test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "import re\n",
    "import string\n",
    "\n",
    "# Function to clean the text data\n",
    "def clean_text(text):\n",
    "    text = text.lower()  # Convert to lowercase\n",
    "    text = re.sub(r'\\[.*?\\]', '', text)\n",
    "    text = re.sub(r'https?://\\S+|www\\.\\S+', '', text)\n",
    "    text = re.sub(r'<.*?>', '', text)\n",
    "    text = re.sub(r'[%s]' % re.escape(string.punctuation), '', text)\n",
    "    text = re.sub(r'\\n', '', text)\n",
    "    text = re.sub(r'\\w*\\d\\w*', '', text)\n",
    "    return text\n",
    "\n",
    "# Load the dataset\n",
    "file_path = './archive/train.csv'  # Update with the correct file path\n",
    "data = pd.read_csv(file_path)\n",
    "\n",
    "# Clean the tweets\n",
    "data['tweet'] = data['tweet'].apply(lambda x: clean_text(x))\n",
    "\n",
    "# Preparing the Tokenizer\n",
    "tokenizer = Tokenizer()\n",
    "tokenizer.fit_on_texts(data['tweet'])\n",
    "\n",
    "# Convert the text to sequences\n",
    "X = tokenizer.texts_to_sequences(data['tweet'])\n",
    "\n",
    "# Pad the sequences\n",
    "X = pad_sequences(X, maxlen=50)\n",
    "\n",
    "data['offensive_language_count_normalized'] = data['offensive_language_count'] / data['count']\n",
    "data['offensive_language_count_normalized'] = data['offensive_language_count_normalized'].apply(lambda x: 1 if x >= 0.9 else 0)\n",
    "\n",
    "# Prepare the target variable\n",
    "y = data['offensive_language_count_normalized'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "310/310 [==============================] - 25s 70ms/step - loss: 0.5668 - accuracy: 0.7050 - val_loss: 0.5173 - val_accuracy: 0.7743\n",
      "Epoch 2/5\n",
      "  2/310 [..............................] - ETA: 18s - loss: 0.3987 - accuracy: 0.8438"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/antoine/.local/lib/python3.10/site-packages/keras/src/engine/training.py:3079: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
      "  saving_api.save_model(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "310/310 [==============================] - 21s 67ms/step - loss: 0.4246 - accuracy: 0.8212 - val_loss: 0.5244 - val_accuracy: 0.7646\n",
      "Epoch 3/5\n",
      "310/310 [==============================] - 21s 66ms/step - loss: 0.3352 - accuracy: 0.8659 - val_loss: 0.5978 - val_accuracy: 0.7323\n",
      "Epoch 4/5\n",
      "310/310 [==============================] - 20s 66ms/step - loss: 0.2529 - accuracy: 0.9033 - val_loss: 0.6611 - val_accuracy: 0.7180\n",
      "Epoch 5/5\n",
      "310/310 [==============================] - 20s 65ms/step - loss: 0.1991 - accuracy: 0.9240 - val_loss: 0.7834 - val_accuracy: 0.7030\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.History at 0x7f427ad6fb50>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Embedding, LSTM, Dense, Dropout\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# Splitting the dataset into training and validation sets\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "\n",
    "# Model parameters\n",
    "vocab_size = len(tokenizer.word_index) + 1  # Plus 1 for padding token\n",
    "embed_size = 16\n",
    "\n",
    "# Building the model\n",
    "model = Sequential()\n",
    "model.add(Embedding(vocab_size, embed_size, input_length=50))\n",
    "model.add(LSTM(60, return_sequences=True))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(LSTM(60))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(1, activation='sigmoid'))  # 'sigmoid' for binary classification\n",
    "\n",
    "# Compile the model\n",
    "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "from sklearn.utils.class_weight import compute_class_weight\n",
    "import numpy as np\n",
    "\n",
    "# Compute class weights\n",
    "class_weights = compute_class_weight(class_weight='balanced', classes=np.unique(y_train), y=y_train)\n",
    "class_weights = dict(enumerate(class_weights))\n",
    "\n",
    "early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "checkpoint = ModelCheckpoint('./models/model-1.h5', monitor='val_accuracy', save_best_only=True)\n",
    "\n",
    "# Train the model with class weights\n",
    "model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=5, batch_size=64, class_weight=class_weights, callbacks=[early_stopping, checkpoint])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_hateness(message):\n",
    "    # Clean and preprocess the message\n",
    "    cleaned_message = clean_text(message)\n",
    "    sequence = tokenizer.texts_to_sequences([cleaned_message])\n",
    "    padded_sequence = pad_sequences(sequence, maxlen=50)\n",
    "\n",
    "    # Predict\n",
    "    prediction = model.predict(padded_sequence, verbose=0)\n",
    "    return prediction[0][0]*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = [\n",
    "    \"This restaurant is not worth your time or money.\",\n",
    "    \"Your work has been consistently outstanding.\",\n",
    "    \"He has a very controversial opinion on that.\",\n",
    "    \"I'm offended by your generalizations about people.\",\n",
    "    \"The team's performance was embarrassingly poor this season.\",\n",
    "    \"Your statement is ignorant and deeply offensive.\",\n",
    "    \"It's irresponsible to spread such misinformation.\",\n",
    "    \"The book was a groundbreaking piece of literature.\",\n",
    "    \"Their approach to problem-solving is quite innovative.\",\n",
    "    \"That comment was uncalled for and hurtful.\",\n",
    "    \"Your dedication to this project is truly inspiring.\",\n",
    "    \"I disagree with your methods and results.\",\n",
    "    \"The movie's plot was predictable and unoriginal.\",\n",
    "    \"Your assistance has been invaluable to our success.\",\n",
    "    \"Their political stance is quite extreme and divisive.\",\n",
    "    \"I've never experienced such terrible customer service.\",\n",
    "    \"Your creativity and vision are truly exceptional.\",\n",
    "    \"That joke was in poor taste and offensive.\",\n",
    "    \"The way they handle issues is highly professional.\",\n",
    "    \"I find your lack of empathy disturbing.\",\n",
    "    \"This new policy could be detrimental to our progress.\",\n",
    "    \"Your leadership has brought us to new heights.\",\n",
    "    \"Such behavior is unacceptable and disappointing.\",\n",
    "    \"Their perspective on this matter is refreshingly honest.\",\n",
    "    \"I can't believe how narrow-minded that statement was.\",\n",
    "    \"The quality of this product exceeded my expectations.\",\n",
    "    \"That speech was inflammatory and reckless.\",\n",
    "    \"Your contribution to the field is groundbreaking.\",\n",
    "    \"They have a tendency to make baseless accusations.\",\n",
    "    \"Her insight into the issue was incredibly enlightening.\",\n",
    "    \"That remark was sexist and inappropriate.\",\n",
    "    \"Your performance has consistently improved.\",\n",
    "    \"Their lack of responsibility in this matter is alarming.\",\n",
    "    \"I've never seen such a well-executed plan.\",\n",
    "    \"That's an outdated and prejudiced viewpoint.\",\n",
    "    \"Your innovative approach is exactly what we needed.\",\n",
    "    \"His comments were racially insensitive.\",\n",
    "    \"The service at this hotel is top-notch.\",\n",
    "    \"Such ignorance about the topic is astounding.\",\n",
    "    \"You have a talent for making complex topics accessible.\",\n",
    "    \"Their accusations were baseless and harmful.\",\n",
    "    \"The artwork you created was deeply moving.\",\n",
    "    \"That kind of language is unacceptable and harmful.\",\n",
    "    \"Your understanding of the subject is impressive.\",\n",
    "    \"Their stance on the issue is misguided and naive.\",\n",
    "    \"The attention to detail in this project is remarkable.\",\n",
    "    \"Your statement was biased and unfounded.\",\n",
    "    \"The innovation in this design is revolutionary.\",\n",
    "    \"That kind of behavior should not be tolerated.\",\n",
    "    \"You have an exceptional ability to motivate people.\",\n",
    "    \"Their argument is flawed and poorly constructed.\",\n",
    "    \"The craftsmanship of this product is exceptional.\",\n",
    "    \"Such a disrespectful comment is unacceptable.\",\n",
    "    \"Your ability to adapt to changes is commendable.\",\n",
    "    \"They have a habit of spreading rumors.\",\n",
    "    \"This painting is a masterpiece of modern art.\",\n",
    "    \"That approach is overly simplistic and ineffective.\",\n",
    "    \"Your problem-solving skills are extraordinary.\",\n",
    "    \"Their remarks were homophobic and unacceptable.\",\n",
    "    \"The level of skill in your work is astounding.\",\n",
    "    \"That sort of attitude is counterproductive and harmful.\",\n",
    "    \"Your vision for the project is innovative and exciting.\",\n",
    "    \"They often make discriminatory remarks.\",\n",
    "    \"This novel is a brilliant exploration of complex themes.\",\n",
    "    \"That methodology is outdated and inefficient.\",\n",
    "    \"Your insights have been crucial to our understanding.\",\n",
    "    \"Their commentary is often sexist and degrading.\",\n",
    "    \"The elegance of this design is unparalleled.\",\n",
    "    \"That kind of talk is divisive and unhelpful.\",\n",
    "    \"You have an impressive knack for leadership.\",\n",
    "    \"They display a concerning lack of judgment.\",\n",
    "    \"This sculpture is a stunning representation of abstract art.\",\n",
    "    \"That opinion is short-sighted and uneducated.\",\n",
    "    \"Your strategic planning has been a game-changer.\",\n",
    "    \"Their views are often extremist and polarizing.\",\n",
    "    \"The complexity of this issue has been well-handled.\",\n",
    "    \"That statement is a gross oversimplification.\",\n",
    "    \"You have a profound understanding of the topic.\",\n",
    "    \"They frequently engage in unethical practices.\",\n",
    "    \"This film is a remarkable achievement in cinema.\",\n",
    "    \"That assumption is both arrogant and incorrect.\",\n",
    "    \"Your ability to overcome obstacles is inspiring.\",\n",
    "    \"They tend to be dismissive of important issues.\",\n",
    "    \"The texture and color in this artwork are extraordinary.\",\n",
    "    \"That kind of statement is fear-mongering.\",\n",
    "    \"You've shown remarkable growth and development.\",\n",
    "    \"Their handling of the situation was inept and careless.\",\n",
    "    \"This song is an anthem for a generation.\",\n",
    "    \"That view is narrow-minded and regressive.\",\n",
    "    \"Your resilience in the face of adversity is admirable.\",\n",
    "    \"They often resort to personal attacks.\",\n",
    "    \"The innovation in this technology is groundbreaking.\",\n",
    "    \"That kind of rhetoric is dangerous and misleading.\",\n",
    "    \"You have a unique ability to connect with people.\",\n",
    "    \"Their interpretation of the data is misleading.\",\n",
    "    \"This artwork is a vivid portrayal of contemporary issues.\",\n",
    "    \"That strategy is unethical and irresponsible.\",\n",
    "    \"Your work ethic is unmatched and commendable.\",\n",
    "    \"They frequently display a lack of empathy.\",\n",
    "    \"This piece is a thought-provoking exploration of identity.\"\n",
    "    ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\"This restaurant is not worth your time or money.\":\t87.71 \n",
      "\"Your work has been consistently outstanding.\":\t6.14 \n",
      "\"He has a very controversial opinion on that.\":\t8.94 \n",
      "\"I'm offended by your generalizations about people.\":\t97.97 \n",
      "\"The team's performance was embarrassingly poor this season.\":\t25.23 \n",
      "\"Your statement is ignorant and deeply offensive.\":\t0.24 \n",
      "\"It's irresponsible to spread such misinformation.\":\t97.30 \n",
      "\"The book was a groundbreaking piece of literature.\":\t89.35 \n",
      "\"Their approach to problem-solving is quite innovative.\":\t6.13 \n",
      "\"That comment was uncalled for and hurtful.\":\t66.97 \n",
      "\"Your dedication to this project is truly inspiring.\":\t96.03 \n",
      "\"I disagree with your methods and results.\":\t13.59 \n",
      "\"The movie's plot was predictable and unoriginal.\":\t95.26 \n",
      "\"Your assistance has been invaluable to our success.\":\t78.13 \n",
      "\"Their political stance is quite extreme and divisive.\":\t0.13 \n",
      "\"I've never experienced such terrible customer service.\":\t96.31 \n",
      "\"Your creativity and vision are truly exceptional.\":\t81.57 \n",
      "\"That joke was in poor taste and offensive.\":\t1.51 \n",
      "\"The way they handle issues is highly professional.\":\t98.44 \n",
      "\"I find your lack of empathy disturbing.\":\t1.32 \n",
      "\"This new policy could be detrimental to our progress.\":\t43.47 \n",
      "\"Your leadership has brought us to new heights.\":\t36.56 \n",
      "\"Such behavior is unacceptable and disappointing.\":\t65.45 \n",
      "\"Their perspective on this matter is refreshingly honest.\":\t92.35 \n",
      "\"I can't believe how narrow-minded that statement was.\":\t31.10 \n",
      "\"The quality of this product exceeded my expectations.\":\t8.27 \n",
      "\"That speech was inflammatory and reckless.\":\t3.88 \n",
      "\"Your contribution to the field is groundbreaking.\":\t17.44 \n",
      "\"They have a tendency to make baseless accusations.\":\t35.04 \n",
      "\"Her insight into the issue was incredibly enlightening.\":\t94.00 \n",
      "\"That remark was sexist and inappropriate.\":\t14.23 \n",
      "\"Your performance has consistently improved.\":\t17.30 \n",
      "\"Their lack of responsibility in this matter is alarming.\":\t2.53 \n",
      "\"I've never seen such a well-executed plan.\":\t78.67 \n",
      "\"That's an outdated and prejudiced viewpoint.\":\t18.04 \n",
      "\"Your innovative approach is exactly what we needed.\":\t35.06 \n",
      "\"His comments were racially insensitive.\":\t79.20 \n",
      "\"The service at this hotel is top-notch.\":\t89.44 \n",
      "\"Such ignorance about the topic is astounding.\":\t89.09 \n",
      "\"You have a talent for making complex topics accessible.\":\t93.14 \n",
      "\"Their accusations were baseless and harmful.\":\t13.96 \n",
      "\"The artwork you created was deeply moving.\":\t41.48 \n",
      "\"That kind of language is unacceptable and harmful.\":\t0.69 \n",
      "\"Your understanding of the subject is impressive.\":\t73.82 \n",
      "\"Their stance on the issue is misguided and naive.\":\t9.96 \n",
      "\"The attention to detail in this project is remarkable.\":\t47.33 \n",
      "\"Your statement was biased and unfounded.\":\t51.91 \n",
      "\"The innovation in this design is revolutionary.\":\t23.42 \n",
      "\"That kind of behavior should not be tolerated.\":\t96.30 \n",
      "\"You have an exceptional ability to motivate people.\":\t10.58 \n",
      "\"Their argument is flawed and poorly constructed.\":\t51.02 \n",
      "\"The craftsmanship of this product is exceptional.\":\t65.71 \n",
      "\"Such a disrespectful comment is unacceptable.\":\t83.82 \n",
      "\"Your ability to adapt to changes is commendable.\":\t73.22 \n",
      "\"They have a habit of spreading rumors.\":\t54.45 \n",
      "\"This painting is a masterpiece of modern art.\":\t97.88 \n",
      "\"That approach is overly simplistic and ineffective.\":\t97.92 \n",
      "\"Your problem-solving skills are extraordinary.\":\t3.48 \n",
      "\"Their remarks were homophobic and unacceptable.\":\t3.80 \n",
      "\"The level of skill in your work is astounding.\":\t92.53 \n",
      "\"That sort of attitude is counterproductive and harmful.\":\t20.38 \n",
      "\"Your vision for the project is innovative and exciting.\":\t10.85 \n",
      "\"They often make discriminatory remarks.\":\t90.20 \n",
      "\"This novel is a brilliant exploration of complex themes.\":\t79.16 \n",
      "\"That methodology is outdated and inefficient.\":\t47.75 \n",
      "\"Your insights have been crucial to our understanding.\":\t89.32 \n",
      "\"Their commentary is often sexist and degrading.\":\t34.16 \n",
      "\"The elegance of this design is unparalleled.\":\t15.61 \n",
      "\"That kind of talk is divisive and unhelpful.\":\t9.84 \n",
      "\"You have an impressive knack for leadership.\":\t0.17 \n",
      "\"They display a concerning lack of judgment.\":\t0.58 \n",
      "\"This sculpture is a stunning representation of abstract art.\":\t91.74 \n",
      "\"That opinion is short-sighted and uneducated.\":\t2.48 \n",
      "\"Your strategic planning has been a game-changer.\":\t71.51 \n",
      "\"Their views are often extremist and polarizing.\":\t63.60 \n",
      "\"The complexity of this issue has been well-handled.\":\t42.94 \n",
      "\"That statement is a gross oversimplification.\":\t25.09 \n",
      "\"You have a profound understanding of the topic.\":\t86.40 \n",
      "\"They frequently engage in unethical practices.\":\t15.38 \n",
      "\"This film is a remarkable achievement in cinema.\":\t84.30 \n",
      "\"That assumption is both arrogant and incorrect.\":\t2.24 \n",
      "\"Your ability to overcome obstacles is inspiring.\":\t3.49 \n",
      "\"They tend to be dismissive of important issues.\":\t1.70 \n",
      "\"The texture and color in this artwork are extraordinary.\":\t54.61 \n",
      "\"That kind of statement is fear-mongering.\":\t9.33 \n",
      "\"You've shown remarkable growth and development.\":\t0.22 \n",
      "\"Their handling of the situation was inept and careless.\":\t25.35 \n",
      "\"This song is an anthem for a generation.\":\t70.24 \n",
      "\"That view is narrow-minded and regressive.\":\t35.95 \n",
      "\"Your resilience in the face of adversity is admirable.\":\t1.21 \n",
      "\"They often resort to personal attacks.\":\t53.12 \n",
      "\"The innovation in this technology is groundbreaking.\":\t93.55 \n",
      "\"That kind of rhetoric is dangerous and misleading.\":\t9.93 \n",
      "\"You have a unique ability to connect with people.\":\t3.25 \n",
      "\"Their interpretation of the data is misleading.\":\t83.49 \n",
      "\"This artwork is a vivid portrayal of contemporary issues.\":\t73.98 \n",
      "\"That strategy is unethical and irresponsible.\":\t78.24 \n",
      "\"Your work ethic is unmatched and commendable.\":\t30.73 \n",
      "\"They frequently display a lack of empathy.\":\t8.76 \n",
      "\"This piece is a thought-provoking exploration of identity.\":\t52.27 \n"
     ]
    }
   ],
   "source": [
    "for t in test:\n",
    "    print(f\"\\\"{t}\\\":\\t{predict_hateness(t):2.2f} \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}