr7YTw5u5Sl0NERCaK4YYaBZ1OwKqDYpfUCz0DIZNxHSkiIqoaww01CvsuZuFKViGclFYY3MVf6nKIiMiEMdxQo1B++/ezYU3hoLSSthgiIjJpDDdk8pJzihB3LgMA8MIDvP2biIjujuGGTN7qQ9cgCMBDrT3QwtNR6nKIiMjEMdyQSSsu0WL9kWQAwBje/k1ERDXAcEMmbevxG8gtKoG/qx36tPOSuhwiImoEGG7IZAmCoB9I/ELPQCjkvP2biIjujeGGTNax5Fycuq6CjZUcQ8MCpC6HiIgaCYYbMlk/7L8KAHg6xA/uDjbSFkNERI0Gww2ZpMx8NX47mQaAA4mJiKh2GG7IJK0/kgSNVofQAFd0auoidTlERNSIMNyQySnV6rDmUBIAYEwvTtpHRES1w3BDJmfX2XSk5hWjiYMN+nfylbocIiJqZBhuyOR8v19c/Xt4jwAorRQSV0NERI0Nww2ZlAvp+ThwORtyGTAqnF1SRERUeww3ZFJWHRBbbZ4M9oGfq13dLpZ1Ebi8B7iVW/fCiIio0bCSugCicqriEvyckAIAGN2zjq022ZeAbx4GSgrFrz3aAP7dbm/eHQErzp1DRGSOGG7IZGyKT0GRRotWXo7o2bLJ/V9IpwN+mSgGG2sH8c+sC+J2fJ14jEIJ+HYG/MOApmGAf1fArTkg4xIPRESNHcMNmQRBEPDDQbFLakzPQMjqEjIOLQWSDgA2jsBrBwBre+B6AnD9KHA9Hkg5ChTnAilHxO1Q2Xl27mKrTtOw2y089u51fm9ERNSwGG7IJPxzMRuXMwvhqLTC4K5N7/9CWReBuPfFx09+CLg2Ex+3eVLcAEAQgJzLt4PO9Xgg7QRwKwe4uFPcyrm3KAs6ZYHHpxNgbXv/9RERUb1juCGT8H3Z6t9DuvrDUXmffy11WuCX14DSW0CLPkC3F6s+TiYDmrQUt85DxX2laiD9FJASf7uFJ/uiGIJyLgMnN4jHya0Bn463w07TMMC9JSDn2HwiIlNhEuFm8eLFmD9/PtLS0hASEoKvvvoKPXr0qPLYRx99FHv27Km0v3///ti2bVt9l0r1IOVmEeLOpgMAXqjLOlIHvwaSDwE2TsDTX9Vu/IyV8nZXFCaI+4pygBvHxKBT3spTlCXuu3EMOPKdeJytC+DXtUJ3Vhjg6Hn/74OIiOpE8nCzfv16REdHY+nSpQgPD0dsbCwiIyNx/vx5eHl5VTp+06ZN0Gg0+q+zs7MREhKC5557riHLJiNacygJOgHo3aoJWnk53t9FMi8Auz8UH/edB7gG1L0we3eg1ePiBojdWbnXyoJOWQtP6nGgOA+4/Ke4lXNpBjSt0J3lGwLY2Ne9JiIiuieZIAiClAWEh4eje/fuWLRoEQBAp9MhICAAkydPxrRp0+55fmxsLGbNmoXU1FQ4ODjc83iVSgUXFxfk5eXB2dm5zvVT3RSXaNHr493IKdTgmxe6IbKDT+0votMCyyPFwcGtIoBRGxvuridtCZB++nbrzvV4IPM8gDv+WckUgHewYXeWRxtAzhmYiYhqojaf35K23Gg0GsTHx2P69On6fXK5HBEREThw4ECNrrFs2TIMHz682mCjVquhVqv1X6tUqroVTUa17UQqcgo18He1w+PtKrfU1ciBRWKwUboAUQsb9nZuhTXgFypu3ceJ+4rzbndnlbfwFKQDaSfFLX6FeJyNk3hexe4sZ66lRURUV5KGm6ysLGi1Wnh7exvs9/b2xrlz5+55/uHDh3Hq1CksW7as2mNiYmIwd+7cOtdK9eOHsoHEI8ObwUpxH4NyM88Duz8SH/edB7j4G6+4+2XrArR4VNwAsTtLdb3C3VkJYvjR5ANX/xa3ck5+ht1Zfl0A5X121RERWSjJx9zUxbJly9CpU6dqBx8DwPTp0xEdHa3/WqVSISDACOMxqM4Sk3NxPCUPNgo5hne/j5+JthTY8iqgVQOtnwRCRxm/SGOQyQCXpuIWPFDcpy0FMs+VdWUdFVt4Ms8C+TeAszeAs/8rO1cOeLa7Pdi5aRjg2R5QNOp/ukRE9UrS35AeHh5QKBRIT0832J+eng4fn7uPvSgsLMSPP/6I999//67HKZVKKJXKOtdKxlfeavNUiC+aON7Hz2j/QjEcKF2AqC8b1+zCCivxlnKfjkC3MeI+dQGQmmjYwqNKATLOiNuxVeJx1vaAb2hZC09ZK49L08b1/omI6pGk4cbGxgbdunVDXFwcBg0aBEAcUBwXF4dJkybd9dwNGzZArVbj+eefb4BKydiyC9T49XgqAGD0/dz+nX4G+CtGfNzvE8DZz3jFSUXpCAQ9KG7l8tNuTzR4/Shwvaw7K2m/uJVz8DKcWdm/q9g9RkRkgSRv246OjsaYMWMQFhaGHj16IDY2FoWFhRg7diwAYPTo0fD390dMTIzBecuWLcOgQYPQpEkd1iAiyaw/mgyNVoeQpi4IDXCt3cnakrLuKA3Qph8QMrxeajQJTj5A+6fEDRDXzcq6UCHsxIt3axVmAOd/E7dyHm3Kxu6UzcHj1YGLhRKRRZA83AwbNgyZmZmYNWsW0tLSEBoaiu3bt+sHGSclJUF+x+yv58+fx759+/DHH39IUTLVUalWhzUHkwDc56R9/8SK3Te2rkBUrGV1x8jlgFc7cetSNsao5JY4307F5SRyr1VYLHSteJxCKc63U7GFxy3Isr5/RGQRJJ/npqFxnhvp7Tidhv+sioebvTUOTH8ctta1mOsl/TTwzSOArgR45rvbyyeQoYJM4EZCWdgpCzzFeZWPs29SoSurrJWHi4USkQlqNPPckGVadUBc/Xt4j2a1CzbaEmDzK2KwaTsA6MRZqavl6Am0iRQ3QLwdPfuSYXdW6gmgKBv49w9xK+fewnCyQZ9O4vIURESNRK3DTVBQEF566SW8+OKLaNasWX3URGbsYkY+9l3MglwGjAqv5d+ffV+Iq3fbuQFPfcHulNqQyQCPVuIWMkzcV6oWJxWs2J2Vc6nCYqE/icfJrcWAU3GyQfcWXCyUiExWrcPN1KlTsXLlSrz//vvo06cPxo0bh8GDB/N2a6qR8labx9t7o6lbLdZaSj0B7PlEfNz/M8DJ++7H071ZKcXA0jQMCP+PuK8op6w7q0ILT1G2uO9Gwu1zbV3u6M7qxsVCichk3PeYm4SEBKxcuRLr1q2DVqvFyJEj8dJLL6Fr167GrtGoOOZGOgXqUjwwLw4F6lKsHheOB1t71OzEUg3w3WNA+kmgfRQwdBVbbRqKIAA3rxqunZV6HCgtrnysa7M7urM6c7FQIjKa2nx+13lAcUlJCb7++mu88847KCkpQadOnfD6669j7NixkJngBxDDjXRWHbiKmb+cRgtPB8RFP1Lzvx9/xgB7Pgbs3IGJhwDH+1yDioxDWwKkn6qwdlY8kHW+8nEyBeDdwbA7y6MNu7OI6L40yIDikpISbN68GStWrMDOnTvxwAMPYNy4cUhJScG7776LXbt2Ye3atfd7eTIzgiDg+7IuqdEPBNY82KQeB/7+THw84HMGG1OgsBbXvPLrAnR/WdxXvlho+czK+sVCT4jb0eXicTZOgH8Xw+4sLhZKREZW63CTkJCAFStWYN26dZDL5Rg9ejS++OILtGvXTn/M4MGD0b17d6MWSo3bgcvZuJhRAAcbBYZ0a1qzk0o1wOZXAV0pEDwI6PhMvdZIdVDVYqF5KRXuzqqwWOiVveJWztnfcO0s31AuFkpEdVLrcNO9e3c88cQTWLJkCQYNGgRra+tKxzRv3hzDh5vxrLFUaz/sF1ttBnf1h5Nt5b8zVdr7KZBxGrD3EFttqPGQyQDXAHHrMEjcpy0VFwetuHZW5llxxXTVdeDs1rJz5eLioOUzK/uHiYuHcrFQIqqhWo+5uXbtGgIDA+urnnrHMTcN70buLTz4yW7oBOCPNx5GG2+nGpx0DPjucUDQAkN/uL2aNpmX8sVC9ZMNJohB507W9mI3mH9XMew0DRNbfExwXB8R1Y96HXOTkZGBtLQ0hIeHG+w/dOgQFAoFwsLCantJMnNrDyVBJwA9WzSpWbApVYvdUYIW6PAMg405q2qxUFWq4WSD5YuFXvtH3Mo5ehuuneXXhYuFEhGA+wg3EydOxNtvv10p3Fy/fh2ffPIJDh06ZLTiqPFTl2qx7rC4jtTonjVs8dvzidhd4eApzmlDlsXZF3CuuFio9vZioSkVFgstSAfObxM3AIBMvBuradjtFh7vDuIAaCKyKLUON2fOnKlyLpsuXbrgzJkzRimKzMdvJ1ORXaiBr4stngiuwcR71+PFmYgBcRZiB676bvHkCsCrvbh1eV7cpykS78KquHZWbpJ4S3rWeSBxjXicla24WGjFAcuugezOIjJztQ43SqUS6enpaNGihcH+1NRUWFlxwB8Z+qHs9u9R4c1gpbjH/CYlxWXdUTpx3aj2UQ1QITVKNvZAswfErVxBxu3b0MsnHCzOA5IPiVs5e48KYacb4MfFQonMTa3TyJNPPonp06fjl19+gYuL2L+dm5uLd999F0888YTRC6TG60RKLo4l5cJaIcOw7jVYR+qvGPF/3Q5eQL9P679AMi+OXkDbvuIGADqduFZWxe6stJNAURbw7w5xK+fe0nCyQZ+OXCyUqBGrdbj57LPP8PDDDyMwMBBdunQBACQmJsLb2xurVq0yeoHUeJW32gzo5AtPp3t8UCQfAfYvFB9HxfJ/0lR3cjng0VrcQsqmpigpFmdXrtidlXO5bMHQS8CJ9eJxChtxsdDysNO0bLFQdmcRNQr3tfxCYWEh1qxZg+PHj8POzg6dO3fGiBEjqpzzxtTwVvCGcbNQg/CYOGhKddj0Wi90beZW/cElt4BvHhYHjXYeDjzzTcMVSlSUY9idlXIUuJVT+ThbV8OxO/7dAIcaro9G9UMQANUN8XeHewtxfTMGUGmVL74rt7o9qaeR1PvyCw4ODpgwYcJ9FUeWYf3RZGhKdejo74wuAa53P/jPj8RfTo4+QL+PG6Q+Ij17d6B1hLgBZYuFXhEDT3l3VupxoDgXuBQnbuVcAw27s3w7A9Z2krwNi1CsEufAKp8TKeUoUJB2+3kHzwqLt5aNp7Jzlaxcs1eqBtJOiT+P8n8rOZfE54IeMnq4qY37HgF85swZJCUlQaPRGOx/+umn61wUNW5anYBV5etI9Qy6+zpSSYeA/YvEx1FfAnZ3aeEhaggymdgK4N4C6PSsuK9Uc3ux0PLWnex/gdxr4nbqZ/E4uZV4+3nF7qwmrblY6P3QlgAZZwzXK8s8D+COzgaZQmyxyUsGCjOBC7+LW7kmrSsE0G6Ad0fAyqZB34pZEAQg+5LhHFRpJwGtpvKx7i2AJq0avsYKat0tdfnyZQwePBgnT56ETCZD+enlH2Bardb4VRoRu6Xq364z6Xj5h6NwtbfGwemPw9ZaUfWBmiJg6YNi0g8ZCQxe0rCFEtXFrdzbrQgpZb/wCzMrH6d0Lptdudvt5SScajAtgiURBPFW/ootMqnHgdJblY91aVZhaY5u4q3+Ng5i93baydstCNePAjevVj5foRRb2Cq28Lg1Z3fWnQqzKgzGL/u5FOdWPs7O/fbfa/9u4s+mnsZM1mu31JQpU9C8eXPExcWhefPmOHz4MLKzs/Hf//4Xn33GCdcI+P7AVQDAsLCA6oMNAOz+UAw2Tr5A35iGKY7IWOxcgZZ9xA0oWyw02XDtrBvHALUKuLJH3Mo5NxU/VMtbePxCxQ9oS3HrZllrTPztrcpg6FK2inxZK5hf1+qDobUdENBD3MoVZlWeHuDWTSDliLiVs3OvPJ7Kkm5qKLklhsmKdxbmXqt8nEIphsmKLWFuQSYZDGvdcuPh4YHdu3ejc+fOcHFxweHDh9G2bVvs3r0b//3vf3Hs2LH6qtUo2HJTvy5nFuCxz/dAJgP2vtUHAe72VR947QCwoh8AARi1EWjNaQTIDGlLxa6ViqujZ5xF5a4VOeAVbPgB69lOnMCwsSvVAOkny1q3yr4P2RcrHye3MrxDzb+b2LVhzC49QRDvjjOYHuBE1V0rbs3vmB6gE2Bta7xapKLT3Z7xuzz0pZ8GdKWVj/VoY7jEiVcHSbv06rXlRqvVwslJXB/Iw8MDN27cQNu2bREYGIjz58/fX8VkNlYdFNP+4+28qg82miLgl9cACOKMsww2ZK4UVmIXiG9nIGysuE+dD9xIrDAIMwHIvyGO6Uk/BSR8Lx5n7SB2Z1Vs4XH2M8n/JetVDA/lAeJu4aFiV11DhAeZDGjSUtw6DxX3larLpge4I3zdvCJuJzeIx8mtxfmP9N1ZYeL8SKY+nio/zTDMlbcm3snBy7BFxr9ro16rrdbhpmPHjjh+/DiaN2+O8PBwfPrpp7CxscG3335badZisiyF6lJsPJoCAHihZ1D1B8a9L/4CdPYHIuc1THFEpkLpBDR/SNzKqW5U/gDSFADX9olbOUcfww8gvy6ArYQt0IXZFbqWjt7u9rmTnZthKPDrajpLq1gpb38/y93ZbZZyVJz88cYxcTvynXicrYv4Xiq28Dh6SvM+AEBdUNa9VCE4q1IqH2dtD/iGGgZnl6amHZxrqdbdUjt27EBhYSGeeeYZXLx4EU899RQuXLiAJk2aYP369Xjsscfqq1ajYLdU/Vl98Bre23IKzT0cEBf9COTyKv6hXP0HWNlffPz8z0CriIYtkqgxKF8stOJkg+lnAOHOGzZkgGfbO7oOgutnsdCSYrEVpuIg07sO2C2/Y8wMBuxWHPBc3sKTmgiUFlc+1qWZYWjwDRGXCzE2nVbs4jTo8jwjLl9TkUwOeLY3HITt2V5sVWxkavP5fV+T+N0pJycHbm5ud7/l10Qw3NQPQRDQN/ZvnE/Px8yngjHuweaVD9IUAkt6ib8Qu44Gnv6qweskarQ0hUDqiQqTDcYDeUmVj7OyqzDos2x19NpObqfTiV0z5R+cKUfFrpuqxmVY6q3WtblV3TvYsOXKo03txlMJAqC6Xnmweklh5WOd/W//3P27iYPVlU51eacmo97CTUlJCezs7JCYmIiOHTvWuVApMNzUj4OXszH824Ows1bg4LuPw8Wuiv85/vYWcPhb8U6R1w5I25xOZA4KMgy7s64nAOq8ysc5eFYYqNtVfFxxcruCDMNbqK8fu8t1OElete41yWA5GycxdFTsznL2rfo65S1F1V3Hv4vhIOyK1zEz9Tag2NraGs2aNTP5uWyo4f1Qdvv34K7+VQebK3vFYAMAAxcx2BAZg6MX0LafuAG3Fwut2J2Vdqpscrvt4lauSWtxsrWMs9W3APmFGt7B5RLQuLuX6putM9DiEXEDbi8PUbG17cYxQJMPXP1b3Mo5+YmDqnOv3aUFqIPh7eq1bQGyILXullq2bBk2bdqEVatWwd298c0DwJYb40vLK0bvT3ZDqxOwfepDaOdzx/dVXQAs6Sn2WXcbKy6MSUQNo6RYnNyu4tpZN6/ccZBMvPW84lgRr/b1M3bH0um0QOa5Ci1l8VWPlXFtZtiV5dO5fsbuNCL1eiv4okWLcPHiRfj5+SEwMBAODoYTTyUkJNT2ktTIrT10DVqdgB7N3SsHGwDYOUsMNi7NgCc/aPgCiSyZtS0Q0F3cyhXllK0DdAXwaifedWUm4zJMnrysBca7A9BtjLhPUyhOD5BxRmwd8+8qtsrRfat1uBk0aFA9lEGNlaZUh7WHkwEAY6q6/fvyX8DRZeLjgYv4C5TIFNi7c34pU2LjAAT1FjcyilqHm9mzZ9dHHdRI/X4qFVkFang7K/FkhzumRVfnA79MFh93f/l2PzQREVE9MvGpFcnU/VC2+vfIHoGwVtzx1+mPmeJARddAIGKuBNUREZElqnXLjVwuv+t8NryTynKcup6H+Gs3Ya2QYUR4gOGTF+OA+BXi44GLAaVjwxdIREQWqdbhZvPmzQZfl5SU4NixY/j+++8xdy7/d25JVpW12vTt6AsvpwprwhTnAVtfFx/3mGA4zTwREVE9q3W4GThwYKV9zz77LDp06ID169dj3LhxRimMTFtukQZbEq8DAMb0DDR88o/3xPVM3IKAiDkNXhsREVk2o425eeCBBxAXF2esy5GJ23A0BepSHYJ9ndEt0O32E//uAhJ+ACADBi0R7wIgIiJqQEYJN7du3cLChQvh7+9vjMuRidPqBKw6KHZJje4ZeHsM1q1cYGvZ3VHhrwCBvaQpkIiILFqtu6XuXCBTEATk5+fD3t4eq1evNmpxZJr2XMhAUk4RnG2tMDC0QqDdMQPIvyFO6f74LOkKJCIii1brcPPFF18YhBu5XA5PT0+Eh4fDzc3tLmdWbfHixZg/fz7S0tIQEhKCr776Cj169Kj2+NzcXMyYMQObNm1CTk4OAgMDERsbi/79+9f6ten+lN/+Pax7AOxsytY1ufAHkLgat7ujLHuacCIikk6tw82LL75otBdfv349oqOjsXTpUoSHhyM2NhaRkZE4f/48vLwqTz2t0WjwxBNPwMvLCxs3boS/vz+uXbsGV1dXo9VEd3c1qxB/nc+ETAY8/0DZQOJbN4H/ld0d1XMi0OwB6QokIiKLV+tws2LFCjg6OuK5554z2L9hwwYUFRVhzJgxNb7WggULMH78eIwdOxYAsHTpUmzbtg3Lly/HtGnTKh2/fPly5OTkYP/+/bC2Fhd0CwoKqu1boDooH2vzaBtPBDYpGyy8/V0gPxVo0gp47D0JqyMiIrqPAcUxMTHw8PCotN/Lywvz5s2r8XU0Gg3i4+MRERFxuxi5HBEREThw4ECV52zduhU9e/bExIkT4e3tjY4dO2LevHl3nThQrVZDpVIZbHR/ijSl2HBUXEdqdK8gcef534HjawGZXOyOsraTrkAiIiLcR7hJSkpC8+bNK+0PDAxEUlJSja+TlZUFrVYLb2/D9Yi8vb2RlpZW5TmXL1/Gxo0bodVq8dtvv2HmzJn4/PPP8eGHH1b7OjExMXBxcdFvAQEB1R5Ld/dL4g2oiksR2MQej7T2FFcW/t8U8cmeE4GA6sdKERERNZRahxsvLy+cOHGi0v7jx4+jSZMmRimqOjqdDl5eXvj222/RrVs3DBs2DDNmzMDSpUurPWf69OnIy8vTb8nJyfVao7kSBAHf778KAHjhgUDI5TJg+zSgIB3waAP0mSFtgURERGVqPeZmxIgReP311+Hk5ISHH34YALBnzx5MmTIFw4cPr/F1PDw8oFAokJ6ebrA/PT0dPj4+VZ7j6+sLa2trKBQK/b727dsjLS0NGo0GNjY2lc5RKpVQKpU1rouqdvTaTZxLy4ettRzPdQsAzm0DTqxndxQREZmcWrfcfPDBBwgPD8fjjz8OOzs72NnZ4cknn8Rjjz1WqzE3NjY26Natm8GsxjqdDnFxcejZs2eV5/Tu3RsXL16ETqfT77tw4QJ8fX2rDDZkPOWtNoNC/eGCfOB/U8Uner0ONA2TrC4iIqI71Trc2NjYYP369Th//jzWrFmDTZs24dKlS1i+fHmtA0Z0dDS+++47fP/99zh79ixeffVVFBYW6u+eGj16NKZPn64//tVXX0VOTg6mTJmCCxcuYNu2bZg3bx4mTpxY27dBtZChKsb2U+I4qBd6BgK/vQUUZgCe7YBHp9/jbCIiooZV626pcq1bt0br1q3r9OLDhg1DZmYmZs2ahbS0NISGhmL79u36QcZJSUmQy2/nr4CAAOzYsQNvvPEGOnfuDH9/f0yZMgXvvPNOneqgu1t7OAmlOgFhgW7okLsHOLURkCmAQV8D1rb3vgAREVEDkgmCINTmhCFDhqBHjx6VAsWnn36KI0eOYMOGDUYt0NhUKhVcXFyQl5cHZ2dnqcsxeZpSHXp/shuZ+WosfSYQff8aCBRlAQ/9l0ssEBFRg6nN53etu6X27t1b5VIH/fr1w969e2t7OTJxO06nITNfDU8nJZ64+pkYbLyCgUfYWkZERKap1uGmoKCgyrE11tbWnCDPDK0qW0dqVosLUJzZfLs7yop3oBERkWmqdbjp1KkT1q9fX2n/jz/+iODgYKMURabhzA0VDl/NgZdchQFJn4k7H4oG/LpIWxgREdFd1HpA8cyZM/HMM8/g0qVLeOyxxwAAcXFxWLt2LTZu3Gj0Akk6qw5eBSDgG/e1kBdkA94dgYfflrosIiKiu6p1uImKisKWLVswb948bNy4EXZ2dggJCcHu3bvh7u5eHzWSBPKKSrDl2A08JT+ILgV7AblVWXcU5xMiIiLTdl+3gg8YMAADBgwAII5eXrduHd58803Ex8ffdRFLajw2xCfDoSQHH9mtBAQAD70J+IZIXRYREdE91XrMTbm9e/dizJgx8PPzw+eff47HHnsMBw8eNGZtJBGdTsCqA1fxofVyuAj5gE8n8dZvIiKiRqBWLTdpaWlYuXIlli1bBpVKhaFDh0KtVmPLli0cTGxG9v6biZDcXehrcwSC3AqyQUvYHUVERI1GjVtuoqKi0LZtW5w4cQKxsbG4ceMGvvrqq/qsjSTyy98JeN96JQBA9sg7YssNERFRI1Hjlpvff/8dr7/+Ol599dU6L7tApislpxD9r30KV0Uh1J6doHzwDalLIiIiqpUat9zs27cP+fn56NatG8LDw7Fo0SJkZWXVZ20kgbN/LMMTiniUwgrKZ78BFNZSl0RERFQrNQ43DzzwAL777jukpqbiP//5D3788Uf4+flBp9Nh586dyM/Pr886qQEIqht44NzHAICzbV4FvDtIXBEREVHt1XrhzIrOnz+PZcuWYdWqVcjNzcUTTzyBrVu3GrM+o+PCmdUQBKhWDIFzUhxOCi0Q+M5+ONvbSV0VERERgHpeOLOitm3b4tNPP0VKSgrWrVtXl0uR1I6vg3NSHNSCFX4JfI/BhoiIGq06hZtyCoUCgwYNMvlWG6qG6gaE38VVvmNLn0X4Aw9KXBAREdH9M0q4oUZMEICtr0OmViFR1xIbbAbhkTaeUldFRER03xhuLN2x1cDFnSiR2eC/Ja+gX0gAbKz414KIiBovfopZsrwUYMe7AIAvtc/hkuCPQV38JS6KiIiobhhuLFVZdxTUKuS4heBrTT80c7dH12auUldGRERUJww3lirhB+BSHGBli0/tpkAHOQZ18YdMJpO6MiIiojphuLFEucnAjhkAgILe07Dhqnjb96BQPymrIiIiMgqGG0sjCMDWSYAmHwgIx0brKGh1AkICXNHC01Hq6oiIiOqM4cbSxK8ALv8FWNkCA7/G5sQ0AMBgttoQEZGZYLixJDevAX/MFB8/PhuXBB8cT8mDQi5DVAjDDRERmQeGG0uh05V1RxUAzXoB4a9gy7HrAIBH2niiiaNS4gKJiIiMg+HGUhxdBlzZC1jZAQMXQZDJsLks3HBuGyIiMicMN5Yg5wqwc7b4+Im5QJOWiL92Eyk3b8HBRoEn2ntLWx8REZERMdyYO50O+GUSUFIIBD4IdB8PAPpWm74dfWFno5CyQiIiIqNiuDF3R/4PuLYPsHYABi4C5HJoSnX49UQqAGAwu6SIiMjMMNyYs5zLwK4K3VHuzQEAf53PQN6tEng7K9GzZRMJCyQiIjI+hhtzpdMBWyYCJUVA0ENA2Dj9U1sSxS6pgaH+UMi53AIREZkXhhtzdfgbIGm/QXcUAOTdKsGusxkAgEGh7JIiIiLzw3BjjrIvAbvmio+f/ABwC9I/9fvJVGhKdWjr7YT2vk7S1EdERFSPGG7MjU4LbHkNKL0FtHgUCHvJ4OmKc9twBXAiIjJHDDfm5uASIPkgYOMEPP0VUCHAXM+9hUNXcgAAA7mWFBERmSmGG3OS9S+w+wPxceSHgGszg6d/KRtI/EALd/i52jV0dURERA2C4cZc6LujioGWjwFdxxg8LQgCNieI4YZz2xARkTljuDEXBxYDKYcBpXOl7igAOJOqwr8ZBbCxkqNvR1+JiiQiIqp/JhFuFi9ejKCgINja2iI8PByHDx+u9tiVK1dCJpMZbLa2tg1YrQnKPA/s/lB8HPkR4NK00iHlK4A/0d4bLnbWDVkdERFRg5I83Kxfvx7R0dGYPXs2EhISEBISgsjISGRkZFR7jrOzM1JTU/XbtWvXGrBiE6MtBba8CmjVQKsIoMsLlQ/RCfgl8QYArgBORETmT/Jws2DBAowfPx5jx45FcHAwli5dCnt7eyxfvrzac2QyGXx8fPSbt7cFr2p94CvgejygdAGiFlbqjgKA/ZeykJGvhqu9NR5p4ylBkURERA1H0nCj0WgQHx+PiIgI/T65XI6IiAgcOHCg2vMKCgoQGBiIgIAADBw4EKdPn672WLVaDZVKZbCZjYyzwJ/zxMd9YwCXqltlyue2eaqzL2ysJM+zRERE9UrST7qsrCxotdpKLS/e3t5IS0ur8py2bdti+fLl+OWXX7B69WrodDr06tULKSkpVR4fExMDFxcX/RYQEGD09yEJfXeUBmgdCYSOrPKwIk0pdpwSv5e8S4qIiCxBo/tvfM+ePTF69GiEhobikUcewaZNm+Dp6YlvvvmmyuOnT5+OvLw8/ZacnNzAFdeTf2KBG8cAWxcg6ssqu6MAYOeZdBRqtAhwt0PXZm4NWyMREZEErKR8cQ8PDygUCqSnpxvsT09Ph4+PT42uYW1tjS5duuDixYtVPq9UKqFUKutcq0lJPw389bH4uN+ngHP1t3aX3yU1OJTLLRARkWWQtOXGxsYG3bp1Q1xcnH6fTqdDXFwcevbsWaNraLVanDx5Er6+FjJ3i7ZE7I7SlQBt+gGdh1V7aFaBGnv/zQLAu6SIiMhySNpyAwDR0dEYM2YMwsLC0KNHD8TGxqKwsBBjx44FAIwePRr+/v6IiYkBALz//vt44IEH0KpVK+Tm5mL+/Pm4du0aXn75ZSnfRsPZFwukHgdsXYGo2Gq7owDg1+M3oNUJCAlwRQtPx4aqkIiISFKSh5thw4YhMzMTs2bNQlpaGkJDQ7F9+3b9IOOkpCTI5bcbmG7evInx48cjLS0Nbm5u6NatG/bv34/g4GCp3kLDSTsF7PlEfNx/PuB09667zWVz2wzmIplERGRBZIIgCFIX0ZBUKhVcXFyQl5cHZ2dnqcupOW0J8F0fIO0k0O4pYNjqu7baXMoswOOf74FCLsOhdx+Hh6OZjTsiIiKLUpvP70Z3t5TF+vtzMdjYuQNPfXHXYAMAv5QNJH64tQeDDRERWRSGm8Yg9QSwd774uP98wNHrrocLgoDNiWK44UBiIiKyNAw3pq5UU3Z3VCnQ/mmg45B7npKQdBPJObfgYKPAk8E1u6WeiIjIXDDcmLq984H0U4B9E2DAgnt2RwG3l1uI7OgDOxtFfVdIRERkUhhuTNmNRHGsDQAM+BxwvPeil5pSHX49kQoAeKZL03osjoiIyDQx3JiqUrXYHSVogeBBQIfBNTptz4VM5BaVwMtJiZ4tm9RvjURERCaI4cZU7fkUyDgD2HuIrTY1VL7cwsBQPyjkXG6BiIgsD8ONKbqeAOz7Qnz81ALAwaNGp+XdKsHOs+I6XbxLioiILBXDjamp2B3VcQgQPLDGp24/lQpNqQ5tvB0R7NuIJigkIiIyIoYbU/NXDJB5DnDwAvp/VqtTy++SGtSFK4ATEZHlYrgxJSlHgX++FB8/9QVg717jU6/n3sLByzkAgIGh7JIiIiLLxXBjKkqKy7qjdECnoUD7p2p1+tayRTLDm7vD39WuPiokIiJqFBhuTMWfHwFZFwBHb6DfJ7U6VRAEbD6WAgB4pitbbYiIyLIx3JiC5MPAgUXi46dia9UdBQBnU/NxIb0ANlZy9O3oa/z6iIiIGhGGG6mV3LrdHdV5ONCuf60vsaVskcyI9l5wsbM2doVERESNCsON1HZ/CGRfBBx9gH4f1/p0rU7AL+UrgHMgMREREcONpJIOAgcWi4+fXgjYudX6EgcuZSNdpYarvTUebetl5AKJiIgaH4YbqWiKgC2vARCA0FFAm8j7ukz53DYDOvnCxoo/TiIiIn4aSmX3B0DOJcDJD4icd1+XuKXRYvspcQXwwVxugYiICADDjTSu7QcOLhEfP70QsHO9r8vsPJuOQo0WAe526BZY+y4tIiIic8Rw09A0hbe7o7q8ALR+4r4vVb4C+OBQLrdARERUjuGmoe2aC9y8Ajj7A5Ef3fdlsgvU2HMhEwAwkF1SREREegw3DenqPuDwN+Ljp78CbF3u+1K/nkiFVicgpKkLWno6GqlAIiKixo/hpqGoC8q6owB0HQO0erxOl6u4AjgRERHdxnDTUHbNAXKvAS4BwJMf1ulSlzMLkJicC4Vchqc6+xmnPiIiIjPBcNMQLu8BjnwnPn76K8DWuU6X21K2AvhDrT3g6aSsa3VERERmheGmvqnzga2TxMdhLwEt+9TpcoIg3L5Lil1SRERElTDc1Leds4DcJMC1GfDE+3W+XEJSLpJyiuBgo8CTwT5GKJCIiMi8MNzUp0t/AkeXi48HLgaUTnW+ZHmrTWRHH9jZKOp8PSIiInPDcFNfilXA1sni4+7jgeYP1/mSmlIdfj0hjrdhlxQREVHVGG7qyx/vAXnJgGsgEDHHKJfceyETN4tK4OmkRK+WHka5JhERkblhuKkPF+OAhO/Fx4O+BpTGmWRvc6LYJTUwxA8KOZdbICIiqgrDjbEV593ujurxHyDoQaNcVlVcgp1n0gFw4j4iIqK7Ybgxth0zANV1wK05EDHbaJfdfjINmlIdWns5ooNf3ebJISIiMmcMN8b0707g2CoAMrE7ysbBaJeuuNwCVwAnIiKqHsONsdzKBba+Lj5+4FUgsJfRLn0j9xYOXskGwC4pIiKie2G4MZYLO4D8G4B7S+CxmUa99NbjNyAIQHhzd/i72hn12kRERObGSuoCzEbIMMDJG7B2AGzsjXppLrdARERUcybRcrN48WIEBQXB1tYW4eHhOHz4cI3O+/HHHyGTyTBo0KD6LbCmWjwKBHQ36iXPpqpwLi0fNgo5+nXyNeq1iYiIzJHk4Wb9+vWIjo7G7NmzkZCQgJCQEERGRiIjI+Ou5129ehVvvvkmHnrooQaqVBrlrTaPt/eCi521xNUQERGZPsnDzYIFCzB+/HiMHTsWwcHBWLp0Kezt7bF8+fJqz9FqtRg1ahTmzp2LFi1aNGC1DUurE7Al8fZdUkRERHRvkoYbjUaD+Ph4RERE6PfJ5XJERETgwIED1Z73/vvvw8vLC+PGjWuIMiVz8HI20lVquNhZ49G2nlKXQ0RE1ChIOqA4KysLWq0W3t7eBvu9vb1x7ty5Ks/Zt28fli1bhsTExBq9hlqthlqt1n+tUqnuu96GVj63zYDOvlBacQVwIiKimpC8W6o28vPz8cILL+C7776Dh0fNFo6MiYmBi4uLfgsICKjnKo3jlkaL7afSAADPsEuKiIioxiRtufHw8IBCoUB6errB/vT0dPj4+FQ6/tKlS7h69SqioqL0+3Q6HQDAysoK58+fR8uWLQ3OmT59OqKjo/Vfq1SqRhFwdp1NR4G6FE3d7NAt0E3qcoiIiBoNScONjY0NunXrhri4OP3t3DqdDnFxcZg0aVKl49u1a4eTJ08a7HvvvfeQn5+PL7/8ssrQolQqoVQq66X++lRxbhsut0BERFRzkk/iFx0djTFjxiAsLAw9evRAbGwsCgsLMXbsWADA6NGj4e/vj5iYGNja2qJjx44G57u6ugJApf2NWXaBGnsuZAIABoayS4qIiKg2JA83w4YNQ2ZmJmbNmoW0tDSEhoZi+/bt+kHGSUlJkMsb1dCgOtt2MhWlOgGdm7qglZej1OUQERE1KjJBEASpi2hIKpUKLi4uyMvLg7Ozs9TlVGnQ4n+QmJyLWU8F46UHm0tdDhERkeRq8/ltWU0ijcCVrEIkJudCIZchKsRP6nKIiIgaHYYbE1M+kPjBVh7wdGp8A6GJiIikxnBjQgTh9nILz3TlQGIiIqL7wXBjQo4l5+JadhHsbRR4Itj73icQERFRJQw3JqS8S6pvBx/Y20h+IxsREVGjxHBjIkq0Ovzv+A0AXAGciIioLhhuTMTeC5m4WVQCTyclerVsInU5REREjRbDjYkoXwH86RA/WCn4YyEiIrpf/BQ1AariEuw8Iy4eOphdUkRERHXCcGMCtp9Kg7pUh9ZejujgZ5qzJhMRETUWDDcmoPwuqUFcAZyIiKjOGG4klpp3CwcuZwMABoZyuQUiIqK6YriR2NbEGxAEoEdzdzR1s5e6HCIiokaP4UZi5XdJcSAxERGRcTDcSOhsqgrn0vJho5Cjf0dfqcshIiIyCww3EipfJPOxdl5wsbeWuBoiIiLzwHAjEa1OwC/HuNwCERGRsTHcSOTQ5WykqYrhYmeNPu08pS6HiIjIbDDcSKR8IPGAzr5QWikkroaIiMh8MNxIoLhEi99PpQHgXVJERETGxnAjgV1n01GgLkVTNzt0a+YmdTlERERmheFGAvrlFkL9IZdzuQUiIiJjYrhpYDmFGvx1PhMAMKgLl1sgIiIyNoabBrbtxA2U6gR08ndBKy8nqcshIiIyOww3DWxThRXAiYiIyPgYbhrQ1axCHEvKhUIuw9Mh7JIiIiKqDww3Dah8uYUHW3nA00kpcTVERETmieGmgQiCoL9LinPbEBER1R+GmwaSmJyLq9lFsLdR4MkO3lKXQ0REZLYYbhpIeatNZAcf2NtYSVwNERGR+WK4aQAlWh3+dyIVAO+SIiIiqm8MNw3g738zkVOogYejEr1bNpG6HCIiIrPGcNMANiWIXVIDQ/1gpeC3nIiIqD7xk7ae5ReXYOeZdAC8S4qIiKghMNzUs+2n0qAu1aGVlyM6+DlLXQ4REZHZY7ipZ+UT9w3u4g+ZjCuAExER1TeGm3qUlleM/ZeyAYDLLRARETUQhpt6tPX4dQgC0CPIHQHu9lKXQ0REZBFMItwsXrwYQUFBsLW1RXh4OA4fPlztsZs2bUJYWBhcXV3h4OCA0NBQrFq1qgGrrbnNx24A4Nw2REREDUnycLN+/XpER0dj9uzZSEhIQEhICCIjI5GRkVHl8e7u7pgxYwYOHDiAEydOYOzYsRg7dix27NjRwJXf3bk0Fc6mqmCjkGNAJ1+pyyEiIrIYkoebBQsWYPz48Rg7diyCg4OxdOlS2NvbY/ny5VUe/+ijj2Lw4MFo3749WrZsiSlTpqBz587Yt29fA1d+d5vLllt4rJ0XXOytJa6GiIjIckgabjQaDeLj4xEREaHfJ5fLERERgQMHDtzzfEEQEBcXh/Pnz+Phhx+uz1JrRacT8Au7pIiIiCQh6QqOWVlZ0Gq18PY2XCXb29sb586dq/a8vLw8+Pv7Q61WQ6FQ4Ouvv8YTTzxR5bFqtRpqtVr/tUqlMk7xd3HwSjbSVMVwtrVCn3ae9f56REREdFujXJ7ayckJiYmJKCgoQFxcHKKjo9GiRQs8+uijlY6NiYnB3LlzG7S+8hXAB3T2g9JK0aCvTUREZOkkDTceHh5QKBRIT0832J+eng4fH59qz5PL5WjVqhUAIDQ0FGfPnkVMTEyV4Wb69OmIjo7Wf61SqRAQEGCcN1CF4hItfj+ZBoDLLRAREUlB0jE3NjY26NatG+Li4vT7dDod4uLi0LNnzxpfR6fTGXQ9VaRUKuHs7Gyw1ae4sxnIV5fC39UOYYFu9fpaREREVJnk3VLR0dEYM2YMwsLC0KNHD8TGxqKwsBBjx44FAIwePRr+/v6IiYkBIHYzhYWFoWXLllCr1fjtt9+watUqLFmyRMq3oVd+l9SgLn6Qy7ncAhERUUOTPNwMGzYMmZmZmDVrFtLS0hAaGort27frBxknJSVBLr/dwFRYWIjXXnsNKSkpsLOzQ7t27bB69WoMGzZMqregl1OowV/nxfl52CVFREQkDZkgCILURTQklUoFFxcX5OXlGb2LatWBq5j5y2l08nfB/yY/aNRrExERWbLafH5LPomfObndJcVWGyIiIqkw3BjJtexCJCTlQi4DokK43AIREZFUJB9zYy6uZRfB00mJ9r7O8HKylbocIiIii8VwYyQPt/HEwemPI6dQI3UpREREFo3dUkakkMvg6aSUugwiIiKLxnBDREREZoXhhoiIiMwKww0RERGZFYYbIiIiMisMN0RERGRWGG6IiIjIrDDcEBERkVlhuCEiIiKzwnBDREREZoXhhoiIiMwKww0RERGZFYYbIiIiMisMN0RERGRWrKQuoKEJggAAUKlUEldCRERENVX+uV3+OX43Fhdu8vPzAQABAQESV0JERES1lZ+fDxcXl7seIxNqEoHMiE6nw40bN+Dk5ASZTGbUa6tUKgQEBCA5ORnOzs5GvTbVHn8epoU/D9PCn4fp4c/k7gRBQH5+Pvz8/CCX331UjcW13MjlcjRt2rReX8PZ2Zl/MU0Ifx6mhT8P08Kfh+nhz6R692qxKccBxURERGRWGG6IiIjIrDDcGJFSqcTs2bOhVCqlLoXAn4ep4c/DtPDnYXr4MzEeixtQTEREROaNLTdERERkVhhuiIiIyKww3BAREZFZYbghIiIis8JwYySLFy9GUFAQbG1tER4ejsOHD0tdksWKiYlB9+7d4eTkBC8vLwwaNAjnz5+Xuiwq8/HHH0Mmk2Hq1KlSl2Kxrl+/jueffx5NmjSBnZ0dOnXqhKNHj0pdlkXSarWYOXMmmjdvDjs7O7Rs2RIffPBBjdZPouox3BjB+vXrER0djdmzZyMhIQEhISGIjIxERkaG1KVZpD179mDixIk4ePAgdu7ciZKSEjz55JMoLCyUujSLd+TIEXzzzTfo3Lmz1KVYrJs3b6J3796wtrbG77//jjNnzuDzzz+Hm5ub1KVZpE8++QRLlizBokWLcPbsWXzyySf49NNP8dVXX0ldWqPGW8GNIDw8HN27d8eiRYsAiOtXBQQEYPLkyZg2bZrE1VFmZia8vLywZ88ePPzww1KXY7EKCgrQtWtXfP311/jwww8RGhqK2NhYqcuyONOmTcM///yDv//+W+pSCMBTTz0Fb29vLFu2TL9vyJAhsLOzw+rVqyWsrHFjy00daTQaxMfHIyIiQr9PLpcjIiICBw4ckLAyKpeXlwcAcHd3l7gSyzZx4kQMGDDA4N8KNbytW7ciLCwMzz33HLy8vNClSxd89913UpdlsXr16oW4uDhcuHABAHD8+HHs27cP/fr1k7iyxs3iFs40tqysLGi1Wnh7exvs9/b2xrlz5ySqisrpdDpMnToVvXv3RseOHaUux2L9+OOPSEhIwJEjR6QuxeJdvnwZS5YsQXR0NN59910cOXIEr7/+OmxsbDBmzBipy7M406ZNg0qlQrt27aBQKKDVavHRRx9h1KhRUpfWqDHckFmbOHEiTp06hX379kldisVKTk7GlClTsHPnTtja2kpdjsXT6XQICwvDvHnzAABdunTBqVOnsHTpUoYbCfz0009Ys2YN1q5diw4dOiAxMRFTp06Fn58ffx51wHBTRx4eHlAoFEhPTzfYn56eDh8fH4mqIgCYNGkSfv31V+zduxdNmzaVuhyLFR8fj4yMDHTt2lW/T6vVYu/evVi0aBHUajUUCoWEFVoWX19fBAcHG+xr3749fv75Z4kqsmxvvfUWpk2bhuHDhwMAOnXqhGvXriEmJobhpg445qaObGxs0K1bN8TFxen36XQ6xMXFoWfPnhJWZrkEQcCkSZOwefNm7N69G82bN5e6JIv2+OOP4+TJk0hMTNRvYWFhGDVqFBITExlsGljv3r0rTY1w4cIFBAYGSlSRZSsqKoJcbvhRrFAooNPpJKrIPLDlxgiio6MxZswYhIWFoUePHoiNjUVhYSHGjh0rdWkWaeLEiVi7di1++eUXODk5IS0tDQDg4uICOzs7iauzPE5OTpXGOzk4OKBJkyYcByWBN954A7169cK8efMwdOhQHD58GN9++y2+/fZbqUuzSFFRUfjoo4/QrFkzdOjQAceOHcOCBQvw0ksvSV1ao8ZbwY1k0aJFmD9/PtLS0hAaGoqFCxciPDxc6rIskkwmq3L/ihUr8OKLLzZsMVSlRx99lLeCS+jXX3/F9OnT8e+//6J58+aIjo7G+PHjpS7LIuXn52PmzJnYvHkzMjIy4OfnhxEjRmDWrFmwsbGRurxGi+GGiIiIzArH3BAREZFZYbghIiIis8JwQ0RERGaF4YaIiIjMCsMNERERmRWGGyIiIjIrDDdERERkVhhuiMjiyWQybNmyReoyiMhIGG6ISFIvvvgiZDJZpa1v375Sl0ZEjRTXliIiyfXt2xcrVqww2KdUKiWqhogaO7bcEJHklEolfHx8DDY3NzcAYpfRkiVL0K9fP9jZ2aFFixbYuHGjwfknT57EY489Bjs7OzRp0gQTJkxAQUGBwTHLly9Hhw4doFQq4evri0mTJhk8n5WVhcGDB8Pe3h6tW7fG1q1b6/dNE1G9YbghIpM3c+ZMDBkyBMePH8eoUaMwfPhwnD17FgBQWFiIyMhIuLm54ciRI9iwYQN27dplEF6WLFmCiRMnYsKECTh58iS2bt2KVq1aGbzG3LlzMXToUJw4cQL9+/fHqFGjkJOT06Dvk4iMRCAiktCYMWMEhUIhODg4GGwfffSRIAiCAEB45ZVXDM4JDw8XXn31VUEQBOHbb78V3NzchIKCAv3z27ZtE+RyuZCWliYIgiD4+fkJM2bMqLYGAMJ7772n/7qgoEAAIPz+++9Ge59E1HA45oaIJNenTx8sWbLEYJ+7u7v+cc+ePQ2e69mzJxITEwEAZ8+eRUhICBwcHPTP9+7dGzqdDufPn4dMJsONGzfw+OOP37WGzp076x87ODjA2dkZGRkZ9/uWiEhCDDdEJDkHB4dK3UTGYmdnV6PjrK2tDb6WyWTQ6XT1URIR1TOOuSEik3fw4MFKX7dv3x4A0L59exw/fhyFhYX65//55x/I5XK0bdsWTk5OCAoKQlxcXIPWTETSYcsNEUlOrVYjLS3NYJ+VlRU8PDwAABs2bEBYWBgefPBBrFmzBocPH8ayZcsAAKNGjcLs2bMxZswYzJkzB5mZmZg8eTJeeOEFeHt7AwDmzJmDV155BV5eXujXrx/y8/Pxzz//YPLkyQ37RomoQTDcEJHktm/fDl9fX4N9bdu2xblz5wCIdzL9+OOPeO211+Dr64t169YhODgYAGBvb48dO3ZgypQp6N69O+zt7TFkyBAsWLBAf60xY8aguLgYX3zxBd588014eHjg2Wefbbg3SEQNSiYIgiB1EURE1ZHJZNi8eTMGDRokdSlE1EhwzA0RERGZFYYbIiIiMiscc0NEJo0950RUW2y5ISIiIrPCcENERERmheGGiIiIzArDDREREZkVhhsiIiIyKww3REREZFYYboiIiMisMNwQERGRWWG4ISIiIrPy/1hbqQKsbCWoAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "evaluate.get_accuracy_plot(history)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 42,
      "id": "e25d1d63-46b1-4c11-a8ce-a6377bf6e595",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "e25d1d63-46b1-4c11-a8ce-a6377bf6e595",
        "outputId": "d8e1d427-e131-4372-eba7-a08e517fbe79"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "evaluate.get_loss_plot(history)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "id": "fea80130-1c91-4f23-b0b4-8598172f2b2f",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fea80130-1c91-4f23-b0b4-8598172f2b2f",
        "outputId": "70c14edb-e537-4f39-a0b6-17a159e8c4f1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "                  intent      predicted_intent\n",
            "39            i_get_back               how_are\n",
            "30          project_want               how_are\n",
            "26              thanking              thanking\n",
            "9           payment_link          offer_trials\n",
            "34            give_email                annual\n",
            "8               location          offer_trials\n",
            "25       demo_discussion      interview_resume\n",
            "26              thanking              thanking\n",
            "39            i_get_back   refund_process_time\n",
            "26              thanking              thanking\n",
            "25       demo_discussion                  cost\n",
            "10                Paypal                annual\n",
            "7                  MLOPS          offer_trials\n",
            "31           refund_info           refund_info\n",
            "26              thanking              thanking\n",
            "26              thanking              thanking\n",
            "2           download_can           emi_options\n",
            "30          project_want             placement\n",
            "24           ProjectDemo               how_are\n",
            "28      discount_student       demo_discussion\n",
            "26              thanking              thanking\n",
            "36                 noise  monthly_subscription\n",
            "29            enrollment             placement\n",
            "37             greetings            give_email\n",
            "31           refund_info   refund_process_time\n",
            "27              discount       demo_discussion\n",
            "30          project_want             placement\n",
            "18      monthly_payments              location\n",
            "18      monthly_payments              location\n",
            "19              12_month           ProjectDemo\n",
            "34            give_email            enrollment\n",
            "17  monthly_subscription      monthly_payments\n",
            "34            give_email            enrollment\n",
            "15                  cost         certification\n",
            "31           refund_info              6_months\n",
            "26              thanking              thanking\n",
            "17  monthly_subscription              6_months\n",
            "40          you_get_back            enrollment\n",
            "2           download_can      discount_student\n",
            "26              thanking              thanking\n",
            "13                 renew              6_months\n",
            "17  monthly_subscription      monthly_payments\n",
            "27              discount              12_month\n",
            "31           refund_info           refund_info\n",
            "39            i_get_back               how_are\n",
            "17  monthly_subscription              6_months\n"
          ]
        }
      ],
      "source": [
        "evaluate.compare_predicted_intents()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 44,
      "id": "c58cd045-6d6f-4956-ba97-062664643822",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "c58cd045-6d6f-4956-ba97-062664643822",
        "outputId": "aa361a1c-843e-47f7-de6c-100459a98b32"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.10/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              precision    recall  f1-score    support\n",
              "0              0.000000  0.000000  0.000000   1.000000\n",
              "2              0.000000  0.000000  0.000000   0.000000\n",
              "4              0.000000  0.000000  0.000000   1.000000\n",
              "5              0.000000  0.000000  0.000000   1.000000\n",
              "6              0.000000  0.000000  0.000000   1.000000\n",
              "7              0.000000  0.000000  0.000000   0.000000\n",
              "10             0.000000  0.000000  0.000000   0.000000\n",
              "11             0.000000  0.000000  0.000000   1.000000\n",
              "12             0.000000  0.000000  0.000000   2.000000\n",
              "13             0.000000  0.000000  0.000000   2.000000\n",
              "14             0.000000  0.000000  0.000000   1.000000\n",
              "15             0.000000  0.000000  0.000000   2.000000\n",
              "18             0.000000  0.000000  0.000000   0.000000\n",
              "19             0.000000  0.000000  0.000000   1.000000\n",
              "20             0.000000  0.000000  0.000000   3.000000\n",
              "21             0.000000  0.000000  0.000000   1.000000\n",
              "22             0.000000  0.000000  0.000000   0.000000\n",
              "23             0.000000  0.000000  0.000000   3.000000\n",
              "25             0.000000  0.000000  0.000000   0.000000\n",
              "27             0.000000  0.000000  0.000000   1.000000\n",
              "28             0.000000  0.000000  0.000000   2.000000\n",
              "29             0.000000  0.000000  0.000000   4.000000\n",
              "30             0.000000  0.000000  0.000000   1.000000\n",
              "31             0.000000  0.000000  0.000000   0.000000\n",
              "32             0.000000  0.000000  0.000000   1.000000\n",
              "33             0.000000  0.000000  0.000000   0.000000\n",
              "34             0.000000  0.000000  0.000000   3.000000\n",
              "36             1.000000  0.500000  0.666667   4.000000\n",
              "37             0.000000  0.000000  0.000000   0.000000\n",
              "38             0.000000  0.000000  0.000000   1.000000\n",
              "39             1.000000  1.000000  1.000000   8.000000\n",
              "40             0.000000  0.000000  0.000000   1.000000\n",
              "accuracy       0.217391  0.217391  0.217391   0.217391\n",
              "macro avg      0.062500  0.046875  0.052083  46.000000\n",
              "weighted avg   0.260870  0.217391  0.231884  46.000000"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-7e00bc74-1963-474a-9f76-847266969520\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>f1-score</th>\n",
              "      <th>support</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>3.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>3.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>4.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>3.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.500000</td>\n",
              "      <td>0.666667</td>\n",
              "      <td>4.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>8.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>accuracy</th>\n",
              "      <td>0.217391</td>\n",
              "      <td>0.217391</td>\n",
              "      <td>0.217391</td>\n",
              "      <td>0.217391</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>macro avg</th>\n",
              "      <td>0.062500</td>\n",
              "      <td>0.046875</td>\n",
              "      <td>0.052083</td>\n",
              "      <td>46.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weighted avg</th>\n",
              "      <td>0.260870</td>\n",
              "      <td>0.217391</td>\n",
              "      <td>0.231884</td>\n",
              "      <td>46.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7e00bc74-1963-474a-9f76-847266969520')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-7e00bc74-1963-474a-9f76-847266969520 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-7e00bc74-1963-474a-9f76-847266969520');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-5c379338-b150-4caa-8073-4f85a854578b\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-5c379338-b150-4caa-8073-4f85a854578b')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-5c379338-b150-4caa-8073-4f85a854578b button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ],
      "source": [
        "evaluate.get_fscore()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "id": "c70c430d-61aa-4073-bc38-bbcd5aece29b",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 455
        },
        "id": "c70c430d-61aa-4073-bc38-bbcd5aece29b",
        "outputId": "41a879c4-0b9d-48b6-a385-caa1ffd14284"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "evaluate.get_confusion_matrix()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Extraction"
      ],
      "metadata": {
        "id": "eATitZjfFmS_"
      },
      "id": "eATitZjfFmS_"
    },
    {
      "cell_type": "code",
      "source": [
        "# extraction of the necessary data\n",
        "with open('/content/drive/MyDrive/code/intents.json') as file:\n",
        "    data = json.load(file)\n",
        "\n",
        "training_sentences = []\n",
        "training_labels = []\n",
        "labels = []\n",
        "responses = []\n",
        "\n",
        "for intent in data['intents']:\n",
        "    for pattern in intent['patterns']:\n",
        "        training_sentences.append(pattern)\n",
        "        training_labels.append(intent['tag'])\n",
        "    responses.append(intent['responses'])\n",
        "\n",
        "    if intent['tag'] not in labels:\n",
        "        labels.append(intent['tag'])\n",
        "\n",
        "num_classes = len(labels)"
      ],
      "metadata": {
        "id": "_TkbpaXfFpnc"
      },
      "id": "_TkbpaXfFpnc",
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# transformation of tags into normalized encoding\n",
        "lbl_encoder = LabelEncoder()\n",
        "lbl_encoder.fit(training_labels)\n",
        "training_labels = lbl_encoder.transform(training_labels)"
      ],
      "metadata": {
        "id": "YknjA8h_Gbtg"
      },
      "id": "YknjA8h_Gbtg",
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# text data vectorization\n",
        "# when using this class, all punctuation will be removed, transforming the texts into a sequence\n",
        "# of words separated by spaces.\n",
        "vocab_size = 1000\n",
        "embedding_dim = 16\n",
        "max_len = 20\n",
        "oov_token = \"<OOV>\"\n",
        "\n",
        "tokenizer = Tokenizer(num_words=vocab_size, oov_token=oov_token)\n",
        "tokenizer.fit_on_texts(training_sentences)\n",
        "word_index = tokenizer.word_index\n",
        "sequences = tokenizer.texts_to_sequences(training_sentences)\n",
        "padded_sequences = pad_sequences(sequences, truncating='post', maxlen=max_len)"
      ],
      "metadata": {
        "id": "SZnXKF-THExG"
      },
      "id": "SZnXKF-THExG",
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# neural network architecture using the Keras “Sequential” model.\n",
        "model = Sequential()\n",
        "model.add(Embedding(vocab_size, embedding_dim, input_length=max_len))\n",
        "model.add(GlobalAveragePooling1D())\n",
        "model.add(Dense(16, activation='relu'))\n",
        "model.add(Dense(16, activation='relu'))\n",
        "model.add(Dense(num_classes, activation='softmax'))\n",
        "\n",
        "model.compile(loss='sparse_categorical_crossentropy',\n",
        "              optimizer='adam', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tiZFkro9HN_u",
        "outputId": "99a75bd4-975a-46d3-e56e-c97013dcad98"
      },
      "id": "tiZFkro9HN_u",
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 20, 16)            16000     \n",
            "                                                                 \n",
            " global_average_pooling1d (  (None, 16)                0         \n",
            " GlobalAveragePooling1D)                                         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 16)                272       \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 16)                272       \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 97)                1649      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 18193 (71.07 KB)\n",
            "Trainable params: 18193 (71.07 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# neural network training\n",
        "epochs = 800\n",
        "history = model.fit(padded_sequences, np.array(training_labels), epochs=epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ezyRitHwHUni",
        "outputId": "89f89f42-ecf1-4a9f-ba12-8374f312a08a"
      },
      "id": "ezyRitHwHUni",
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/800\n",
            "11/11 [==============================] - 1s 4ms/step - loss: 4.5726 - accuracy: 0.0093\n",
            "Epoch 2/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.5656 - accuracy: 0.0340\n",
            "Epoch 3/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.5578 - accuracy: 0.0617\n",
            "Epoch 4/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.5476 - accuracy: 0.0617\n",
            "Epoch 5/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.5335 - accuracy: 0.0617\n",
            "Epoch 6/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.5142 - accuracy: 0.0617\n",
            "Epoch 7/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.4864 - accuracy: 0.0617\n",
            "Epoch 8/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.4510 - accuracy: 0.0617\n",
            "Epoch 9/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.4069 - accuracy: 0.0617\n",
            "Epoch 10/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.3580 - accuracy: 0.0617\n",
            "Epoch 11/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.3103 - accuracy: 0.0617\n",
            "Epoch 12/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2746 - accuracy: 0.0617\n",
            "Epoch 13/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2517 - accuracy: 0.0617\n",
            "Epoch 14/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2402 - accuracy: 0.0617\n",
            "Epoch 15/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2291 - accuracy: 0.0617\n",
            "Epoch 16/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2200 - accuracy: 0.0617\n",
            "Epoch 17/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2108 - accuracy: 0.0617\n",
            "Epoch 18/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.2031 - accuracy: 0.0617\n",
            "Epoch 19/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1993 - accuracy: 0.0617\n",
            "Epoch 20/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.1929 - accuracy: 0.0617\n",
            "Epoch 21/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.1863 - accuracy: 0.0617\n",
            "Epoch 22/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1808 - accuracy: 0.0617\n",
            "Epoch 23/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1759 - accuracy: 0.0617\n",
            "Epoch 24/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1703 - accuracy: 0.0617\n",
            "Epoch 25/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1655 - accuracy: 0.0617\n",
            "Epoch 26/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1597 - accuracy: 0.0617\n",
            "Epoch 27/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1540 - accuracy: 0.0617\n",
            "Epoch 28/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1468 - accuracy: 0.0617\n",
            "Epoch 29/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1401 - accuracy: 0.0617\n",
            "Epoch 30/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1340 - accuracy: 0.0617\n",
            "Epoch 31/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1260 - accuracy: 0.0617\n",
            "Epoch 32/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1193 - accuracy: 0.0617\n",
            "Epoch 33/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.1106 - accuracy: 0.0617\n",
            "Epoch 34/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.1019 - accuracy: 0.0617\n",
            "Epoch 35/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0926 - accuracy: 0.0617\n",
            "Epoch 36/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0831 - accuracy: 0.0617\n",
            "Epoch 37/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.0731 - accuracy: 0.0617\n",
            "Epoch 38/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 4.0624 - accuracy: 0.0617\n",
            "Epoch 39/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0516 - accuracy: 0.0617\n",
            "Epoch 40/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0383 - accuracy: 0.0617\n",
            "Epoch 41/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0244 - accuracy: 0.0617\n",
            "Epoch 42/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 4.0128 - accuracy: 0.0617\n",
            "Epoch 43/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.9968 - accuracy: 0.0617\n",
            "Epoch 44/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.9797 - accuracy: 0.0617\n",
            "Epoch 45/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.9640 - accuracy: 0.0617\n",
            "Epoch 46/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.9464 - accuracy: 0.0617\n",
            "Epoch 47/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.9305 - accuracy: 0.0617\n",
            "Epoch 48/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.9122 - accuracy: 0.0648\n",
            "Epoch 49/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.8941 - accuracy: 0.0679\n",
            "Epoch 50/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.8755 - accuracy: 0.0710\n",
            "Epoch 51/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.8550 - accuracy: 0.0802\n",
            "Epoch 52/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.8323 - accuracy: 0.0772\n",
            "Epoch 53/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.8090 - accuracy: 0.0802\n",
            "Epoch 54/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 3.7859 - accuracy: 0.0864\n",
            "Epoch 55/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.7642 - accuracy: 0.0957\n",
            "Epoch 56/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.7371 - accuracy: 0.0926\n",
            "Epoch 57/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.7096 - accuracy: 0.0926\n",
            "Epoch 58/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.6804 - accuracy: 0.0957\n",
            "Epoch 59/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.6533 - accuracy: 0.0957\n",
            "Epoch 60/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.6230 - accuracy: 0.0957\n",
            "Epoch 61/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.5945 - accuracy: 0.1142\n",
            "Epoch 62/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.5652 - accuracy: 0.1265\n",
            "Epoch 63/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.5352 - accuracy: 0.1265\n",
            "Epoch 64/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.5058 - accuracy: 0.1173\n",
            "Epoch 65/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.4764 - accuracy: 0.1142\n",
            "Epoch 66/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.4450 - accuracy: 0.1204\n",
            "Epoch 67/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.4171 - accuracy: 0.1142\n",
            "Epoch 68/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.3913 - accuracy: 0.1265\n",
            "Epoch 69/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.3621 - accuracy: 0.1265\n",
            "Epoch 70/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.3370 - accuracy: 0.1235\n",
            "Epoch 71/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.3038 - accuracy: 0.1327\n",
            "Epoch 72/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.2754 - accuracy: 0.1265\n",
            "Epoch 73/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.2487 - accuracy: 0.1327\n",
            "Epoch 74/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.2175 - accuracy: 0.1451\n",
            "Epoch 75/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.1906 - accuracy: 0.1512\n",
            "Epoch 76/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.1607 - accuracy: 0.1543\n",
            "Epoch 77/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.1356 - accuracy: 0.1790\n",
            "Epoch 78/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.1090 - accuracy: 0.1759\n",
            "Epoch 79/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.0849 - accuracy: 0.2068\n",
            "Epoch 80/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 3.0554 - accuracy: 0.1852\n",
            "Epoch 81/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 3.0251 - accuracy: 0.1914\n",
            "Epoch 82/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.9989 - accuracy: 0.1914\n",
            "Epoch 83/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.9687 - accuracy: 0.2068\n",
            "Epoch 84/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.9408 - accuracy: 0.2160\n",
            "Epoch 85/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.9212 - accuracy: 0.2191\n",
            "Epoch 86/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.8911 - accuracy: 0.2191\n",
            "Epoch 87/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.8683 - accuracy: 0.2068\n",
            "Epoch 88/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.8355 - accuracy: 0.2315\n",
            "Epoch 89/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.8147 - accuracy: 0.2222\n",
            "Epoch 90/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.7878 - accuracy: 0.2253\n",
            "Epoch 91/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.7658 - accuracy: 0.2222\n",
            "Epoch 92/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.7416 - accuracy: 0.2315\n",
            "Epoch 93/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.7163 - accuracy: 0.2377\n",
            "Epoch 94/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.6911 - accuracy: 0.2407\n",
            "Epoch 95/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.6682 - accuracy: 0.2469\n",
            "Epoch 96/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.6527 - accuracy: 0.2685\n",
            "Epoch 97/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.6256 - accuracy: 0.2901\n",
            "Epoch 98/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.5996 - accuracy: 0.2901\n",
            "Epoch 99/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.5818 - accuracy: 0.2623\n",
            "Epoch 100/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.5601 - accuracy: 0.2623\n",
            "Epoch 101/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.5349 - accuracy: 0.2901\n",
            "Epoch 102/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.5111 - accuracy: 0.2963\n",
            "Epoch 103/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.4911 - accuracy: 0.3025\n",
            "Epoch 104/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.4700 - accuracy: 0.3025\n",
            "Epoch 105/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.4497 - accuracy: 0.2994\n",
            "Epoch 106/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.4290 - accuracy: 0.3025\n",
            "Epoch 107/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.4092 - accuracy: 0.2932\n",
            "Epoch 108/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.3886 - accuracy: 0.2932\n",
            "Epoch 109/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.3710 - accuracy: 0.3148\n",
            "Epoch 110/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.3537 - accuracy: 0.3117\n",
            "Epoch 111/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.3321 - accuracy: 0.3210\n",
            "Epoch 112/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.3133 - accuracy: 0.3148\n",
            "Epoch 113/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.2958 - accuracy: 0.3364\n",
            "Epoch 114/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.2721 - accuracy: 0.3364\n",
            "Epoch 115/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.2628 - accuracy: 0.3364\n",
            "Epoch 116/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.2373 - accuracy: 0.3580\n",
            "Epoch 117/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.2224 - accuracy: 0.3549\n",
            "Epoch 118/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.2072 - accuracy: 0.3395\n",
            "Epoch 119/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.1983 - accuracy: 0.3148\n",
            "Epoch 120/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.1685 - accuracy: 0.3704\n",
            "Epoch 121/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.1483 - accuracy: 0.3704\n",
            "Epoch 122/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.1422 - accuracy: 0.3827\n",
            "Epoch 123/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.1174 - accuracy: 0.4012\n",
            "Epoch 124/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.1026 - accuracy: 0.3951\n",
            "Epoch 125/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.0870 - accuracy: 0.4105\n",
            "Epoch 126/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.0739 - accuracy: 0.4012\n",
            "Epoch 127/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.0600 - accuracy: 0.3951\n",
            "Epoch 128/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 2.0411 - accuracy: 0.4136\n",
            "Epoch 129/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.0179 - accuracy: 0.4074\n",
            "Epoch 130/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 2.0039 - accuracy: 0.4012\n",
            "Epoch 131/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.9877 - accuracy: 0.4198\n",
            "Epoch 132/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.9728 - accuracy: 0.4167\n",
            "Epoch 133/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.9608 - accuracy: 0.4259\n",
            "Epoch 134/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.9447 - accuracy: 0.4444\n",
            "Epoch 135/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.9338 - accuracy: 0.4352\n",
            "Epoch 136/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.9112 - accuracy: 0.4290\n",
            "Epoch 137/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.9012 - accuracy: 0.4475\n",
            "Epoch 138/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.8955 - accuracy: 0.4599\n",
            "Epoch 139/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.8765 - accuracy: 0.4352\n",
            "Epoch 140/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.8584 - accuracy: 0.4568\n",
            "Epoch 141/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.8447 - accuracy: 0.4691\n",
            "Epoch 142/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.8323 - accuracy: 0.4969\n",
            "Epoch 143/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.8184 - accuracy: 0.4815\n",
            "Epoch 144/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.8030 - accuracy: 0.4969\n",
            "Epoch 145/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.7889 - accuracy: 0.5123\n",
            "Epoch 146/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.7706 - accuracy: 0.5278\n",
            "Epoch 147/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.7649 - accuracy: 0.5123\n",
            "Epoch 148/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.7534 - accuracy: 0.5093\n",
            "Epoch 149/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.7372 - accuracy: 0.5123\n",
            "Epoch 150/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.7241 - accuracy: 0.5185\n",
            "Epoch 151/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.7108 - accuracy: 0.5370\n",
            "Epoch 152/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.7009 - accuracy: 0.5278\n",
            "Epoch 153/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 1.6842 - accuracy: 0.5525\n",
            "Epoch 154/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.6676 - accuracy: 0.5309\n",
            "Epoch 155/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.6531 - accuracy: 0.5494\n",
            "Epoch 156/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.6445 - accuracy: 0.5525\n",
            "Epoch 157/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.6367 - accuracy: 0.5463\n",
            "Epoch 158/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.6276 - accuracy: 0.5247\n",
            "Epoch 159/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.6086 - accuracy: 0.5525\n",
            "Epoch 160/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5975 - accuracy: 0.5833\n",
            "Epoch 161/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5931 - accuracy: 0.5648\n",
            "Epoch 162/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5753 - accuracy: 0.5988\n",
            "Epoch 163/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5576 - accuracy: 0.6111\n",
            "Epoch 164/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5493 - accuracy: 0.5926\n",
            "Epoch 165/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5399 - accuracy: 0.5833\n",
            "Epoch 166/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5217 - accuracy: 0.5895\n",
            "Epoch 167/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5242 - accuracy: 0.6142\n",
            "Epoch 168/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.5087 - accuracy: 0.6049\n",
            "Epoch 169/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.4889 - accuracy: 0.5988\n",
            "Epoch 170/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4745 - accuracy: 0.6173\n",
            "Epoch 171/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4748 - accuracy: 0.6019\n",
            "Epoch 172/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4625 - accuracy: 0.5926\n",
            "Epoch 173/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4409 - accuracy: 0.6296\n",
            "Epoch 174/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4265 - accuracy: 0.6420\n",
            "Epoch 175/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4223 - accuracy: 0.6389\n",
            "Epoch 176/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.4095 - accuracy: 0.6296\n",
            "Epoch 177/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3964 - accuracy: 0.6296\n",
            "Epoch 178/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3901 - accuracy: 0.6327\n",
            "Epoch 179/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3764 - accuracy: 0.6358\n",
            "Epoch 180/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3701 - accuracy: 0.6173\n",
            "Epoch 181/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.3570 - accuracy: 0.6204\n",
            "Epoch 182/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.3421 - accuracy: 0.6605\n",
            "Epoch 183/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3289 - accuracy: 0.6605\n",
            "Epoch 184/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3237 - accuracy: 0.6636\n",
            "Epoch 185/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3137 - accuracy: 0.6512\n",
            "Epoch 186/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.3075 - accuracy: 0.6481\n",
            "Epoch 187/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2945 - accuracy: 0.6512\n",
            "Epoch 188/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2856 - accuracy: 0.6574\n",
            "Epoch 189/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.2710 - accuracy: 0.6636\n",
            "Epoch 190/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2635 - accuracy: 0.6759\n",
            "Epoch 191/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2542 - accuracy: 0.6728\n",
            "Epoch 192/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2457 - accuracy: 0.6914\n",
            "Epoch 193/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.2282 - accuracy: 0.6944\n",
            "Epoch 194/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.2277 - accuracy: 0.6667\n",
            "Epoch 195/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.2103 - accuracy: 0.6790\n",
            "Epoch 196/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2076 - accuracy: 0.6883\n",
            "Epoch 197/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.2000 - accuracy: 0.6975\n",
            "Epoch 198/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1853 - accuracy: 0.6975\n",
            "Epoch 199/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1771 - accuracy: 0.6914\n",
            "Epoch 200/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1705 - accuracy: 0.7006\n",
            "Epoch 201/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1571 - accuracy: 0.7099\n",
            "Epoch 202/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.1504 - accuracy: 0.7407\n",
            "Epoch 203/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1363 - accuracy: 0.7130\n",
            "Epoch 204/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1287 - accuracy: 0.6975\n",
            "Epoch 205/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.1146 - accuracy: 0.7222\n",
            "Epoch 206/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.1130 - accuracy: 0.7315\n",
            "Epoch 207/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.1056 - accuracy: 0.7346\n",
            "Epoch 208/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.0920 - accuracy: 0.7500\n",
            "Epoch 209/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.0803 - accuracy: 0.7438\n",
            "Epoch 210/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0754 - accuracy: 0.7191\n",
            "Epoch 211/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0672 - accuracy: 0.7377\n",
            "Epoch 212/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.0533 - accuracy: 0.7469\n",
            "Epoch 213/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0501 - accuracy: 0.7716\n",
            "Epoch 214/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0416 - accuracy: 0.7623\n",
            "Epoch 215/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0322 - accuracy: 0.7531\n",
            "Epoch 216/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0238 - accuracy: 0.7531\n",
            "Epoch 217/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 1.0167 - accuracy: 0.7500\n",
            "Epoch 218/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.0138 - accuracy: 0.7778\n",
            "Epoch 219/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 1.0023 - accuracy: 0.7654\n",
            "Epoch 220/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9934 - accuracy: 0.7963\n",
            "Epoch 221/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9864 - accuracy: 0.7840\n",
            "Epoch 222/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9759 - accuracy: 0.7716\n",
            "Epoch 223/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9633 - accuracy: 0.7747\n",
            "Epoch 224/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9601 - accuracy: 0.7809\n",
            "Epoch 225/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9538 - accuracy: 0.7716\n",
            "Epoch 226/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9439 - accuracy: 0.7963\n",
            "Epoch 227/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.9375 - accuracy: 0.7901\n",
            "Epoch 228/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.9256 - accuracy: 0.7994\n",
            "Epoch 229/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.9243 - accuracy: 0.7778\n",
            "Epoch 230/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.9121 - accuracy: 0.7994\n",
            "Epoch 231/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.9097 - accuracy: 0.7994\n",
            "Epoch 232/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.9023 - accuracy: 0.7901\n",
            "Epoch 233/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8885 - accuracy: 0.8056\n",
            "Epoch 234/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.8848 - accuracy: 0.7994\n",
            "Epoch 235/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8740 - accuracy: 0.8148\n",
            "Epoch 236/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8655 - accuracy: 0.8086\n",
            "Epoch 237/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8644 - accuracy: 0.8025\n",
            "Epoch 238/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8585 - accuracy: 0.8148\n",
            "Epoch 239/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8482 - accuracy: 0.7963\n",
            "Epoch 240/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8424 - accuracy: 0.8025\n",
            "Epoch 241/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8318 - accuracy: 0.8148\n",
            "Epoch 242/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.8226 - accuracy: 0.8179\n",
            "Epoch 243/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8257 - accuracy: 0.8210\n",
            "Epoch 244/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8243 - accuracy: 0.8025\n",
            "Epoch 245/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8054 - accuracy: 0.8117\n",
            "Epoch 246/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.8069 - accuracy: 0.7963\n",
            "Epoch 247/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7986 - accuracy: 0.8210\n",
            "Epoch 248/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.7974 - accuracy: 0.8241\n",
            "Epoch 249/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7838 - accuracy: 0.8272\n",
            "Epoch 250/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.7841 - accuracy: 0.8210\n",
            "Epoch 251/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7795 - accuracy: 0.8302\n",
            "Epoch 252/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.7687 - accuracy: 0.8395\n",
            "Epoch 253/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.7632 - accuracy: 0.8488\n",
            "Epoch 254/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7550 - accuracy: 0.8333\n",
            "Epoch 255/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.7574 - accuracy: 0.8086\n",
            "Epoch 256/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7380 - accuracy: 0.8333\n",
            "Epoch 257/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7450 - accuracy: 0.8210\n",
            "Epoch 258/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7379 - accuracy: 0.8302\n",
            "Epoch 259/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7233 - accuracy: 0.8395\n",
            "Epoch 260/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7172 - accuracy: 0.8333\n",
            "Epoch 261/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7091 - accuracy: 0.8426\n",
            "Epoch 262/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6992 - accuracy: 0.8519\n",
            "Epoch 263/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7015 - accuracy: 0.8519\n",
            "Epoch 264/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.7059 - accuracy: 0.8488\n",
            "Epoch 265/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6915 - accuracy: 0.8580\n",
            "Epoch 266/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6822 - accuracy: 0.8704\n",
            "Epoch 267/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6803 - accuracy: 0.8611\n",
            "Epoch 268/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.6707 - accuracy: 0.8642\n",
            "Epoch 269/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6705 - accuracy: 0.8519\n",
            "Epoch 270/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6637 - accuracy: 0.8488\n",
            "Epoch 271/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6533 - accuracy: 0.8642\n",
            "Epoch 272/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6440 - accuracy: 0.8704\n",
            "Epoch 273/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6442 - accuracy: 0.8765\n",
            "Epoch 274/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6460 - accuracy: 0.8704\n",
            "Epoch 275/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6397 - accuracy: 0.8735\n",
            "Epoch 276/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6295 - accuracy: 0.8827\n",
            "Epoch 277/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6271 - accuracy: 0.8765\n",
            "Epoch 278/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6219 - accuracy: 0.8920\n",
            "Epoch 279/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6170 - accuracy: 0.8735\n",
            "Epoch 280/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6089 - accuracy: 0.8827\n",
            "Epoch 281/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6085 - accuracy: 0.8642\n",
            "Epoch 282/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6038 - accuracy: 0.8889\n",
            "Epoch 283/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.6034 - accuracy: 0.8827\n",
            "Epoch 284/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5933 - accuracy: 0.8858\n",
            "Epoch 285/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5831 - accuracy: 0.9012\n",
            "Epoch 286/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5826 - accuracy: 0.9043\n",
            "Epoch 287/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.5891 - accuracy: 0.8920\n",
            "Epoch 288/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5877 - accuracy: 0.8704\n",
            "Epoch 289/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5896 - accuracy: 0.8920\n",
            "Epoch 290/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5714 - accuracy: 0.8858\n",
            "Epoch 291/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5567 - accuracy: 0.9043\n",
            "Epoch 292/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5551 - accuracy: 0.8889\n",
            "Epoch 293/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5484 - accuracy: 0.9012\n",
            "Epoch 294/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5451 - accuracy: 0.9105\n",
            "Epoch 295/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5491 - accuracy: 0.9105\n",
            "Epoch 296/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5426 - accuracy: 0.8920\n",
            "Epoch 297/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5425 - accuracy: 0.8889\n",
            "Epoch 298/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5474 - accuracy: 0.8889\n",
            "Epoch 299/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5248 - accuracy: 0.8981\n",
            "Epoch 300/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.5239 - accuracy: 0.9136\n",
            "Epoch 301/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5151 - accuracy: 0.9167\n",
            "Epoch 302/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5173 - accuracy: 0.9074\n",
            "Epoch 303/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5131 - accuracy: 0.8951\n",
            "Epoch 304/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5056 - accuracy: 0.9136\n",
            "Epoch 305/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.5026 - accuracy: 0.9198\n",
            "Epoch 306/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4952 - accuracy: 0.9136\n",
            "Epoch 307/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4981 - accuracy: 0.8981\n",
            "Epoch 308/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4852 - accuracy: 0.9228\n",
            "Epoch 309/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.4857 - accuracy: 0.9167\n",
            "Epoch 310/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.4814 - accuracy: 0.9136\n",
            "Epoch 311/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4735 - accuracy: 0.9383\n",
            "Epoch 312/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4720 - accuracy: 0.9290\n",
            "Epoch 313/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4766 - accuracy: 0.9136\n",
            "Epoch 314/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4692 - accuracy: 0.9198\n",
            "Epoch 315/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4646 - accuracy: 0.9136\n",
            "Epoch 316/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4587 - accuracy: 0.9259\n",
            "Epoch 317/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4604 - accuracy: 0.9167\n",
            "Epoch 318/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4482 - accuracy: 0.9228\n",
            "Epoch 319/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4470 - accuracy: 0.9259\n",
            "Epoch 320/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4415 - accuracy: 0.9228\n",
            "Epoch 321/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4402 - accuracy: 0.9198\n",
            "Epoch 322/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4361 - accuracy: 0.9321\n",
            "Epoch 323/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4345 - accuracy: 0.9259\n",
            "Epoch 324/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4345 - accuracy: 0.9228\n",
            "Epoch 325/800\n",
            "11/11 [==============================] - 0s 2ms/step - loss: 0.4260 - accuracy: 0.9167\n",
            "Epoch 326/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4263 - accuracy: 0.9383\n",
            "Epoch 327/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4278 - accuracy: 0.9198\n",
            "Epoch 328/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.4109 - accuracy: 0.9259\n",
            "Epoch 329/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.4177 - accuracy: 0.9198\n",
            "Epoch 330/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4260 - accuracy: 0.9228\n",
            "Epoch 331/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.4111 - accuracy: 0.9321\n",
            "Epoch 332/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.4026 - accuracy: 0.9352\n",
            "Epoch 333/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3989 - accuracy: 0.9383\n",
            "Epoch 334/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3932 - accuracy: 0.9414\n",
            "Epoch 335/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3903 - accuracy: 0.9414\n",
            "Epoch 336/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.4147 - accuracy: 0.9136\n",
            "Epoch 337/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3955 - accuracy: 0.9321\n",
            "Epoch 338/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3876 - accuracy: 0.9383\n",
            "Epoch 339/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3825 - accuracy: 0.9290\n",
            "Epoch 340/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3770 - accuracy: 0.9352\n",
            "Epoch 341/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3706 - accuracy: 0.9506\n",
            "Epoch 342/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3719 - accuracy: 0.9506\n",
            "Epoch 343/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3757 - accuracy: 0.9321\n",
            "Epoch 344/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3656 - accuracy: 0.9444\n",
            "Epoch 345/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3671 - accuracy: 0.9383\n",
            "Epoch 346/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3646 - accuracy: 0.9321\n",
            "Epoch 347/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3617 - accuracy: 0.9352\n",
            "Epoch 348/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3568 - accuracy: 0.9290\n",
            "Epoch 349/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3479 - accuracy: 0.9414\n",
            "Epoch 350/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3472 - accuracy: 0.9475\n",
            "Epoch 351/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3504 - accuracy: 0.9475\n",
            "Epoch 352/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3486 - accuracy: 0.9444\n",
            "Epoch 353/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3518 - accuracy: 0.9444\n",
            "Epoch 354/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3443 - accuracy: 0.9352\n",
            "Epoch 355/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3357 - accuracy: 0.9475\n",
            "Epoch 356/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3374 - accuracy: 0.9537\n",
            "Epoch 357/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3317 - accuracy: 0.9444\n",
            "Epoch 358/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3231 - accuracy: 0.9414\n",
            "Epoch 359/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3190 - accuracy: 0.9475\n",
            "Epoch 360/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3170 - accuracy: 0.9630\n",
            "Epoch 361/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3200 - accuracy: 0.9630\n",
            "Epoch 362/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3218 - accuracy: 0.9475\n",
            "Epoch 363/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3159 - accuracy: 0.9444\n",
            "Epoch 364/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3083 - accuracy: 0.9599\n",
            "Epoch 365/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3083 - accuracy: 0.9537\n",
            "Epoch 366/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.3040 - accuracy: 0.9537\n",
            "Epoch 367/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.3051 - accuracy: 0.9506\n",
            "Epoch 368/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.2968 - accuracy: 0.9630\n",
            "Epoch 369/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.2961 - accuracy: 0.9599\n",
            "Epoch 370/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2935 - accuracy: 0.9537\n",
            "Epoch 371/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2886 - accuracy: 0.9599\n",
            "Epoch 372/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2878 - accuracy: 0.9630\n",
            "Epoch 373/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.2888 - accuracy: 0.9537\n",
            "Epoch 374/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2867 - accuracy: 0.9568\n",
            "Epoch 375/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2812 - accuracy: 0.9568\n",
            "Epoch 376/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2811 - accuracy: 0.9568\n",
            "Epoch 377/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2795 - accuracy: 0.9630\n",
            "Epoch 378/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2749 - accuracy: 0.9599\n",
            "Epoch 379/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2743 - accuracy: 0.9660\n",
            "Epoch 380/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.2695 - accuracy: 0.9568\n",
            "Epoch 381/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2693 - accuracy: 0.9537\n",
            "Epoch 382/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2694 - accuracy: 0.9537\n",
            "Epoch 383/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.2720 - accuracy: 0.9506\n",
            "Epoch 384/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2698 - accuracy: 0.9599\n",
            "Epoch 385/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2671 - accuracy: 0.9660\n",
            "Epoch 386/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2605 - accuracy: 0.9630\n",
            "Epoch 387/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.2579 - accuracy: 0.9660\n",
            "Epoch 388/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.2592 - accuracy: 0.9568\n",
            "Epoch 389/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2608 - accuracy: 0.9568\n",
            "Epoch 390/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2567 - accuracy: 0.9568\n",
            "Epoch 391/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2516 - accuracy: 0.9660\n",
            "Epoch 392/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2575 - accuracy: 0.9537\n",
            "Epoch 393/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2470 - accuracy: 0.9660\n",
            "Epoch 394/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2449 - accuracy: 0.9660\n",
            "Epoch 395/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2456 - accuracy: 0.9568\n",
            "Epoch 396/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2465 - accuracy: 0.9537\n",
            "Epoch 397/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2456 - accuracy: 0.9537\n",
            "Epoch 398/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2411 - accuracy: 0.9568\n",
            "Epoch 399/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2437 - accuracy: 0.9660\n",
            "Epoch 400/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2434 - accuracy: 0.9691\n",
            "Epoch 401/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2365 - accuracy: 0.9630\n",
            "Epoch 402/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2384 - accuracy: 0.9691\n",
            "Epoch 403/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2437 - accuracy: 0.9722\n",
            "Epoch 404/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2283 - accuracy: 0.9691\n",
            "Epoch 405/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2238 - accuracy: 0.9691\n",
            "Epoch 406/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2241 - accuracy: 0.9630\n",
            "Epoch 407/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2240 - accuracy: 0.9568\n",
            "Epoch 408/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.2201 - accuracy: 0.9599\n",
            "Epoch 409/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.2205 - accuracy: 0.9660\n",
            "Epoch 410/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2169 - accuracy: 0.9691\n",
            "Epoch 411/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2149 - accuracy: 0.9691\n",
            "Epoch 412/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2226 - accuracy: 0.9568\n",
            "Epoch 413/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2186 - accuracy: 0.9599\n",
            "Epoch 414/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2106 - accuracy: 0.9660\n",
            "Epoch 415/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.2144 - accuracy: 0.9691\n",
            "Epoch 416/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2221 - accuracy: 0.9568\n",
            "Epoch 417/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.2134 - accuracy: 0.9568\n",
            "Epoch 418/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.2046 - accuracy: 0.9660\n",
            "Epoch 419/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.2102 - accuracy: 0.9722\n",
            "Epoch 420/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2033 - accuracy: 0.9660\n",
            "Epoch 421/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1999 - accuracy: 0.9722\n",
            "Epoch 422/800\n",
            "11/11 [==============================] - 0s 18ms/step - loss: 0.2013 - accuracy: 0.9660\n",
            "Epoch 423/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.2056 - accuracy: 0.9660\n",
            "Epoch 424/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1993 - accuracy: 0.9660\n",
            "Epoch 425/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1941 - accuracy: 0.9660\n",
            "Epoch 426/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1931 - accuracy: 0.9722\n",
            "Epoch 427/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1943 - accuracy: 0.9660\n",
            "Epoch 428/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1915 - accuracy: 0.9722\n",
            "Epoch 429/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1911 - accuracy: 0.9722\n",
            "Epoch 430/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1907 - accuracy: 0.9660\n",
            "Epoch 431/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1957 - accuracy: 0.9691\n",
            "Epoch 432/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1896 - accuracy: 0.9753\n",
            "Epoch 433/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1833 - accuracy: 0.9722\n",
            "Epoch 434/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.1833 - accuracy: 0.9630\n",
            "Epoch 435/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1853 - accuracy: 0.9660\n",
            "Epoch 436/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1831 - accuracy: 0.9722\n",
            "Epoch 437/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1976 - accuracy: 0.9568\n",
            "Epoch 438/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1835 - accuracy: 0.9630\n",
            "Epoch 439/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1801 - accuracy: 0.9691\n",
            "Epoch 440/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1778 - accuracy: 0.9660\n",
            "Epoch 441/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1792 - accuracy: 0.9691\n",
            "Epoch 442/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1729 - accuracy: 0.9722\n",
            "Epoch 443/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1710 - accuracy: 0.9660\n",
            "Epoch 444/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1748 - accuracy: 0.9722\n",
            "Epoch 445/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1727 - accuracy: 0.9691\n",
            "Epoch 446/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1716 - accuracy: 0.9630\n",
            "Epoch 447/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1738 - accuracy: 0.9691\n",
            "Epoch 448/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.1746 - accuracy: 0.9691\n",
            "Epoch 449/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1714 - accuracy: 0.9630\n",
            "Epoch 450/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1706 - accuracy: 0.9691\n",
            "Epoch 451/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1661 - accuracy: 0.9691\n",
            "Epoch 452/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1619 - accuracy: 0.9722\n",
            "Epoch 453/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.1576 - accuracy: 0.9753\n",
            "Epoch 454/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1715 - accuracy: 0.9722\n",
            "Epoch 455/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1684 - accuracy: 0.9753\n",
            "Epoch 456/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1636 - accuracy: 0.9660\n",
            "Epoch 457/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1613 - accuracy: 0.9691\n",
            "Epoch 458/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1580 - accuracy: 0.9784\n",
            "Epoch 459/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1558 - accuracy: 0.9722\n",
            "Epoch 460/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1545 - accuracy: 0.9722\n",
            "Epoch 461/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1600 - accuracy: 0.9691\n",
            "Epoch 462/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1554 - accuracy: 0.9691\n",
            "Epoch 463/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1515 - accuracy: 0.9753\n",
            "Epoch 464/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1484 - accuracy: 0.9753\n",
            "Epoch 465/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1479 - accuracy: 0.9660\n",
            "Epoch 466/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1504 - accuracy: 0.9753\n",
            "Epoch 467/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1475 - accuracy: 0.9753\n",
            "Epoch 468/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1490 - accuracy: 0.9722\n",
            "Epoch 469/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1463 - accuracy: 0.9722\n",
            "Epoch 470/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1422 - accuracy: 0.9691\n",
            "Epoch 471/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1496 - accuracy: 0.9722\n",
            "Epoch 472/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1489 - accuracy: 0.9784\n",
            "Epoch 473/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1400 - accuracy: 0.9753\n",
            "Epoch 474/800\n",
            "11/11 [==============================] - 0s 13ms/step - loss: 0.1395 - accuracy: 0.9691\n",
            "Epoch 475/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1409 - accuracy: 0.9722\n",
            "Epoch 476/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1395 - accuracy: 0.9722\n",
            "Epoch 477/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1401 - accuracy: 0.9660\n",
            "Epoch 478/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1363 - accuracy: 0.9753\n",
            "Epoch 479/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1443 - accuracy: 0.9722\n",
            "Epoch 480/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1414 - accuracy: 0.9753\n",
            "Epoch 481/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1409 - accuracy: 0.9784\n",
            "Epoch 482/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1385 - accuracy: 0.9722\n",
            "Epoch 483/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1341 - accuracy: 0.9753\n",
            "Epoch 484/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1310 - accuracy: 0.9753\n",
            "Epoch 485/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1326 - accuracy: 0.9722\n",
            "Epoch 486/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.1285 - accuracy: 0.9722\n",
            "Epoch 487/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1288 - accuracy: 0.9691\n",
            "Epoch 488/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1281 - accuracy: 0.9753\n",
            "Epoch 489/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1271 - accuracy: 0.9753\n",
            "Epoch 490/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.1303 - accuracy: 0.9722\n",
            "Epoch 491/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1269 - accuracy: 0.9722\n",
            "Epoch 492/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1319 - accuracy: 0.9753\n",
            "Epoch 493/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1255 - accuracy: 0.9722\n",
            "Epoch 494/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.1258 - accuracy: 0.9691\n",
            "Epoch 495/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1249 - accuracy: 0.9722\n",
            "Epoch 496/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1219 - accuracy: 0.9753\n",
            "Epoch 497/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1222 - accuracy: 0.9691\n",
            "Epoch 498/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1210 - accuracy: 0.9753\n",
            "Epoch 499/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1217 - accuracy: 0.9722\n",
            "Epoch 500/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1221 - accuracy: 0.9722\n",
            "Epoch 501/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1189 - accuracy: 0.9722\n",
            "Epoch 502/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1216 - accuracy: 0.9660\n",
            "Epoch 503/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1398 - accuracy: 0.9722\n",
            "Epoch 504/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1297 - accuracy: 0.9691\n",
            "Epoch 505/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1224 - accuracy: 0.9753\n",
            "Epoch 506/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1207 - accuracy: 0.9753\n",
            "Epoch 507/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1155 - accuracy: 0.9722\n",
            "Epoch 508/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1150 - accuracy: 0.9753\n",
            "Epoch 509/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1171 - accuracy: 0.9722\n",
            "Epoch 510/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1185 - accuracy: 0.9691\n",
            "Epoch 511/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1178 - accuracy: 0.9660\n",
            "Epoch 512/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1131 - accuracy: 0.9722\n",
            "Epoch 513/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1109 - accuracy: 0.9753\n",
            "Epoch 514/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1112 - accuracy: 0.9784\n",
            "Epoch 515/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1150 - accuracy: 0.9691\n",
            "Epoch 516/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1093 - accuracy: 0.9815\n",
            "Epoch 517/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1096 - accuracy: 0.9722\n",
            "Epoch 518/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1068 - accuracy: 0.9722\n",
            "Epoch 519/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1058 - accuracy: 0.9753\n",
            "Epoch 520/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1145 - accuracy: 0.9691\n",
            "Epoch 521/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1123 - accuracy: 0.9722\n",
            "Epoch 522/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1073 - accuracy: 0.9753\n",
            "Epoch 523/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1101 - accuracy: 0.9753\n",
            "Epoch 524/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1076 - accuracy: 0.9722\n",
            "Epoch 525/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1062 - accuracy: 0.9753\n",
            "Epoch 526/800\n",
            "11/11 [==============================] - 0s 13ms/step - loss: 0.1093 - accuracy: 0.9722\n",
            "Epoch 527/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1027 - accuracy: 0.9722\n",
            "Epoch 528/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1030 - accuracy: 0.9753\n",
            "Epoch 529/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1040 - accuracy: 0.9722\n",
            "Epoch 530/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.1044 - accuracy: 0.9691\n",
            "Epoch 531/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.1059 - accuracy: 0.9722\n",
            "Epoch 532/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.1041 - accuracy: 0.9722\n",
            "Epoch 533/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0996 - accuracy: 0.9753\n",
            "Epoch 534/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.1084 - accuracy: 0.9753\n",
            "Epoch 535/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1024 - accuracy: 0.9753\n",
            "Epoch 536/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0962 - accuracy: 0.9784\n",
            "Epoch 537/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0962 - accuracy: 0.9722\n",
            "Epoch 538/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0973 - accuracy: 0.9753\n",
            "Epoch 539/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0951 - accuracy: 0.9691\n",
            "Epoch 540/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0999 - accuracy: 0.9753\n",
            "Epoch 541/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0971 - accuracy: 0.9722\n",
            "Epoch 542/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0944 - accuracy: 0.9753\n",
            "Epoch 543/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0937 - accuracy: 0.9722\n",
            "Epoch 544/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0924 - accuracy: 0.9784\n",
            "Epoch 545/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0920 - accuracy: 0.9753\n",
            "Epoch 546/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0910 - accuracy: 0.9753\n",
            "Epoch 547/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0908 - accuracy: 0.9753\n",
            "Epoch 548/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0910 - accuracy: 0.9691\n",
            "Epoch 549/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0956 - accuracy: 0.9660\n",
            "Epoch 550/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0981 - accuracy: 0.9722\n",
            "Epoch 551/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0931 - accuracy: 0.9753\n",
            "Epoch 552/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0933 - accuracy: 0.9784\n",
            "Epoch 553/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1131 - accuracy: 0.9722\n",
            "Epoch 554/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1077 - accuracy: 0.9784\n",
            "Epoch 555/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.1070 - accuracy: 0.9784\n",
            "Epoch 556/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0970 - accuracy: 0.9784\n",
            "Epoch 557/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0908 - accuracy: 0.9784\n",
            "Epoch 558/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0879 - accuracy: 0.9815\n",
            "Epoch 559/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0903 - accuracy: 0.9784\n",
            "Epoch 560/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0915 - accuracy: 0.9722\n",
            "Epoch 561/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0884 - accuracy: 0.9753\n",
            "Epoch 562/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0845 - accuracy: 0.9784\n",
            "Epoch 563/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0830 - accuracy: 0.9753\n",
            "Epoch 564/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0841 - accuracy: 0.9784\n",
            "Epoch 565/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0875 - accuracy: 0.9691\n",
            "Epoch 566/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0838 - accuracy: 0.9784\n",
            "Epoch 567/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0857 - accuracy: 0.9784\n",
            "Epoch 568/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0865 - accuracy: 0.9753\n",
            "Epoch 569/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0809 - accuracy: 0.9784\n",
            "Epoch 570/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0806 - accuracy: 0.9784\n",
            "Epoch 571/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0813 - accuracy: 0.9784\n",
            "Epoch 572/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0795 - accuracy: 0.9753\n",
            "Epoch 573/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0807 - accuracy: 0.9722\n",
            "Epoch 574/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0885 - accuracy: 0.9753\n",
            "Epoch 575/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0836 - accuracy: 0.9784\n",
            "Epoch 576/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0803 - accuracy: 0.9815\n",
            "Epoch 577/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0816 - accuracy: 0.9753\n",
            "Epoch 578/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0835 - accuracy: 0.9753\n",
            "Epoch 579/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0797 - accuracy: 0.9753\n",
            "Epoch 580/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0794 - accuracy: 0.9753\n",
            "Epoch 581/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0760 - accuracy: 0.9784\n",
            "Epoch 582/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0784 - accuracy: 0.9753\n",
            "Epoch 583/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0910 - accuracy: 0.9691\n",
            "Epoch 584/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.1008 - accuracy: 0.9722\n",
            "Epoch 585/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0901 - accuracy: 0.9722\n",
            "Epoch 586/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0789 - accuracy: 0.9784\n",
            "Epoch 587/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0765 - accuracy: 0.9784\n",
            "Epoch 588/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0757 - accuracy: 0.9784\n",
            "Epoch 589/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0747 - accuracy: 0.9784\n",
            "Epoch 590/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0774 - accuracy: 0.9753\n",
            "Epoch 591/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0799 - accuracy: 0.9753\n",
            "Epoch 592/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0744 - accuracy: 0.9815\n",
            "Epoch 593/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0739 - accuracy: 0.9784\n",
            "Epoch 594/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0775 - accuracy: 0.9784\n",
            "Epoch 595/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0764 - accuracy: 0.9722\n",
            "Epoch 596/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0755 - accuracy: 0.9722\n",
            "Epoch 597/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0773 - accuracy: 0.9784\n",
            "Epoch 598/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0840 - accuracy: 0.9784\n",
            "Epoch 599/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0749 - accuracy: 0.9722\n",
            "Epoch 600/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0717 - accuracy: 0.9753\n",
            "Epoch 601/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0731 - accuracy: 0.9784\n",
            "Epoch 602/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0738 - accuracy: 0.9753\n",
            "Epoch 603/800\n",
            "11/11 [==============================] - 0s 13ms/step - loss: 0.0703 - accuracy: 0.9753\n",
            "Epoch 604/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0812 - accuracy: 0.9784\n",
            "Epoch 605/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0725 - accuracy: 0.9784\n",
            "Epoch 606/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0714 - accuracy: 0.9691\n",
            "Epoch 607/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0697 - accuracy: 0.9784\n",
            "Epoch 608/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0695 - accuracy: 0.9784\n",
            "Epoch 609/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0695 - accuracy: 0.9722\n",
            "Epoch 610/800\n",
            "11/11 [==============================] - 0s 13ms/step - loss: 0.0681 - accuracy: 0.9753\n",
            "Epoch 611/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0751 - accuracy: 0.9784\n",
            "Epoch 612/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0720 - accuracy: 0.9753\n",
            "Epoch 613/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0695 - accuracy: 0.9784\n",
            "Epoch 614/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0693 - accuracy: 0.9753\n",
            "Epoch 615/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0689 - accuracy: 0.9722\n",
            "Epoch 616/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0740 - accuracy: 0.9815\n",
            "Epoch 617/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0797 - accuracy: 0.9722\n",
            "Epoch 618/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0703 - accuracy: 0.9815\n",
            "Epoch 619/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0673 - accuracy: 0.9784\n",
            "Epoch 620/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0704 - accuracy: 0.9753\n",
            "Epoch 621/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0658 - accuracy: 0.9753\n",
            "Epoch 622/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0672 - accuracy: 0.9784\n",
            "Epoch 623/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0649 - accuracy: 0.9784\n",
            "Epoch 624/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0657 - accuracy: 0.9753\n",
            "Epoch 625/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0674 - accuracy: 0.9691\n",
            "Epoch 626/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0692 - accuracy: 0.9784\n",
            "Epoch 627/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0653 - accuracy: 0.9753\n",
            "Epoch 628/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0637 - accuracy: 0.9753\n",
            "Epoch 629/800\n",
            "11/11 [==============================] - 0s 16ms/step - loss: 0.0639 - accuracy: 0.9784\n",
            "Epoch 630/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0625 - accuracy: 0.9815\n",
            "Epoch 631/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0825 - accuracy: 0.9753\n",
            "Epoch 632/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0736 - accuracy: 0.9753\n",
            "Epoch 633/800\n",
            "11/11 [==============================] - 0s 13ms/step - loss: 0.0678 - accuracy: 0.9784\n",
            "Epoch 634/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0624 - accuracy: 0.9784\n",
            "Epoch 635/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0638 - accuracy: 0.9784\n",
            "Epoch 636/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0629 - accuracy: 0.9753\n",
            "Epoch 637/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0659 - accuracy: 0.9784\n",
            "Epoch 638/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0647 - accuracy: 0.9753\n",
            "Epoch 639/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0614 - accuracy: 0.9753\n",
            "Epoch 640/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0626 - accuracy: 0.9753\n",
            "Epoch 641/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0611 - accuracy: 0.9753\n",
            "Epoch 642/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.1049 - accuracy: 0.9599\n",
            "Epoch 643/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0839 - accuracy: 0.9753\n",
            "Epoch 644/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0650 - accuracy: 0.9691\n",
            "Epoch 645/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0649 - accuracy: 0.9784\n",
            "Epoch 646/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0644 - accuracy: 0.9784\n",
            "Epoch 647/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0615 - accuracy: 0.9753\n",
            "Epoch 648/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0609 - accuracy: 0.9753\n",
            "Epoch 649/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0700 - accuracy: 0.9784\n",
            "Epoch 650/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0682 - accuracy: 0.9753\n",
            "Epoch 651/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0628 - accuracy: 0.9722\n",
            "Epoch 652/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0613 - accuracy: 0.9753\n",
            "Epoch 653/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0633 - accuracy: 0.9722\n",
            "Epoch 654/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0613 - accuracy: 0.9722\n",
            "Epoch 655/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0603 - accuracy: 0.9753\n",
            "Epoch 656/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0601 - accuracy: 0.9784\n",
            "Epoch 657/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0592 - accuracy: 0.9753\n",
            "Epoch 658/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0679 - accuracy: 0.9753\n",
            "Epoch 659/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0677 - accuracy: 0.9753\n",
            "Epoch 660/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0605 - accuracy: 0.9784\n",
            "Epoch 661/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0583 - accuracy: 0.9784\n",
            "Epoch 662/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0588 - accuracy: 0.9660\n",
            "Epoch 663/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0911 - accuracy: 0.9691\n",
            "Epoch 664/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0841 - accuracy: 0.9660\n",
            "Epoch 665/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0606 - accuracy: 0.9784\n",
            "Epoch 666/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0593 - accuracy: 0.9691\n",
            "Epoch 667/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0559 - accuracy: 0.9753\n",
            "Epoch 668/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0615 - accuracy: 0.9784\n",
            "Epoch 669/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0857 - accuracy: 0.9722\n",
            "Epoch 670/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0689 - accuracy: 0.9753\n",
            "Epoch 671/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0662 - accuracy: 0.9753\n",
            "Epoch 672/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0586 - accuracy: 0.9846\n",
            "Epoch 673/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0613 - accuracy: 0.9753\n",
            "Epoch 674/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0571 - accuracy: 0.9722\n",
            "Epoch 675/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0615 - accuracy: 0.9753\n",
            "Epoch 676/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0680 - accuracy: 0.9784\n",
            "Epoch 677/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0597 - accuracy: 0.9753\n",
            "Epoch 678/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0580 - accuracy: 0.9753\n",
            "Epoch 679/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0575 - accuracy: 0.9753\n",
            "Epoch 680/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0566 - accuracy: 0.9753\n",
            "Epoch 681/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0554 - accuracy: 0.9660\n",
            "Epoch 682/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0550 - accuracy: 0.9660\n",
            "Epoch 683/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0575 - accuracy: 0.9753\n",
            "Epoch 684/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0573 - accuracy: 0.9753\n",
            "Epoch 685/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0573 - accuracy: 0.9722\n",
            "Epoch 686/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0543 - accuracy: 0.9691\n",
            "Epoch 687/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0541 - accuracy: 0.9784\n",
            "Epoch 688/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0545 - accuracy: 0.9753\n",
            "Epoch 689/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0543 - accuracy: 0.9722\n",
            "Epoch 690/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0597 - accuracy: 0.9753\n",
            "Epoch 691/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0565 - accuracy: 0.9753\n",
            "Epoch 692/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0552 - accuracy: 0.9722\n",
            "Epoch 693/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0537 - accuracy: 0.9753\n",
            "Epoch 694/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0556 - accuracy: 0.9753\n",
            "Epoch 695/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0527 - accuracy: 0.9753\n",
            "Epoch 696/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0540 - accuracy: 0.9784\n",
            "Epoch 697/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0567 - accuracy: 0.9722\n",
            "Epoch 698/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0563 - accuracy: 0.9722\n",
            "Epoch 699/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0551 - accuracy: 0.9722\n",
            "Epoch 700/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0539 - accuracy: 0.9722\n",
            "Epoch 701/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0536 - accuracy: 0.9722\n",
            "Epoch 702/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0533 - accuracy: 0.9753\n",
            "Epoch 703/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0532 - accuracy: 0.9753\n",
            "Epoch 704/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0526 - accuracy: 0.9722\n",
            "Epoch 705/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0514 - accuracy: 0.9753\n",
            "Epoch 706/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0518 - accuracy: 0.9722\n",
            "Epoch 707/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0555 - accuracy: 0.9784\n",
            "Epoch 708/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0579 - accuracy: 0.9722\n",
            "Epoch 709/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0560 - accuracy: 0.9722\n",
            "Epoch 710/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0540 - accuracy: 0.9691\n",
            "Epoch 711/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0539 - accuracy: 0.9691\n",
            "Epoch 712/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0526 - accuracy: 0.9722\n",
            "Epoch 713/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0502 - accuracy: 0.9753\n",
            "Epoch 714/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0533 - accuracy: 0.9722\n",
            "Epoch 715/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0511 - accuracy: 0.9784\n",
            "Epoch 716/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0506 - accuracy: 0.9753\n",
            "Epoch 717/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0510 - accuracy: 0.9722\n",
            "Epoch 718/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0502 - accuracy: 0.9753\n",
            "Epoch 719/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0615 - accuracy: 0.9722\n",
            "Epoch 720/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0578 - accuracy: 0.9753\n",
            "Epoch 721/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0538 - accuracy: 0.9722\n",
            "Epoch 722/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0541 - accuracy: 0.9722\n",
            "Epoch 723/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0628 - accuracy: 0.9722\n",
            "Epoch 724/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0579 - accuracy: 0.9784\n",
            "Epoch 725/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0564 - accuracy: 0.9784\n",
            "Epoch 726/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0506 - accuracy: 0.9753\n",
            "Epoch 727/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0520 - accuracy: 0.9722\n",
            "Epoch 728/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0502 - accuracy: 0.9722\n",
            "Epoch 729/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0497 - accuracy: 0.9753\n",
            "Epoch 730/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0481 - accuracy: 0.9815\n",
            "Epoch 731/800\n",
            "11/11 [==============================] - 0s 14ms/step - loss: 0.0503 - accuracy: 0.9784\n",
            "Epoch 732/800\n",
            "11/11 [==============================] - 0s 12ms/step - loss: 0.0518 - accuracy: 0.9722\n",
            "Epoch 733/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0475 - accuracy: 0.9753\n",
            "Epoch 734/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0499 - accuracy: 0.9753\n",
            "Epoch 735/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0484 - accuracy: 0.9753\n",
            "Epoch 736/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0616 - accuracy: 0.9722\n",
            "Epoch 737/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0569 - accuracy: 0.9722\n",
            "Epoch 738/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0495 - accuracy: 0.9753\n",
            "Epoch 739/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0531 - accuracy: 0.9784\n",
            "Epoch 740/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0501 - accuracy: 0.9691\n",
            "Epoch 741/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0524 - accuracy: 0.9691\n",
            "Epoch 742/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0551 - accuracy: 0.9660\n",
            "Epoch 743/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0493 - accuracy: 0.9722\n",
            "Epoch 744/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0485 - accuracy: 0.9722\n",
            "Epoch 745/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0474 - accuracy: 0.9753\n",
            "Epoch 746/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0542 - accuracy: 0.9784\n",
            "Epoch 747/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0498 - accuracy: 0.9784\n",
            "Epoch 748/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0495 - accuracy: 0.9722\n",
            "Epoch 749/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0480 - accuracy: 0.9753\n",
            "Epoch 750/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0474 - accuracy: 0.9753\n",
            "Epoch 751/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0481 - accuracy: 0.9784\n",
            "Epoch 752/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0491 - accuracy: 0.9660\n",
            "Epoch 753/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0483 - accuracy: 0.9691\n",
            "Epoch 754/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0658 - accuracy: 0.9660\n",
            "Epoch 755/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0580 - accuracy: 0.9722\n",
            "Epoch 756/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0518 - accuracy: 0.9722\n",
            "Epoch 757/800\n",
            "11/11 [==============================] - 0s 3ms/step - loss: 0.0483 - accuracy: 0.9753\n",
            "Epoch 758/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0501 - accuracy: 0.9691\n",
            "Epoch 759/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0467 - accuracy: 0.9753\n",
            "Epoch 760/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0563 - accuracy: 0.9722\n",
            "Epoch 761/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0505 - accuracy: 0.9753\n",
            "Epoch 762/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0505 - accuracy: 0.9722\n",
            "Epoch 763/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0498 - accuracy: 0.9753\n",
            "Epoch 764/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0472 - accuracy: 0.9784\n",
            "Epoch 765/800\n",
            "11/11 [==============================] - 0s 4ms/step - loss: 0.0469 - accuracy: 0.9784\n",
            "Epoch 766/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0616 - accuracy: 0.9753\n",
            "Epoch 767/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0594 - accuracy: 0.9753\n",
            "Epoch 768/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0519 - accuracy: 0.9722\n",
            "Epoch 769/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0503 - accuracy: 0.9753\n",
            "Epoch 770/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0466 - accuracy: 0.9753\n",
            "Epoch 771/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0484 - accuracy: 0.9691\n",
            "Epoch 772/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0472 - accuracy: 0.9753\n",
            "Epoch 773/800\n",
            "11/11 [==============================] - 0s 9ms/step - loss: 0.0482 - accuracy: 0.9722\n",
            "Epoch 774/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0474 - accuracy: 0.9722\n",
            "Epoch 775/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0497 - accuracy: 0.9722\n",
            "Epoch 776/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0482 - accuracy: 0.9784\n",
            "Epoch 777/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0476 - accuracy: 0.9691\n",
            "Epoch 778/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0482 - accuracy: 0.9753\n",
            "Epoch 779/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0489 - accuracy: 0.9722\n",
            "Epoch 780/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0460 - accuracy: 0.9722\n",
            "Epoch 781/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0456 - accuracy: 0.9784\n",
            "Epoch 782/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0729 - accuracy: 0.9753\n",
            "Epoch 783/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0566 - accuracy: 0.9722\n",
            "Epoch 784/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0499 - accuracy: 0.9784\n",
            "Epoch 785/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0471 - accuracy: 0.9722\n",
            "Epoch 786/800\n",
            "11/11 [==============================] - 0s 5ms/step - loss: 0.0463 - accuracy: 0.9722\n",
            "Epoch 787/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0489 - accuracy: 0.9784\n",
            "Epoch 788/800\n",
            "11/11 [==============================] - 0s 6ms/step - loss: 0.0445 - accuracy: 0.9784\n",
            "Epoch 789/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0459 - accuracy: 0.9753\n",
            "Epoch 790/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0477 - accuracy: 0.9722\n",
            "Epoch 791/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0614 - accuracy: 0.9815\n",
            "Epoch 792/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0541 - accuracy: 0.9753\n",
            "Epoch 793/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0527 - accuracy: 0.9691\n",
            "Epoch 794/800\n",
            "11/11 [==============================] - 0s 8ms/step - loss: 0.0583 - accuracy: 0.9722\n",
            "Epoch 795/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0492 - accuracy: 0.9753\n",
            "Epoch 796/800\n",
            "11/11 [==============================] - 0s 12ms/step - loss: 0.0441 - accuracy: 0.9753\n",
            "Epoch 797/800\n",
            "11/11 [==============================] - 0s 11ms/step - loss: 0.0442 - accuracy: 0.9660\n",
            "Epoch 798/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0452 - accuracy: 0.9722\n",
            "Epoch 799/800\n",
            "11/11 [==============================] - 0s 7ms/step - loss: 0.0453 - accuracy: 0.9784\n",
            "Epoch 800/800\n",
            "11/11 [==============================] - 0s 10ms/step - loss: 0.0682 - accuracy: 0.9753\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# to save the trained model\n",
        "model.save(\"chat_model\")\n",
        "\n",
        "# to save the fitted tokenizer\n",
        "with open('/content/drive/MyDrive/code/tokenizer.pickle', 'wb') as handle:\n",
        "    pickle.dump(tokenizer, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
        "\n",
        "# to save the fitted label encoder\n",
        "with open('/content/drive/MyDrive/code/label_encoder.pickle', 'wb') as ecn_file:\n",
        "    pickle.dump(lbl_encoder, ecn_file, protocol=pickle.HIGHEST_PROTOCOL)"
      ],
      "metadata": {
        "id": "hXasrqlcHaFK"
      },
      "id": "hXasrqlcHaFK",
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "id": "c4171997-a102-4e28-9d68-0e7bda7dc6c9",
      "metadata": {
        "id": "c4171997-a102-4e28-9d68-0e7bda7dc6c9"
      },
      "source": [
        "## Run chatbot"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 46,
      "id": "5cd023f7-61e8-4fae-8dc1-914ea75eb17b",
      "metadata": {
        "id": "5cd023f7-61e8-4fae-8dc1-914ea75eb17b"
      },
      "outputs": [],
      "source": [
        "class Predict:\n",
        "    '''Class for model inference.'''\n",
        "\n",
        "    def __init__(self, model, intents, intent_data, utt):\n",
        "        '''Args:\n",
        "            model (str): Variable name for model path.\n",
        "            intents (:obj:`list`): Ordered list of intent labels.\n",
        "            intent_data (:obj:, `dict`): Dictionary containing intents as keys\n",
        "            and and dictionaries of queries and responses as values.\n",
        "            utt (str): Utterance to be used to predict intent.\n",
        "        '''\n",
        "\n",
        "        self.intent_data = intent_data\n",
        "\n",
        "        self.intents = intents\n",
        "\n",
        "        self.model = model\n",
        "\n",
        "        # instantiate Utt object\n",
        "        parsed_utt = Utt(utt, query=True)\n",
        "\n",
        "        # preprocess utterance\n",
        "        self.utt = parsed_utt.parse_utt()\n",
        "\n",
        "        if self.utt:\n",
        "            self.predicted_intent = self._predict_intent()\n",
        "\n",
        "            if not self.predicted_intent:\n",
        "                self.response = \"Sorry, I didn't understand that. Please can you rephrase your message.\"\n",
        "            else:\n",
        "                self.response = self._get_response()\n",
        "        else:\n",
        "            self.response = \"Sorry, I didn't understand that. Please can you rephrase your message.\"\n",
        "\n",
        "    def _predict_intent(self):\n",
        "        '''Given utterance, predicts intent label.'''\n",
        "\n",
        "        query = np.array([self.utt], dtype=object)\n",
        "        predict = self.model.predict(query, batch_size=32, verbose=0)\n",
        "        predict_logit = predict.argmax(axis=1)\n",
        "        predicted_label = self.intents[predict_logit[0]]\n",
        "        if predicted_label:\n",
        "            return predicted_label\n",
        "        return None\n",
        "\n",
        "    def _get_response(self):\n",
        "        '''Given label, generates random response from list of available responses.'''\n",
        "\n",
        "        response_list = self.intent_data[self.predicted_intent]['2']\n",
        "        return random.choice(response_list)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "1c17dc7c-6adb-417d-813c-1784d8e50450",
      "metadata": {
        "id": "1c17dc7c-6adb-417d-813c-1784d8e50450"
      },
      "outputs": [],
      "source": [
        "# bot_name = \"ChatBot\"\n",
        "# print(\"How may we help you? (type 'quit' to exit)\")\n",
        "# while True:\n",
        "#     sentence = input(\"You: \")\n",
        "#     if sentence == \"quit\":\n",
        "#         break\n",
        "\n",
        "#     predict = Predict(model, intents, intent_data, sentence)\n",
        "#     print(f\"{bot_name}: {predict.response}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def chat():\n",
        "    # load trained model\n",
        "    model = keras.models.load_model('chat_model')\n",
        "\n",
        "    # load tokenizer object\n",
        "    with open('/content/drive/MyDrive/code/tokenizer.pickle', 'rb') as handle:\n",
        "        tokenizer = pickle.load(handle)\n",
        "\n",
        "    # load label encoder object\n",
        "    with open('/content/drive/MyDrive/code/label_encoder.pickle', 'rb') as enc:\n",
        "        lbl_encoder = pickle.load(enc)\n",
        "\n",
        "    # parameters\n",
        "    max_len = 20\n",
        "\n",
        "    while True:\n",
        "        print(Fore.LIGHTBLUE_EX + \"User: \" + Style.RESET_ALL, end=\"\")\n",
        "        inp = input()\n",
        "        if inp.lower() == \"quit\":\n",
        "            break\n",
        "\n",
        "        result = model.predict(keras.preprocessing.sequence.pad_sequences(tokenizer.texts_to_sequences([inp]),\n",
        "                                             truncating='post', maxlen=max_len))\n",
        "        tag = lbl_encoder.inverse_transform([np.argmax(result)])\n",
        "\n",
        "        for i in data['intents']:\n",
        "            if i['tag'] == tag:\n",
        "                print(Fore.GREEN + \"ChatBot:\" + Style.RESET_ALL , np.random.choice(i['responses']))\n",
        "\n",
        "print(Fore.YELLOW + \"Start messaging with the bot (type quit to stop)!\" + Style.RESET_ALL)\n",
        "chat()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8B-Ji7vqH3JI",
        "outputId": "20a4ac1b-97e6-4e8c-e1a9-181ee0f3e6b8"
      },
      "id": "8B-Ji7vqH3JI",
      "execution_count": 70,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Start messaging with the bot (type quit to stop)!\n",
            "User: Hello\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "ChatBot: Hola human, how can I help?\n",
            "User: What is your favorite programming language\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "ChatBot:  Python is the best language for creating chat robots.\n",
            "User: Tell me some gossip\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "ChatBot: Mike said he launched his browser and entered his name into the little slot when I asked him to.\n",
            "User: What do you call me?\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "ChatBot: Your name is  <HUMAN>, how can I help you?\n",
            "User: Thank You\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "ChatBot: Any time!\n",
            "User: See you later\n",
            "1/1 [==============================] - 0s 19ms/step\n",
            "ChatBot: Have a nice day\n",
            "User: quit\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "chatbot-final",
      "language": "python",
      "name": "chatbot-final"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.13"
    },
    "colab": {
      "provenance": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}