model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from giza_datasets import DatasetsLoader\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "import tensorflow as tf\n",
    "\n",
    "plt.style.use(\"seaborn-v0_8\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset read from cache.\n",
      "Loading dataset tokens-ohcl from cache.\n",
      "Min price: 0.0226, Max price: 82070.7575\n",
      "Min price: 0.0, Max price: 1.0\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>token</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-02-03</th>\n",
       "      <td>0.041895</td>\n",
       "      <td>0.042308</td>\n",
       "      <td>0.041895</td>\n",
       "      <td>0.042171</td>\n",
       "      <td>WBTC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-07</th>\n",
       "      <td>0.042258</td>\n",
       "      <td>0.042480</td>\n",
       "      <td>0.041743</td>\n",
       "      <td>0.041743</td>\n",
       "      <td>WBTC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-11</th>\n",
       "      <td>0.041278</td>\n",
       "      <td>0.045940</td>\n",
       "      <td>0.041278</td>\n",
       "      <td>0.045940</td>\n",
       "      <td>WBTC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-15</th>\n",
       "      <td>0.043936</td>\n",
       "      <td>0.044509</td>\n",
       "      <td>0.043932</td>\n",
       "      <td>0.043932</td>\n",
       "      <td>WBTC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-19</th>\n",
       "      <td>0.044033</td>\n",
       "      <td>0.046684</td>\n",
       "      <td>0.044033</td>\n",
       "      <td>0.046684</td>\n",
       "      <td>WBTC</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close token\n",
       "date                                                    \n",
       "2019-02-03  0.041895  0.042308  0.041895  0.042171  WBTC\n",
       "2019-02-07  0.042258  0.042480  0.041743  0.041743  WBTC\n",
       "2019-02-11  0.041278  0.045940  0.041278  0.045940  WBTC\n",
       "2019-02-15  0.043936  0.044509  0.043932  0.043932  WBTC\n",
       "2019-02-19  0.044033  0.046684  0.044033  0.046684  WBTC"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>token</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2024-01-07</th>\n",
       "      <td>0.026857</td>\n",
       "      <td>0.027587</td>\n",
       "      <td>0.026857</td>\n",
       "      <td>0.027287</td>\n",
       "      <td>FRXETH</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-01-11</th>\n",
       "      <td>0.027025</td>\n",
       "      <td>0.030961</td>\n",
       "      <td>0.027025</td>\n",
       "      <td>0.030961</td>\n",
       "      <td>FRXETH</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-01-15</th>\n",
       "      <td>0.031913</td>\n",
       "      <td>0.031913</td>\n",
       "      <td>0.030238</td>\n",
       "      <td>0.030238</td>\n",
       "      <td>FRXETH</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-01-19</th>\n",
       "      <td>0.030656</td>\n",
       "      <td>0.031251</td>\n",
       "      <td>0.029902</td>\n",
       "      <td>0.029902</td>\n",
       "      <td>FRXETH</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2024-01-23</th>\n",
       "      <td>0.030072</td>\n",
       "      <td>0.030072</td>\n",
       "      <td>0.029845</td>\n",
       "      <td>0.029845</td>\n",
       "      <td>FRXETH</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close   token\n",
       "date                                                      \n",
       "2024-01-07  0.026857  0.027587  0.026857  0.027287  FRXETH\n",
       "2024-01-11  0.027025  0.030961  0.027025  0.030961  FRXETH\n",
       "2024-01-15  0.031913  0.031913  0.030238  0.030238  FRXETH\n",
       "2024-01-19  0.030656  0.031251  0.029902  0.029902  FRXETH\n",
       "2024-01-23  0.030072  0.030072  0.029845  0.029845  FRXETH"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Most recent date: 2024-01-23 00:00:00\n"
     ]
    }
   ],
   "source": [
    "loader = DatasetsLoader()\n",
    "df = loader.load('tokens-ohcl')\n",
    "data = df.to_pandas()\n",
    "data.set_index(\"date\", inplace=True)    # index the date column\n",
    "\n",
    "prices = ['Open', 'High', 'Low', 'Close']   # the prices to predict\n",
    "\n",
    "# normalize the 4 columns\n",
    "min_price = data[prices].values.min()\n",
    "max_price = data[prices].values.max()\n",
    "print(f\"Min price: {min_price}, Max price: {max_price}\")\n",
    "\n",
    "data[prices] = (data[prices] - min_price) / (max_price - min_price)\n",
    "print(f\"Min price: {data[prices].values.min()}, Max price: {data[prices].values.max()}\")\n",
    "\n",
    "# min_price = min_price.values\n",
    "# max_price = max_price.values\n",
    "\n",
    "display(data.head())\n",
    "display(data.tail())\n",
    "\n",
    "# get the most recent date in the dataset\n",
    "most_recent_date = data.index.max()\n",
    "print(f\"Most recent date: {most_recent_date}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (11_138, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>date</th><th>Open</th><th>High</th><th>Low</th><th>Close</th><th>token</th></tr><tr><td>date</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>str</td></tr></thead><tbody><tr><td>2019-02-03</td><td>3438.3604</td><td>3472.2433</td><td>3438.3604</td><td>3461.0583</td><td>&quot;WBTC&quot;</td></tr><tr><td>2019-02-07</td><td>3468.16</td><td>3486.4073</td><td>3425.8603</td><td>3425.8603</td><td>&quot;WBTC&quot;</td></tr><tr><td>2019-02-11</td><td>3387.7629</td><td>3770.3402</td><td>3387.7629</td><td>3770.3402</td><td>&quot;WBTC&quot;</td></tr><tr><td>2019-02-15</td><td>3605.9233</td><td>3652.9015</td><td>3605.5237</td><td>3605.5237</td><td>&quot;WBTC&quot;</td></tr><tr><td>2019-02-19</td><td>3613.8624</td><td>3831.385</td><td>3613.8624</td><td>3831.385</td><td>&quot;WBTC&quot;</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>2024-01-07</td><td>2204.1786</td><td>2264.1205</td><td>2204.1786</td><td>2239.5087</td><td>&quot;FRXETH&quot;</td></tr><tr><td>2024-01-11</td><td>2217.9515</td><td>2540.9835</td><td>2217.9515</td><td>2540.9835</td><td>&quot;FRXETH&quot;</td></tr><tr><td>2024-01-15</td><td>2619.135</td><td>2619.135</td><td>2481.6893</td><td>2481.6893</td><td>&quot;FRXETH&quot;</td></tr><tr><td>2024-01-19</td><td>2515.9513</td><td>2564.8498</td><td>2454.1049</td><td>2454.1049</td><td>&quot;FRXETH&quot;</td></tr><tr><td>2024-01-23</td><td>2468.0286</td><td>2468.0286</td><td>2449.4172</td><td>2449.4172</td><td>&quot;FRXETH&quot;</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (11_138, 6)\n",
       "┌────────────┬───────────┬───────────┬───────────┬───────────┬────────┐\n",
       "│ date       ┆ Open      ┆ High      ┆ Low       ┆ Close     ┆ token  │\n",
       "│ ---        ┆ ---       ┆ ---       ┆ ---       ┆ ---       ┆ ---    │\n",
       "│ date       ┆ f64       ┆ f64       ┆ f64       ┆ f64       ┆ str    │\n",
       "╞════════════╪═══════════╪═══════════╪═══════════╪═══════════╪════════╡\n",
       "│ 2019-02-03 ┆ 3438.3604 ┆ 3472.2433 ┆ 3438.3604 ┆ 3461.0583 ┆ WBTC   │\n",
       "│ 2019-02-07 ┆ 3468.16   ┆ 3486.4073 ┆ 3425.8603 ┆ 3425.8603 ┆ WBTC   │\n",
       "│ 2019-02-11 ┆ 3387.7629 ┆ 3770.3402 ┆ 3387.7629 ┆ 3770.3402 ┆ WBTC   │\n",
       "│ 2019-02-15 ┆ 3605.9233 ┆ 3652.9015 ┆ 3605.5237 ┆ 3605.5237 ┆ WBTC   │\n",
       "│ 2019-02-19 ┆ 3613.8624 ┆ 3831.385  ┆ 3613.8624 ┆ 3831.385  ┆ WBTC   │\n",
       "│ …          ┆ …         ┆ …         ┆ …         ┆ …         ┆ …      │\n",
       "│ 2024-01-07 ┆ 2204.1786 ┆ 2264.1205 ┆ 2204.1786 ┆ 2239.5087 ┆ FRXETH │\n",
       "│ 2024-01-11 ┆ 2217.9515 ┆ 2540.9835 ┆ 2217.9515 ┆ 2540.9835 ┆ FRXETH │\n",
       "│ 2024-01-15 ┆ 2619.135  ┆ 2619.135  ┆ 2481.6893 ┆ 2481.6893 ┆ FRXETH │\n",
       "│ 2024-01-19 ┆ 2515.9513 ┆ 2564.8498 ┆ 2454.1049 ┆ 2454.1049 ┆ FRXETH │\n",
       "│ 2024-01-23 ┆ 2468.0286 ┆ 2468.0286 ┆ 2449.4172 ┆ 2449.4172 ┆ FRXETH │\n",
       "└────────────┴───────────┴───────────┴───────────┴───────────┴────────┘"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30 308 4\n",
      "30 208 101 4\n",
      "(8138, 101, 4)\n",
      "X_train shape: (6510, 100, 4)\n",
      "y_train shape: (6510, 4)\n",
      "X_test shape: (1628, 100, 4)\n",
      "y_test shape: (1628, 4)\n"
     ]
    }
   ],
   "source": [
    "records_per_token = data.groupby(\"token\")[prices].apply(lambda x: x.values.tolist()).to_list()\n",
    "print(len(records_per_token), len(records_per_token[0]), len(records_per_token[0][0]))      # shape: (n_tokens, n_records_per_token, n_columns)\n",
    "\n",
    "# create sequences of length len_sequence each for example [a,b,c,d,e] -> [[a,b,c],[b,c,d],[c,d,e]] where len_sequence = 3\n",
    "len_sequence = 101\n",
    "sequences = []\n",
    "for token_records in records_per_token:\n",
    "    token_sequences = []\n",
    "    for i in range(len(token_records) - len_sequence + 1):\n",
    "        sequence = token_records[i:i+len_sequence]\n",
    "        token_sequences.append(sequence)\n",
    "    sequences.append(token_sequences)\n",
    "\n",
    "print(len(sequences), len(sequences[0]), len(sequences[0][0]), len(sequences[0][0][0]))      # shape: (n_tokens, n_sequences_per_token, len_sequence, n_columns)\n",
    "\n",
    "sequences = [sequence for token_sequences in sequences for sequence in token_sequences] # concatenate all the sequences of all the tokens\n",
    "sequences = np.array(sequences)\n",
    "print(sequences.shape)    # shape: (n_total_sequences, len_sequence-1, n_columns)\n",
    "\n",
    "# Split the data into training and testing sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(sequences[:, :len_sequence-1], sequences[:, len_sequence-1], test_size=0.2, random_state=42)\n",
    "print(\"X_train shape:\", X_train.shape)  # shape: (n_train_sequences, len_sequence-1, n_columns)\n",
    "print(\"y_train shape:\", y_train.shape)  # shape: (n_train_sequences, n_columns)\n",
    "print(\"X_test shape:\", X_test.shape)    # shape: (n_test_sequences, len_sequence-1, n_columns)\n",
    "print(\"y_test shape:\", y_test.shape)    # shape: (n_test_sequences, n_columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 40ms/step - loss: 0.0012 - val_loss: 2.2565e-04\n",
      "Epoch 2/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 39ms/step - loss: 1.4774e-04 - val_loss: 1.4421e-04\n",
      "Epoch 3/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 40ms/step - loss: 1.1300e-04 - val_loss: 1.3631e-04\n",
      "Epoch 4/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 37ms/step - loss: 9.7490e-05 - val_loss: 1.0050e-04\n",
      "Epoch 5/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 37ms/step - loss: 8.2726e-05 - val_loss: 7.3798e-05\n",
      "Epoch 6/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 37ms/step - loss: 6.5381e-05 - val_loss: 5.3217e-05\n",
      "Epoch 7/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 40ms/step - loss: 8.4367e-05 - val_loss: 8.9898e-05\n",
      "Epoch 8/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 40ms/step - loss: 7.3519e-05 - val_loss: 6.1565e-05\n",
      "Epoch 9/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 37ms/step - loss: 7.8533e-05 - val_loss: 4.8012e-05\n",
      "Epoch 10/10\n",
      "\u001b[1m407/407\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 37ms/step - loss: 5.4335e-05 - val_loss: 5.2286e-05\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Loss Evolution')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAF0CAYAAAAKKLFNAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8fJSN1AAAACXBIWXMAAA9hAAAPYQGoP6dpAABd7ElEQVR4nO3deXhTZd7/8XeWZum+r7RlX0QspQVRURZFwXGQQZwRfdxAQQVx40ERfywiohVkRnEDBXHkUQZlXJDBdURFBSm0FbHKTuma0n1LmuX3R9pAKLUtpE3Sfl/XxdX0nJOcO19TPzn3uc99FDabzYYQQgghvIbS3Q0QQgghRNtIeAshhBBeRsJbCCGE8DIS3kIIIYSXkfAWQgghvIyEtxBCCOFlJLyFEEIILyPhLYQQQngZCW8hhFeS+aVEVybhLUQHu/XWW7n11lvd3Yxm9evX7w//LV++3KX7e+yxxxgzZkybnnPgwAGmTJnitKxfv368+OKLrmyaEB5L7e4GCCE8z+TJk7nxxhvPui4qKqqDW9PUtm3b2Lt3r9OyjRs3Eh0d7aYWCdGxJLyFEE1ER0czePBgdzejTbytvUKcD+k2F8JD7dixg5tvvpmUlBQuvvhiHnnkEfLz8x3rrVYrK1euZMyYMVx44YWMGTOGFStWUF9f79hmy5YtTJgwgYsuuojhw4czZ84cCgsLXdK+a665htmzZzdZfv3113Pvvfc6ft+6dSuTJk0iOTmZyy67jAULFlBeXt7s656t+/vFF1+kX79+jserVq1qsu2ZzysqKmLevHmMHDmSiy66iMmTJ/Pll1822deGDRuYP38+w4YNIzk5mQceeIDi4uI2VkOIjiXhLYQH+uCDD5g6dSoxMTE8//zzzJs3j7179/K3v/2NkydPArBmzRreeecdZs6cydq1a5kyZQpvvPEGr7zyCgDp6enMnTuXq6++mjVr1jBv3jx+/PFHHnnkkRb3b7VaMZvNZ/3XaMKECWzfvp2qqirHskOHDpGdnc31118PwMsvv8zDDz/M4MGDeeGFF5g5cyaffvopt956K3V1dedUmxtvvJHJkycD9q7ys3XvFxcXM3nyZHbv3s1DDz3Eiy++SFxcHDNnzuSjjz5y2nblypVYrVaef/555s6dy3//+1+efvrpc2qbEB1Fus2F8DBWq5Xly5czYsQIVqxY4Vg+ZMgQrr32Wt544w3mzp3Lrl27uPDCC7nhhhsAGDZsGHq9noCAAMAe3jqdjunTp6PRaAAIDg7m559/xmazoVAomm3Dyy+/zMsvv3zWdT/88AOhoaFMmDCBF198kS+++IKJEycC9iP9wMBAxowZQ3l5Oa+88gp//etfWbBggeP5ffv25ZZbbuH999/nlltuaXN9oqOjHee2m+sqX7duHSUlJXz66afExcUBMHLkSO644w7S0tK47rrrUCqVjvYsW7bM8dysrCy2bdvW5nYJ0ZEkvIXwMEeOHMFgMDQ5Qk5ISCA5OZldu3YBcPHFF7NixQpuvvlmxowZw6hRo/if//kfx/ZDhw5l5cqVXHfddVxzzTWMHDmSESNGMHLkyBbb8Ne//pW//vWvZ10XGBgIQHx8PEOGDGHr1q2O8P7kk08YN24cGo2GH374AZPJxHXXXef0/NTUVOLi4ti1a9c5hXdr7Nq1i+TkZEdwN5owYQLz5s3j8OHD9O7dG2j6BSA6Opra2tp2aZcQriLd5kJ4mLKyMgDCw8ObrAsPD6eyshKAu+66iwULFlBXV8fy5cv505/+xHXXXcePP/4IQHJyMqtXryY+Pp5169Zxyy23cMUVV/DPf/6zxTZERkYyaNCgs/5Tq09957/++uv5/vvvKS0t5eeff+bYsWOOLvPG89otvY/2UF5eTkRExFn3C1BRUeFYptfrnbZRKpVyDbnweBLeQniY4OBggLMOmjIYDISEhAD2kLnlllvYvHkzO3bsYNmyZZhMJu6//35MJhMAl19+OW+88QY//fQTr776Kn379uWpp54iKyvLJW0dP348CoWCL774gq1btxIXF0dKSgoAQUFBrXofZ2OxWJx+r6mpaVO7goKCMBgMZ90v8If7FsIbSHgL4WF69OhBREQEW7ZscVqek5NDRkYGQ4YMAeCmm27iqaeeAiAsLIxJkyZxyy23UFFRQVVVFc8++yw33HADNpsNvV7P6NGjefTRRwHIy8tzSVsDAwMZPXo0X375JZ9++ikTJkxwnEtPSkpCo9E0eR+7d+8mLy/P8T7O5O/v32RE/J49e5x+bzxf3ZyhQ4eyd+9ecnNznZZ/9NFHREREkJiY2Kr3J4SnknPeQrhBQUEBb775ZpPlffv25dJLL+Xhhx9m3rx5PPLII0yYMIHS0lJWrVpFUFAQd955J2APqLVr1xIeHk5ycjKFhYWsW7eOYcOGERoayvDhw1m3bh2PPfYYEyZMoL6+ntdff53g4GCGDx/eYvsyMjLOuk6v1zsu2wL7eeTZs2djsVgcXeZg70GYPn06L730Ej4+PowePZoTJ07wj3/8g969e/OXv/zlrK8/atQoPvnkE5KSkkhMTGTz5s0cO3bMaZvG8+5btmwhKSmJ+Ph4p/V33nknH330EXfccQezZs0iODiYDz74gB9//JGnn366xfAXwtNJeAvhBsePH3ca4dxo8uTJXHrppUyaNAk/Pz9ee+01Zs6cib+/P5dffjkPP/yw41zuAw88gEaj4f333+ell14iICCAMWPGOAa6jRw5kuXLl7N27VpmzZqFQqEgJSWFt956y9E135z33nuP995776zr+vfvz4cffuj4feTIkQQEBBAfH0+PHj2ctr3//vsJDw/n7bffZuPGjQQHBzNu3DgefPBBfH19z/r68+bNw2w28+yzz6JWq7n22mt55JFHeOKJJxzbXH311Xz44Yc89thjTJ48mUWLFjm9RkREBO+88w4rVqzgqaeeor6+nv79+/Pyyy9z5ZVX/uF7F8IbKGwyMkMIIYTwKtJ3JIQQQngZCW8hhBDCy0h4CyGEEF5GwlsIIYTwMhLeQgghhJeR8BZCCCG8jIS3EEII4WW67CQtBoPrboqgVCoIDfWjpKQaq1Uum3cFqanrSU3bh9TV9bpyTSMiAlq1nRx5u4BSqUChUKBUNn9/ZNE2UlPXk5q2D6mr60lNWybhLYQQQngZCW8hhBDCy0h4CyGEEF5GwlsIIYTwMhLeQgghhJeR8BZCCCG8jIS3EEII4WUkvIUQQggvI+EthBBCeBkJbyGEEMLLSHgLIYQQXqbL3pjEVX45WsI7Xxzgpqv7MbhnqLubI4QQHmPp0kX85z9bml3/wguvMmRIaptec9as6SQnpzBt2ow2t2fy5D8zdep0rr32z21+rqeR8D5Pe34zkFdczcbPf2fwjOHubo4QQniMBx6Ywz33zALgyy8/591332bNmvWO9YGBQW1+zaeffg612sdlbfRWEt7nKSpED0BecRV1JjNqpZyJEEIIAH9/f/z9/R2PlUolYWHh5/Wa5xL4nZEkzXlKiLLfe9Vmg5zCKje3RgghvEd+fh4jRqTy5puvM27caJ5//llsNhtvvvkGY8aMYcSIYVx//TjWrl3teM6sWdN5443XAHu3/IsvPs+CBfO48srLmDTpT2zb9sk5t2ffvizuvXcaV101ghtvnMAHH7znWFdQUMBDD81k7NjLue66saxcmYbZbAbgwIHfueeeqVx55WVMnDiedevWnHMbWkuOvM9TQpS/4/HRgkp6xAS6sTVCiK6kps5Mfkl1h+4zJtQPX51royMrK5M33vgnVquVbds+YePG/2PlypUEBobz/fffsXz5M1x22RX069e/yXPff/9f3H33vcyYMZP33tvIc889zYgRIx1H/K119OgRZs++l7/97Wbmzft//PLLPlaseIaQkDBGjhzN3/+ehl7vy7p1/0dpaQlPPDGXxMQeTJp0I089tZCLLhrMggVLOH78GE88MZf+/QdwySUjXFWiJiS8z5OvzofIED1FpbUcLagA4tzdJCFEF1BTZ2buK99TYzR36H59tWrS7r3UpQH+179OIS6uGwAGQxFPPLGISy65hNLSaiZOnMy6dWs4cuTQWcO7d+++3HLL7QDcddcMNm16hyNHDjFoUFKb2vDxx/+mb99+zJgxE4CEhO4cPXqE//u/txg5cjT5+fn069ef6OgYunWL57nn/kFAgP1graAgj8svH0l0dAyxsXH8/e8vExMTez4laZF0m7tAYrS96/xYQaWbWyKEEN7n9KAbMiSVoKBgVqxYwdy5D3PDDddx8uRJrFbrWZ/brVu847Gfn/1ou7E7uy2OHj3KBRcMdFo2aNBFHDt2BIBbbrmNzz77D9dddxULFz5OQUGBo9233non69e/wfXXj2PZsicxmUznfW6/JXLk7QLdowP46dcicg3V1Jut+KjlO5EQon356uxHwJ2h21yj0Tgef/zxB7z44vPceOONjB49hvvue4DZs+9p9rk+Pk1HnttstvNqQyOLxYrFYv/ScPXV40lJGcq3337N999/x//7f49yyy23M336ffzP/9zBmDFj+eab/7Jjx7c88MC9zJ07nz//eWKb29FaEt4ukBht7zqxWG3kFlfRPVrOewsh2p+vTk2v2M41+vqDD95n6tS7uf/++ygtraa0tJySkpPnFMhtkZCQSEbGHqdlv/ySRUJCIgCvvfYSY8aMZeLEyUycOJl//vNNtm3bwu23T+OVV17klltu46ab/oebbvofnnvuab7++isJb0/XvaHbHOB4oYS3EEKcq6CgIH76aSdHjoynoKCYl19ehdlspr7e5JLXP3ToID/++L3TsgEDLuAvf7mRTZve5bXXXmL8+Ov45Zef2bx5Ew89NBeA48ePsnJlGg8//ChKpZIff9xBnz790Gq1ZGVlUFRUyD33zKSmpobMzL1cfvkol7S3ORLeLhDopyEsSMfJ8jr7ee+2jZMQQgjR4IEH5rBs2WKuv/56QkJCGDNmLDqdnt9//80lr79x4wY2btzgtGzlypcYOvRi0tJW8vLL/+Ddd98mKiqaWbMe4k9/mgDAnDnzWLHiGWbNmo7FYuHSSy/jwQf/F4Ann1zG888/y1133Y5KpWLMmKu4445pLmlvcxS29u6L8FAGg+sGl6nVSl7c/DM/7S+kZ2wgT9zWtun+RFNqtZKQED9KS6sxm88+UEW0jdS0fUhdXa8r1zQiIqDljZDR5i7TKy4YgBNFVViaGRUphBBCuIKEt4v06mYfNGIyWyk4WePm1gghhOjMJLxdpGfcqRGfxwrlem8hhBDtR8LbRSKC9fjr7dcbHpc5zoUQQrQjCW8XUSgUjkvGZKY1IYQQ7UnC24Uap0k9XlSJtWsO4hdCCNEBJLxdqDG8a40Wistq3dwaIYQQnZWEtwt1P+12oMfkvLcQQoh2IuHtQpEhenQaFSDnvYUQQrQfmR7VhZQKBQmR/vx+opzjcrmYEKKLW7p0Ef/5z5Zm17/wwqsMGdL2GSltNhv//vd7TJp0Y7P7BZg/f1GbX9tbSHi7WEJ0AL+fKOdYYSU2mw2FQuHuJgkhhFs88MAc7rlnFgBffvk57777NmvWrHesDww8tzuiZWTs4fnnn202vLsCCW8XS4yyD1qrrKmntNJIaKDOzS0SQgj38Pf3x9/f3/FYqVQSFhZ+3q/bRW/J4UTC28Uawxvsk7VIeAshxNkVFhbw/PPPsnv3LkJCQrn22j9z++3TUKuV1NfX8/TTS9i+/StMJhNDhqQyZ848zGYzs2ffA8CIEann1PVeVFTIiy+uZPfuXSiVCsaOHcd99z2ARqPBbDazYsUzfPPNf532GxERSWVlJc888yTp6T8BCi69dASPPPIofn7+7VCdPybh7WIx4b74qJXUm60cK6xkcJ/z/5YphBBnU2uupaDa0KH7jPaLQK/Wn/fr2Gw25s+fS+/efVi3bgPFxcU899zTKJVK7rprOhs2bGDv3nSef/4ldDody5cv44UXnmfRoqUsXZrG/Plz+fDDbW3ueq+vr2f27HuJj49n1arVlJWV8uyzTwEKHnxwDu+/v5G9e/c02e+SJc/wxhuvUVJykpdffgOLxcyTT/4/1q9/g/vue+C869FWEt4uplIq6Rbhx5H8Shm0JoRoN7XmWv7f989Qa+7YOSX0aj1LLn3svAM8Pf0nCgryWb36TZRKJQkJ3Zk580Gefnoxd901nRMnTqDVaomJiSEwMIj58xdRXl6OSqUiIMB+We65dMHv3Pk9xcVFrF79JoGB9td5+OFHefTRh5g+/T7y8/PPul+AgoI89HpfYmPj0Ol0PPVUmtu68CW820FiVABH8ivlBiVCCNGMY8eOUFFRzjXXjHQss1qtGI1GysvL+Nvf/saWLVuYMOEakpNTuOKK0Vx77XXnvd+jR48QH5/gCG6AQYMuwmKxkJubw4QJf+GLLz49635vvHEKjz32CNdddxWpqcMYNepKxo4dd95tOhdtDm+j0cjixYv57LPP0Ol0TJ06lalTp5512/3797Nw4UJ+//13evfuzeLFi7nwwgsd67ds2cLf//53DAYDI0aMYMmSJYSGhgL2LpUVK1bw3nvvYbVamTx5MnPmzEGptF+a/uabb7Js2TKn/U2dOpVHH320rW/J5RIaZlorqTBSWWMiwFfj5hYJITqbxiNgb+02t1gsJCR055lnVjRZ5+fnT0REHP/+9xa++eZbvv/+W157bRWff76Nl15ac1771Wi0Z2mL1fGzT59evPfex3z//XdN9puSMpTNmz/hu++28/3335GW9jS7dv3IggVLzqtN56LN4Z2Wlsa+fftYv349eXl5PProo8TGxjJunPO3j5qaGqZPn86f//xnnnnmGd555x1mzJjB559/jq+vL1lZWcyfP5/FixfTv39/li5dyrx583jttdcAWLduHVu2bGHVqlWYzWb+93//l7CwMKZNmwbAwYMHufnmm7nvvvsc+9Trz/8D5QpnDlob2CPUja0RQnRWerWeHkEJ7m7GOYmPT6SwsIDg4BDHiPSffvqRrVu3sGjREj744ANMJitjxoxlzJir2LfvZ+65505KS0vO6xLchIREcnKOU1FR7jhf/ssvWahUKuLiuvGf/2xBo9Fw5ZVXN9nv559vo1evPowffx3jx1/HF198ytNPP+mW8G7TDGs1NTVs2rSJ+fPnM3DgQMaOHctdd93Fhg0bmmy7detWtFotc+fOpVevXsyfPx8/Pz+2bdsGwNtvv8348eOZOHEi/fv3Jy0tje3bt5OTkwPAW2+9xezZs0lNTWX48OHMmTPHaT+HDh2if//+REREOP41fgDcrVuEH8qGD5d0nQshRFPDhg0nOjqaJ5/8fxw6dJDMzL2kpT2NTqdDpVJRWVnJypXL2b17F3l5uXz++X+IjIwiKCjYcaCWnf0rRqPxrK9vMBTx44/fO/3LyTnO0KEXExsbx5IlCzh06CB79uxm5crnGDt2HAEBAVRXV/GPf6w4636LiopYuTKNfft+JifnOF9//SV9+/bryLI5tOnIOzs7G7PZTHJysmNZSkoKr776Klar1dGlDZCZmUlKSorjG5JCoWDIkCFkZGQwadIkMjMzufvuux3bx8TEEBsbS2ZmJhqNhvz8fIYOHeq0n9zcXIqKioiMjOTw4cN07979XN93u/JRq4gN9+OEoUqmSRVCiLNQqVQ888zz/P3vzzF9+u3o9b6MHn0Vs2bZR27fcsstHD2aw5IlC6isrKBfvwE888wKVCoVPXv2ZujQi7n33qksWrSUkSPHNHn93bt3sXv3Lqdlt902lenT7+OZZ55n5co0pk+/HV9fP66+ehzTp88EYNKkv1JUVHTW/d59971UV1fx2GMPU1tbw+DBKW456oY2hrfBYCAkJASN5tQ53PDwcIxGI2VlZY7z1Y3b9u7d2+n5YWFhHDhwAMARwmeuLygowGCwn8M5fX14uH1UYUFBAUqlkrKyMv79738zb948tFotkydPZurUqa3uTlEqFSiVrpn9TKVSOv0E6B4TwAlDFceLqlCrZQr5tjpbTcX5kZq2D6lr60yYcD0TJlzvtCwxMYGVK19ssq1KpUSpVPLAAw85wvx0arWOF198pdl9LVz4JAsXPtns+oSE+LPu107J7NkPMnv2g03W+Pv78sQTC5t93Y7UpvCura11Cm7A8bvJZGrVto3b1dXVNbu+rq7O6bXP3M/hw4cBe9i/8sor/Prrrzz11FOoVCruuOOOVr2X0FA/l09dGhh46pz7gJ5hfJeVT2FJDVq9Bl+dj0v31VWcXlPhGlLT9iF1dT2pafPaFN5arbZJSDf+rtPpWrVt43bNrdfr9U5BrdVqnfaj1+sZOHAgP/74IyEhIQD069ePkpIS3nnnnVaHd0lJtUuPvAMD9VRU1DpGLUaeNrNaZnYh/RNDXLKvruJsNRXnR2raPqSurteVaxoS4teq7doU3lFRUZSWlmI2m1Gr7U81GAzodDqna+Yaty0uLnZaVlxc7OgKb259REQEUVFRjtfu1q2b4zFAREREwxt0DsNevXpRWFjY6vditdqwWl17cb3FYsVstn/QYsN8UQA24EheBb3jzm0C/q7u9JoK15Catg+pq+tJTZvXppM0AwYMQK1Wk5GR4ViWnp7OoEGDnAarASQlJbF3717H7DM2m409e/aQlJTkWJ+enu7YPj8/n/z8fJKSkoiKiiI2NtZpfXp6OrGxsURGRrJp0yauueYap5ltfv31V3r27NmWt9Ou9Fo1kaG+gIw4F0II4VptCm+9Xs/EiRNZtGgRWVlZfPHFF6xdu5bbbrsNsB8dN56vHjduHBUVFSxdupSDBw+ydOlSamtrGT9+PABTpkzhww8/ZNOmTWRnZzN37lxGjRpFfHy8Y/3y5cvZuXMnO3fuZMWKFY79XHrppRgMBp599lmOHTvGJ598wpo1a7jrrrtcVhhXSIyyX7om4S2EEMKVFLY2TsxaW1vLokWL+Oyzz/D392fatGmO88z9+vVj2bJlTJo0CYCsrCwWLlzIoUOH6NevH4sXL+aCCy5wvNbmzZt54YUXKC8v57LLLmPJkiWO7nCLxUJaWhqbN29GpVIxefJkHnnkEccgs927d/Pcc8+RnZ1NWFgYd999N1OmTGn1+zAYXBeoarWSkBA/Skurnbp4/vPjMTZ9fQilQsHLD1+Bxkflsn12ds3VVJw7qWn7kLq6XleuaUREQMsbcQ7h3Vl0RHj/crSEFe9mAPDEban0jA1s5hXEmbryH297kZq2D6mr63XlmrY2vOXCxHZ0+jSp0nUuhBDCVSS825G/3oewQPulbnJ7UCGEEK4i4d3OEhqOvmWaVCGEEK4i4d3OGrvOTxiqMHexyQaEEEK0DwnvdtZ4b2+zxUb+yRo3t0YIIURnIOHdzpwGrUnXuRBCCBeQ8G5nwf4aAn3tNyWREedCCCFcQcK7nSkUCkfXuYw4F0II4QoS3h2gsev8eGEV1q45J44QQggXkvDuAI3hbay3UFgig9aEEEKcHwnvDtDYbQ72o28hhBDifEh4d4CIIB16rf3+5zJoTQghxPmS8O4ACoXCcXtQGbQmhBDifEl4d5DTp0ntojdyE0II4SIS3h0kseG8d3WdmZMVdW5ujRBCCG8m4d1BEqJk0JoQQgjXkPDuIDGhvmjU9nLLNKlCCCHOh4R3B1EqFcQ3DFqTEedCCCHOh4R3B0qIkmlShRBCnD8J7w7UONNaWZWJ8iqjm1sjhBDCW0l4dyCn24PKoDUhhBDnSMK7A8VF+KFSKgDpOhdCCHHuJLw7kFqlJC7CD5BBa0IIIc6dhHcHSzxtpjUhhBDiXEh4d7DGEefF5XVU19W7uTVCCCG8kYR3B0uU24MKIYQ4TxLeHSw+wh+FfcyadJ0LIYQ4JxLeHUyrUREd6gvA8SIJbyGEEG0n4e0GjV3ncuQthBDiXEh4u0HjiPOCkzUYTRY3t0YIIYS3kfB2g8YR5zYgxyCD1oQQQrSNhLcbJDbcXQyk61wIIUTbSXi7ga/Oh/AgHSAzrQkhhGg7CW83aRy0JnOcCyGEaCsJbzdpHLSWa6jGbLG6uTVCCCG8iYS3mzQOWrNYbeQaqt3cGiGEEN5EwttNTp8mVc57CyGEaAsJbzcJ8tMQ7K8BJLyFEEK0jYS3GzV2ncugNSGEEG0h4e1GjYPWcgqrsFptbm6NEEIIbyHh7UaN571NZiv5JTVubo0QQghvIeHtRgmnzbQmXedCCCFaS8LbjcICdfjp1IBMkyqEEKL12hzeRqORxx9/nNTUVEaMGMHatWub3Xb//v3ceOONJCUlccMNN7Bv3z6n9Vu2bOGqq64iKSmJmTNnUlJS4lhns9lYvnw5w4cPZ9iwYaSlpWG1Np3MxGw2c/311/Piiy+29a24nUKhkJnWhBBCtFmbwzstLY19+/axfv16Fi5cyKpVq9i2bVuT7Wpqapg+fTqpqals3ryZ5ORkZsyYQU2N/dxuVlYW8+fPZ9asWWzcuJGKigrmzZvneP66devYsmULq1at4oUXXuDjjz9m3bp1Tfazdu1asrOz2/o2PEbjiPNjhVXYbDJoTQghRMvaFN41NTVs2rSJ+fPnM3DgQMaOHctdd93Fhg0bmmy7detWtFotc+fOpVevXsyfPx8/Pz9H0L/99tuMHz+eiRMn0r9/f9LS0ti+fTs5OTkAvPXWW8yePZvU1FSGDx/OnDlzmuzn2LFjvPXWW/Tu3ftc37/bNY44rzWaMZTXubk1QgghvEGbwjs7Oxuz2UxycrJjWUpKCpmZmU26tDMzM0lJSUGhUAD2LuIhQ4aQkZHhWJ+amurYPiYmhtjYWDIzMyksLCQ/P5+hQ4c67Sc3N5eioiLHsgULFnD//fcTGhralrfhUU6fae24nPcWQgjRCuq2bGwwGAgJCUGj0TiWhYeHYzQaKSsrcwpRg8HQ5Ig4LCyMAwcOAFBUVERkZGST9QUFBRgMBgCn9eHh4QAUFBQQGRnJ+++/j9Fo5K9//Stbtmxpy9sAQKlUoFQq2vy8s1GplE4/2yI2wg+dRkWdyUKOoYrhF0a7pE3e7nxqKs5Oato+pK6uJzVtWZvCu7a21im4AcfvJpOpVds2bldXV9fs+rq6OqfXPnM/J0+e5Pnnn2fdunWOI/u2Cg31O+fnNicwUH9Oz+sZF8T+IyXknqwhJMTPpW3ydudaU9E8qWn7kLq6ntS0eW0Kb61W2ySkG3/X6XSt2rZxu+bW6/V6p6DWarVO+9Hr9SxdupRJkybRt2/ftjTfSUlJtUuPvAMD9VRU1GI5h9t7dgv3Y/+REg7mlFFSUuXyLxXe6HxrKpqSmrYPqavrdeWatvYArk3hHRUVRWlpKWazGbXa/lSDwYBOpyMwMLDJtsXFxU7LiouLHV3hza2PiIggKirK8drdunVzPAaIiIjgk08+QafT8fbbbwP2o/i9e/eybds2Pvnkk1a9F6vV5vIpSS0WK2bzOYR3hH2ylopqE8VldYQEaF3aLm92rjUVzZOatg+pq+tJTZvXphMKAwYMQK1WOwadAaSnpzNo0CCUSueXSkpKYu/evY7Ln2w2G3v27CEpKcmxPj093bF9fn4++fn5JCUlERUVRWxsrNP69PR0YmNjiYyM5LPPPuOjjz7igw8+4IMPPuDCCy/kpptuYvXq1W0ugCeQ24MKIYRoizaFt16vZ+LEiSxatIisrCy++OIL1q5dy2233QbYj44bz1ePGzeOiooKli5dysGDB1m6dCm1tbWMHz8egClTpvDhhx+yadMmsrOzmTt3LqNGjSI+Pt6xfvny5ezcuZOdO3eyYsUKx34SExOd/ul0OoKCgoiLi3NZYTpSTJgv6oaBGTLiXAghREvaPJRv3rx5DBw4kNtvv53Fixdz//33c/XVVwMwYsQItm7dCoC/vz+vvfYa6enpTJo0iczMTFavXo2vry8AycnJPPnkk7z00ktMmTKFoKAgli1b5tjPtGnTuPbaa5k1axYPPPAA119/PXfccYcL3rLnUauUdIuwn+eQI28hhBAtUdi66LReBoPrQlKtVhIS4kdpafU5n59Zvy2b7Rl5hAVqee6+y1zWNm/lipoKZ1LT9iF1db2uXNOIiICWN0JuTOIxGmdaO1lhpKq23s2tEUII4ckkvD1E4xznIF3nQggh/piEt4foFuGHsuH6bhm0JoQQ4o9IeHsIjY+K2HD7YD458hZCCPFHJLw9yOm3BxVCCCGaI+HtQRoHrRWW1FBrNLu5NUIIITyVhLcHOX2mtZwiOfoWQghxdhLeHiQ+0t/xWM57CyGEaI6EtwfRa9VEhdhvgScjzoUQQjRHwtvDyKA1IYQQLZHw9jCN573ziqupN1vc3BohhBCeSMLbwzSOOLfabJwwVLu5NUIIITyRhLeHSYiSQWtCCCH+mIS3hwnw1RAaqAVk0JoQQoizk/D2QImOQWsS3kIIIZqS8PZAjSPOc4qqMVu61r1shRBCtEzC2wM1HnmbLVYKTta4uTVCCCE8jYS3Bzp9mlTpOhdCCHEmCW8PFOyvIcDXB5DwFkII0ZSEtwdSKBSOrnMZcS6EEOJMEt4eqrHr/FhRFVabzc2tEUII4UkkvD1U44hzo8mCobTWza0RQgjhSSS8PVSizLQmhBCiGRLeHioiWI9eqwbgmJz3FkIIcRoJbw+lUChIiLQffR+XI28hhBCnkfD2YI5Ba4VV2GTQmhBCiAYS3h6s8XKxqtp6SiqMbm6NEEIITyHh7cFOvz2odJ0LIYRoJOHtwaLDfNGo7f+JZMS5EEKIRhLeHkylVBLvGLRW5ebWCCGE8BQS3h4uQe7tLYQQ4gwS3h6uccR5aaWRimqTm1sjhBDCE0h4e7jGEecgg9aEEELYSXh7uNhwP1RKBSBd50IIIewkvD2cj1pJXLgfINOkCiGEsJPw9gIJDee9ZcS5EEIIkPD2Co3nvYvKaqmpq3dza4QQQribhLcXcB60JkffQgjR1Ul4e4FukX4oGh7LiHMhhBAS3l5Ap1ETHeYLyIhzIYQQEt5eIzHq1O1BhRBCdG0S3l6icZrU/JPVGOstbm6NEEIId5Lw9hKJDbcHtdngRJEcfQshRFfW5vA2Go08/vjjpKamMmLECNauXdvstvv37+fGG28kKSmJG264gX379jmt37JlC1dddRVJSUnMnDmTkpISxzqbzcby5csZPnw4w4YNIy0tDavV6lj/888/c9NNN5GUlMQ111zDBx980Na34lUar/UGOe8thBBdXZvDOy0tjX379rF+/XoWLlzIqlWr2LZtW5PtampqmD59OqmpqWzevJnk5GRmzJhBTU0NAFlZWcyfP59Zs2axceNGKioqmDdvnuP569atY8uWLaxatYoXXniBjz/+mHXr1gFQWVnJ3XffTXJyMlu2bGHmzJk88cQTpKenn2sdPJ6fzofwIB0gI86FEKKra1N419TUsGnTJubPn8/AgQMZO3Ysd911Fxs2bGiy7datW9FqtcydO5devXoxf/58/Pz8HEH/9ttvM378eCZOnEj//v1JS0tj+/bt5OTkAPDWW28xe/ZsUlNTGT58OHPmzHHsJz8/nyuuuIK5c+cSHx/PhAkT6NOnD3v27Dnfeng0x6C1Auk2F0KIrqxN4Z2dnY3ZbCY5OdmxLCUlhczMTKcubYDMzExSUlJQKOxXKCsUCoYMGUJGRoZjfWpqqmP7mJgYYmNjyczMpLCwkPz8fIYOHeq0n9zcXIqKiujbty9paWkoFAqsVitfffUVR44ccdq+M2rsOj9hqMJssbawtRBCiM5K3ZaNDQYDISEhaDQax7Lw8HCMRiNlZWWEhoY6bdu7d2+n54eFhXHgwAEAioqKiIyMbLK+oKAAg8EA4LQ+PDwcgIKCAsdyk8nEkCFDqK+v56abbmLw4MGtfi9KpQKlUtHyhq2gUimdfraXnrGBAFisNgpLax33+u6MOqqmXYnUtH1IXV1PatqyNoV3bW2tU3ADjt9NJlOrtm3crq6urtn1dXV1Tq/9R/vZuHEjhw8f5sknn6R79+7ceeedrXovoaF+jl4BVwkM1Lv09c6U1E/leGyoMDJ4QHS77s8TtHdNuyKpafuQurqe1LR5bQpvrVbbJDwbf9fpdK3atnG75tbr9XqnoNZqtU770etP/cfUaDQMHDiQgQMHUlRUxD//+c9Wh3dJSbVLj7wDA/VUVNRiaefu7GB/DWVVJvYfLia1b3i77sudOrKmXYXUtH1IXV2vK9c0JMSvVdu1KbyjoqIoLS3FbDajVtufajAY0Ol0BAYGNtm2uLjYaVlxcbGjy7u59REREURFRTleu1u3bo7HABEREeTk5HD06FEuv/xyx3N79+5NaWlpq9+L1WrDarW1evvWsFismM3t+0FLiAqgrOokR/Mr231fnqAjatrVSE3bh9TV9aSmzWvTCYUBAwagVqsdg84A0tPTGTRoEEql80slJSWxd+9ebDZ7QNpsNvbs2UNSUpJj/emXduXn55Ofn09SUhJRUVHExsY6rU9PTyc2NpbIyEiysrJ46KGHHN3rAPv27aNnz55teTteqXGmteNFlS7/8iGEEMI7tCm89Xo9EydOZNGiRWRlZfHFF1+wdu1abrvtNsB+dNwYqOPGjaOiooKlS5dy8OBBli5dSm1tLePHjwdgypQpfPjhh2zatIns7Gzmzp3LqFGjiI+Pd6xfvnw5O3fuZOfOnaxYscKxn1GjRhEQEMCCBQs4cuQIH3/8Ma+//jr33nuvywrjqRovFzPVWyksrXFza4QQQrhDm4fyzZs3j4EDB3L77bezePFi7r//fq6++moARowYwdatWwHw9/fntddeIz09nUmTJpGZmcnq1avx9bXfHSs5OZknn3ySl156iSlTphAUFMSyZcsc+5k2bRrXXnsts2bN4oEHHuD666/njjvuAMDPz4/XX3+doqIiJk2axMqVK3n88ce56qqrzrceHq9xmlSAYwUyWYsQQnRFCltjv3YXYzC4LvjUaiUhIX6Ulla3+/kZm83G7H98S3WdmWuGxfO3MX3adX/u0pE17Sqkpu1D6up6XbmmERGtuwRYLqLzMgqF4tR5b7k9qBBCdEkS3l7o1DSplXTRjhMhhOjSJLy9UEK0/bx3jdFMcXldC1sLIYTobCS8vVDjkTfIHcaEEKIrkvD2QlGhvmh97FOlyr29hRCi65Hw9kJKhYL4hkvG5PagQgjR9Uh4e6lEx4hzOfIWQoiuRsLbSyU0HHmXV5soqzK6uTVCCCE6koS3lzp90JrMtCaEEF2LhLeXig33Q62y39JUus6FEKJrkfD2UmqVkriIhkFrMtOaEEJ0KRLeXuz0mdaEEEJ0HRLeXqzxDmMnK+qoqq13c2uEEEJ0FAlvL5YQLTOtCSFEVyTh7cXiI/xRKuyD1mSmNSGE6DokvL2YxkdFTJgvILcHFUKIrkTC28slyKA1IYTociS8vVxiw3nvwpIa6kxmN7dGCCFER5Dw9nKNI85tQE6RdJ0LIURXIOHt5eIjZZpUIYToaiS8vZyvTk1kiB6QEedCCNFVSHh3AgmO24NKt7kQQnQFEt6dQON577ziaurNVje3RgghRHuT8O4EGkecW6w2covl6FsIITo7Ce9OIEHu7S2EEF2KhHcnEOirISRAC8jtQYUQoiuQ8O4kEh2D1uTIWwghOjsJ704ioWHQWk5RFRarDFoTQojOTMK7k2gctFZvtpJ/ssbNrRFCCNGeJLw7icQoube3EEJ0FRLenURIgBZ/vQ8Axwpk0JoQQnRmEt6dhEKhcHSdyzSpQgjRuUl4nyeTpZ4sw35q6mvd3RRH13lOUSVWm83NrRFCCNFe1O5ugLf79NhXbDv6JTEHInko5R78VP5ua0vjiPNaowVDWS1RIb5ua4sQQoj2I0fe5ynGNxKA/Koi/p6+mqr6are1pbHbHGSmNSGE6MwkvM9TanQy1/a8CoC8qgJeynidWrN7utAjgvXotSpA7jAmhBCdmYS3C0zodQ3X9h0DwPHKXF7OXIfRYurwdigVCuIjZdCaEEJ0dhLeLqBQKLh98GQu7zYcgMPlR3kt603qLfUd3pbGQWvHCiqxyaA1IYTolCS8XUShUHDzgEkMjRoCwG+lB3l93z8xW80d2o7EaPugtaraekorjR26byGEEB1DwtuFlAoltw64kcERFwKw72Q2b+5/F4vV0mFtcLo9qHSdCyFEpyTh7WIqpYo7B97MBWH9ANhblMWG7Pew2jrmZiExYb74qO3/WWXEuRBCdE4S3u1ArVRz94W30Se4JwA7C9L51+8fdsg5aJVSSXykvetcRpwLIUTn1ObwNhqNPP7446SmpjJixAjWrl3b7Lb79+/nxhtvJCkpiRtuuIF9+/Y5rd+yZQtXXXUVSUlJzJw5k5KSEsc6m83G8uXLGT58OMOGDSMtLQ3rabe6PHToEFOnTmXIkCGMGTOGV1991Wm9u2lUPtxz0R30CEwA4NvcH/j3oU86JMAbu86l21wIITqnNod3Wloa+/btY/369SxcuJBVq1axbdu2JtvV1NQwffp0UlNT2bx5M8nJycyYMYOaGvvtKrOyspg/fz6zZs1i48aNVFRUMG/ePMfz161bx5YtW1i1ahUvvPACH3/8MevWrQOgtraW6dOnExUVxXvvvcfChQtZv34977zzzrnWoV3o1DruS5pGN/9YAL48/g1bj37R7vtNbJhprbTSSEVNx1+yJoQQon21KbxramrYtGkT8+fPZ+DAgYwdO5a77rqLDRs2NNl269ataLVa5s6dS69evZg/fz5+fn6OoH/77bcZP348EydOpH///qSlpbF9+3ZycnIAeOutt5g9ezapqakMHz6cOXPmOPbz008/UV5ezuLFi+nZsycjR47kjjvu4OOPPz7fericr4+eWYPvItovCoCtRz7ni+Pb23WfCXJ7UCGE6NTaFN7Z2dmYzWaSk5Mdy1JSUsjMzGzSZZ2ZmUlKSgoKhQKwX0o1ZMgQMjIyHOtTU1Md28fExBAbG0tmZiaFhYXk5+czdOhQp/3k5uZSVFTEgAEDeOmll9BoNE77rKryzHO8ARp/Zg++m3B9GAD/PvgJ35z4vt321y3CD5XSXncZtCaEEJ1Pm8LbYDAQEhLiFJrh4eEYjUbKysqabBsZGem0LCwsjIKCAgCKioqaXW8wGACc1oeHhwNQUFBAREQEF198sWNdXV0d//rXvxg+fHhb3k6HCtIGMnvwdEK0wQBs/P0Dfszf3S778lGriA33A2TQmhBCdEZtuqtYbW1tk6Pdxt9NJlOrtm3crq6urtn1dXV1Tq/9R/uxWq089thjVFdXM2PGjFa/F6VSgbLh6PR8qVRKp5/NiQoI46HUGSz/6WUqTJW8/esmdD4aUqMHu6Qdp+seHUBOURXHCytRq73vooLW1lS0ntS0fUhdXU9q2rI2hbdWq20Sno2/63S6Vm3buF1z6/V6vVNQa7Vap/3o9XrH9mazmUcffZSvv/6atWvXEhER0er3Ehrq5+jSd5XAQH2L24SE+LEw4EEWffU8laZq1v78f4QEBpAad5FL2zKgZzjfZuVTWFqLRqfBT+/j0tfvKK2pqWgbqWn7kLq6ntS0eW0K76ioKEpLSzGbzajV9qcaDAZ0Oh2BgYFNti0uLnZaVlxc7OgKb259REQEUVFRjtfu1q2b4zHgCOj6+noeeughduzYwerVqxkyZEhb3golJdUuPfIODNRTUVGLxdLy5Wr+BDF7yN2s2P0qdeY6nv9+DbOSpzIgrK9L2gMQGaR1PM76rZD+iSEue+2O0NaaipZJTduH1NX1unJNQ0L8WrVdm8J7wIABqNVqMjIyHIPN0tPTGTRoEEqlc/dGUlISa9aswWazoVAosNls7Nmzh3vuucexPj09nUmTJgGQn59Pfn4+SUlJREVFERsbS3p6uiO809PTiY2NdYT/ggUL2LFjB2vWrHEa+NZaVqsNq9W111xbLFbM5tZ90GJ9Y5mZNJUXM17HZDHx8t51zBx8F72De7ikLbFhvigAG3A4r4LecUEued2O1paaitaRmrYPqavrSU2b16YTCnq9nokTJ7Jo0SKysrL44osvWLt2LbfddhtgPzpuPF89btw4KioqWLp0KQcPHmTp0qXU1tYyfvx4AKZMmcKHH37Ipk2byM7OZu7cuYwaNYr4+HjH+uXLl7Nz50527tzJihUrHPvZsWMHmzdv5rHHHiMxMRGDwYDBYHCa5MUb9Azqzj2D7kCtVGOy1vNK5lqOVeS45LV1GjVRob6AjDgXQojORmFr45RftbW1LFq0iM8++wx/f3+mTZvGHXfcAUC/fv1YtmyZ42g6KyuLhQsXcujQIfr168fixYu54IILHK+1efNmXnjhBcrLy7nssstYsmQJISH27l2LxUJaWhqbN29GpVIxefJkHnnkERQKBQsWLGDjxo1N2hYXF8dXX33VqvdhMLgu0NRqJSEhfpSWVp/Tt8R9xb+y+ue3sNgs+Kl9eWDIDOL8Y867Xa999As79xcSF+HHkmkXt/wED3K+NRVNSU3bh9TV9bpyTSMiAlreiHMI787Ck8IbYE9RFmv3bcCGjQCNPw8NuZco39YPwDub/+w8xqb/HkKhgJcfHonWR3Ver9eRuvIfb3uRmrYPqavrdeWatja8ZRy+hxgSeRG3DvgrChRUmqp4Ye9qTtae32mAxIaZ1mw2OGGQ672FEKKzkPD2IBfHpPC3fn8BoMxYzgt7V1NmLD/n13OeJlXCWwghOgsJbw9zedxwJvW+DoDiuhJe2LuGStO5Ba+/3oewQPt19TJoTQghOg8Jbw90ZcIVXNfjagAKa4p4MWMN1fU15/RaidFye1AhhOhsJLw91LjuVzI2YRQAuVX5vJT5BnXmuja/TuPtQXMNVZi72GQHQgjRWUl4eyiFQsH1vcYzstulAByryOGVrHWYLG27P3fjeW+zxUZecbXL2ymEEKLjSXh7MIVCweQ+ExgeY59B7mDZEVb//Bb1VnOrX6Ox2xyk61wIIToLCW8Pp1QouaX/ZFIikwD4teR31u7bgMVqadXzg/21BPnZb/QiI86FEKJzkPD2AkqFktsvuIlB4fbZ6bKKf+GtXzditbXuHHZj17kceQshROcg4e0lVEoV0wbeQv+QPgDsLszgnez3WxXgidH2QWs5hVUuvxmLEEKIjifh7UV8VD7MuOh2egXZ7zz2ff5PvH/gY1qa4bZxpjVjvYXC0nO75EwIIYTnkPD2MhqVhnuT7iQxwH73ta9P7OCjw9v+8Dmnz7QmXedCCOH9JLy9kF6tY+bgacT6RQPw2bH/su3ol81uHx6kw1drv3X78QIZtCaEEN5OwttL+fn4cn/y3Y47j318+FO+yvn2rNsqFAqZaU0IIToRCW8vFqgJ4P7BdxOmCwXg/QMfsyN351m3TWiYae14YWWL58iFEEJ4NglvLxeiC2Z28nSCtUEAvPPbZnYV7GmyXeOgteo6Mycr2j7NqhBCCM8h4d0JhOtDuX/w3fj7+GHDxj9//RcZRT87bXP6TGvfZOZjNLVukhchhBCeR8K7k4j2i2R28nR81XqsNitrf/k/fjmZ7VgfFeKLvmHQ2pbvj/LwS9/xz89+47icAxdCCK8j4d2JxPnHMGvwXehUWiw2C2t+fovfSw8BoFQqmPanAcSE+QJQa7Tw3z25LFr3E0vW7+bbzDw5GhdCCC+hsHXR0UsGg+uOONVqJSEhfpSWVmM2u/+2mwfLjrAq43XqrfVoVBruH3w3PYMSAbDZbBw4Uc72jFx+yjY43SZUp1FxycBoRg6Odbo23B08raadgdS0fUhdXa8r1zQionX/75XwdgFP/KD9WvI7r2auw2yzoFfreCB5BvEBcU7bVNXW88O+Ar7OyCX/pPPMaz1iAhg5OI5hAyLRadQd2XTAM2vq7aSm7UPq6npduaYS3i3o7OENkGX4hTX7/onVZsXfx48Hh9xDjF9Uk+1OHY3n8VN2UZOj8eEDoxmZFOs06K29eWpNvZnUtH1IXV2vK9dUwrsFXSG8wX4Dkzd/eQcbNoI0ATw45F4ifcOb3b7xaHx7Zh55xdVO6zryaNyTa+qtpKbtQ+rqel25phLeLegq4Q3wQ95PvJ29CYBgbRADw/oRpAkkSHvaP00QARo/lAr7GEabzcbB3FNH4/Xmjj0a9/SaeiOpafuQurpeV65pa8O7409mig53SexQjFYTm37/kDJjOTvydp11O6VCSaAmwCnY4wYG0muQL7n5Fvb9XkNRkZU6kw9f783l6725dI8OYFSy+86NCyFEVyT/t+0iRnW7DF+1nh/yfqLcVEG5sYI6i9FpG6vNSpmxnDJjOZytY6Ib6LuBwqbEatJiNWnJq9ey4Rct7/ysp3dkJKk9E+gbE02QJhC9WodCoeiYNyiEEF2IhHcXMix6CMOihzh+rzPXUW6qpNxoD/PGUC83VlB22u/11nqn17EprCi0tai0tU7LD/Mbh48Dx+2/+yh9GrrkAwnSBpz2OJDg0x7r1Lr2futCCNGpSHh3YTq1Dp1a57gz2dnYbDZqzXVOwV5urKDMVEFFw+PimjIq6yuxKZzPTdVb6ymuPUlx7ck/bIdWpXEK9iBtIKH6YAbW9yZaHeOS9yqEEJ2JhLf4QwqFAl8fPb4++rNeZtbIZrNhqKrg2/2H2XXoOCU15Sh8jCg0RhQ+dWj9zPjoTRitNVhxDnmjxURRTTFFNcVOyzf9BgkBcVyVMJLBEYNQKVXt8h6FEMLbyGhzF+jKIyPPxmazcSivwj6L269FmE6riVajJOWCIAb188MvwOI4ij+z677SVIWNUx/NMF0oVyZcwSUxqWhUGne8La8nn9P2IXV1va5cU7lUrAUS3h2jpq6eH34p5OuMXHINzteNJ0YFMDI5losHRDlumtLIqjCTUZrFh79+RnFtiWO5n48vI+MuZWS3y/DX+HXIe+gs5HPaPqSurteVayrh3QIJ7471h0fjPiouviCKkYNj6RETCJyqafHJCnbnZ/HF8a85XpnreI6P0odLYoZyZcLlhOvDOvz9eCP5nLYPqavrdeWaSni3QMLbfRqPxrdn5HLibEfjg2O57KIYYqODHDW12Wz8XnqIz49/za8lvzu2V6AgOXIQYxNGkRDYraPfileRz2n7kLq6XleuqYR3CyS83c9ms3E4r4LtGXns+rWwydH4JRfF0DsmkD7dgogI1jvWnajM44vj35BelIHVduo5fUN6MzZhJANC+8r15Wchn9P2IXV1va5cUwnvFkh4e5Y/OhoHCAvU0T8hmH4JIfRPCCY8WM/J2lL+e+JbduTtwmQxObaN84/hqoSRpEQmyQj108jntH1IXV2vK9dUwrsFEt6eyWazcTi/gu+y8vnlaCnFZbVn3S48SEe/hGD6J4QQH6thf2UGX+fsoLK+yrFNiDaYMQmXc2nMMHRqbUe9BY8ln9P2IXV1va5cUwnvFkh4eza1WklwsC+/Hynml8MlZB8vI/t4KaWVxrNuHx6ko29CAD4ReRyqz+Ck8dTEML5qPVfEXcLI+MsI1HTcbU09jXxO24fU1fW6ck0lvFsg4e3ZzlZTm82GobyO7GOl/Ha8lOzjZc2EuY3guFLUMUeoVhpOvaZSzcXRKVyZcMUfzirXWcnntH1IXV2vK9dU7iomOh2FQkFksJ7IYD1XJMXaw7ys1nFU/psjzBWU5YZCbghKf3uIq0IMmK1mduTt5Pu8XSRFDOSqhFH0CEpw99sSQog2k/AWXkuhUBAZ4ktkiK8jzIvKavmtIcyzj5VSVhWK6UAoCn0l6uijqMLyQGkjw7CPDMM+In3iuDpxFBfHD3Lcy1wIITyddJu7QFfu4mkvrqipzWajqLTWcVT+6/FSyo0VqKOPoo7MQaGyOLZVGgNIVCYxPC6FgYnhhAR0vgFu8jltH1JX1+vKNZVuc9HlKRQKokJ9iQr1ZeTgOGw2G4UNYb7/eBHZVRmYQw+j0Bixais5wnccPvET5t3dCa3vQ/9ukY7L0zpjmAshvJeEt+gyFAoF0aG+RIf6MmpwHDbbYHJPVvL5wR/5ufInjKpyFBojPgm/UWE+xPdF8XyzPxHqdUSF+jYEeTD94iXMhRDu1eaTfEajkccff5zU1FRGjBjB2rVrm912//793HjjjSQlJXHDDTewb98+p/VbtmzhqquuIikpiZkzZ1JScuoGFDabjeXLlzN8+HCGDRtGWloaVmvT7pOysjIuvfRSTpw40da3Iro4hUJBt/BA7hx+NcuvmseMQbcT72cfwKZQm/GJPYIuaTs+PX6mqKaI7Rl5rP5oP4+8tIN5q39k/bZsdu4vpKzq7JevCSFEe2nzkXdaWhr79u1j/fr15OXl8eijjxIbG8u4ceOctqupqWH69On8+c9/5plnnuGdd95hxowZfP755/j6+pKVlcX8+fNZvHgx/fv3Z+nSpcybN4/XXnsNgHXr1rFlyxZWrVqF2Wzmf//3fwkLC2PatGmOfZSXl3PPPfdw8uRJhDgfSoWSiyIGclHEQA6XH+WLY9vJKt4PShvqiFzUEbkoKqKpO5GItSqEwpIaCktq2J6RB0BkiJ4+3YLo0y2Y3nFBxIT5yhStQoh206bwrqmpYdOmTaxZs4aBAwcycOBADhw4wIYNG5qE99atW9FqtcydOxeFQsH8+fP55ptv2LZtG5MmTeLtt99m/PjxTJw4EbB/KRg9ejQ5OTnEx8fz1ltvMXv2bFJTUwGYM2cO//jHPxzhvXv3bh599FH8/OS2kMK1egZ1Z/pF3SmsLuLLnG/YmZ+O2WbBFliA9oICQlXR+Ff2J++wP5U1ZgCKSmspKq1lx88FAPjrfegdF0TvbkH06RZE9+gAfNQyVasQwjXaFN7Z2dmYzWaSk5Mdy1JSUnj11VexWq0olad64TMzM0lJSXEcfSgUCoYMGUJGRgaTJk0iMzOTu+++27F9TEwMsbGxZGZmotFoyM/PZ+jQoU77yc3NpaioiMjISL777jtuuOEG/vSnP3H11VefcwGEaE6UXyQ395/Mn3pczdcndvBt7g/UmusosRRQ4ltA5NBwrg69BG1VPIdzqzlwopyTFXUAVNXWk3GwmIyDxQCoVQq6xwTSpyHQe8cFEeCrcefbE0J4sTaFt8FgICQkBI3m1P90wsPDMRqNlJWVERoa6rRt7969nZ4fFhbGgQMHABwhfOb6goICDAb7rFinrw8PDwegoKCAyMhIHnzwQQA51y3aXZA2kOt7jeeaxNHsyNvFVznfUmYsp6i2mK25HxOg8efSgcO4bsxgdNZgDuaWc+BEOQdOlJFTVIXNBmaLjYMnyjl4ohx22l83JsyXPt2C6B0XTJ9uQUSG6KWrXQjRKm0K79raWqfgBhy/m0ymVm3buF1dXV2z6+vq6pxe+4/2c66USgVKpWv+R6lSKZ1+ivPniTX1V/tyTc9RXNl9BLsLMvn06H/Jqyqg0lTFp8e+4tNjXxHjF8XQ6MGMu2Iwt/v1p9Zo5lBuOb/nlHHgRDmHcsupM9mvL88/WUP+yRq+ycwHINBPQ59uQfSND6ZvfDCJ0QGoXfj+PbGmnYHU1fWkpi1rU3hrtdom4dn4u06na9W2jds1t16v1zsFtVarddqPXq/HFUJD/Vx+lBMY6Jq2iVM8tabjw65g3AWXk1HwC9sObCerYD8Wm5X86kI+OvQpHx36lB4h8Vwan8ql/VO4PMU+it1isXIkv4Jfj5Tw69ES9h85ycly+5fVimoT6b8ZSP/N3vOkUSvpkxDCBT1CuaBHGP0TQ/B3QVe7p9bU20ldXU9q2rw2hXdUVBSlpaWYzWbUavtTDQYDOp2OwMDAJtsWFxc7LSsuLnZ0hTe3PiIigqioKMdrd+vWzfEYICLCNTeUKCmpdumRd2CgnoqKWiyWrjUbUHvxlpp21/XgnkE9qOpXzd6in/mpIIPfSw5hw8aR0hyOlOawIevf9AxKZGj0YFKikwjzC2TEhVGMuDAKm83GyfI6fj/RcHSeU8aJoipsgMls5ZfDJ/nl8EngAAogLsKPPg1H5n27BRMerGv1l1Bvqam3kbq6XleuaUhI6wZhtym8BwwYgFqtJiMjwzEKPD09nUGDBjkNVgNISkpizZo12Gw2FAoFNpuNPXv2cM899zjWp6enM2nSJADy8/PJz88nKSmJqKgoYmNjSU9Pd4R3eno6sbGxTc6Tnyur1YbV6tqZYS0Wa5ebyq+9eUtNdUo9l0QP45LoYZQbK9lryCK9MJPD5UcBOFx+jMPlx/jXbx/RJ7gnQ6KSSI4YhL/Gj2B/LcP6RzKsv/2zXVNXz6G8Cg6cKOfgiTIO51VgMluxAScM1ZwwVPPfPbkABPlr6NMtmD5xQfSJDyI+0h+V8o+7Gr2lpt5G6up6UtPmtSm89Xo9EydOZNGiRTz99NMUFRWxdu1ali1bBtiPjgMCAtDpdIwbN44VK1awdOlSbrrpJt59911qa2sZP348AFOmTOHWW29l8ODBDBo0iKVLlzJq1Cji4+Md65cvX050dDQAK1asYOrUqa5870K0iyBtAKO6XcaobpdRUlfKnqIs0gszOF6Ziw0bv5cd4veyQ/zr9w/oH9KHlKgkkiIGolfbuwh9dT4M6hnGoJ5hAJgtVnKKqjiQU8aBXPugt/Jq+2mk8ioTu7OL2J1dBIDWR0XP2ED7QLhuQfSKDUKvlYkUhehs2nxjktraWhYtWsRnn32Gv78/06ZN44477gCgX79+LFu2zHE0nZWVxcKFCzl06BD9+vVj8eLFXHDBBY7X2rx5My+88ALl5eVcdtllLFmyhJCQEAAsFgtpaWls3rwZlUrF5MmTeeSRR5p0EZ44cYIrr7ySL7/80nGU3hpyYxLP1hlrWlRTzJ6iTNILM8mrLnBap1aouCCsPylRSQwKvwCtqvlz2423QrWPaC/nYG45ecXVZ91WoYD4CH96dwuiX0IIQy+MRY0czbhCnbmO7/J2kl6YQYhfIEPCB3Nh6AXo1DJ17vnqjH//rdXaG5PIXcVcoCt/0NpLZ69pXlUB6UWZpBdmYKh1niFQo/RhUPgFpEQlcUFoP3xUPi2+XlVtPQcbjsoPnijjcH4l5mbOFcZH+pPUO5zkPuEkRgeglMvT2qTSVMXXJ3aw/cT31JprndZplD4kRQzi4ugh9AvtLbeZPUed/e//j0h4t0DC27N1lZrabDZyKnPZXZTBnsIsSo1lTut1Kh1JEQNJiUqif0gfVMrWzdJWb7ZyrLCSgw3Xmx84UU5VbX2T7YL9NQzuHc7gPhEMSAzBRy1h05ySulK+OP4N3+ftot56qpY9ghKpsdRQWGVw2j5IE0hq9GAujk4hzj+mo5vr1brK3//ZSHi3QMLbs3XFmlptVo5WHGd3YSZ7ijKpNFU5rfdT+zI4chCpUUn0Du7ZpqM6m81GcUUdRwqq+C4zl1+PlmI5Y8CmVqPiwh6hJPcJ56Je4fjrWz7i7wryqwv5/NjX/FS4F6vt1GfxwrD+jE0cTf/wXgQH+5J+dD/f5+5mT2EmNWcckcf5xzAsegipUYMJ1gZ19FvwOl3x77+RhHcLJLw9W1evqdVm5WDZYXYXZpJR9DPV5hqn9YGaAJIjLyI1KonugQmtCvLTa1pRZWLfkZNkHCgm89BJao1mp22VCgV9ugWR3CecwX3CiQzxden78wZHyo/z+bH/kln8i2OZAgUpUUlcnTjacTR95me13mrml5PZ7CrYw77iX7HYLE7P7x/ah2HRQ0iKuPAPxzZ0ZV3571/CuwUS3p5NanqKxWohu/QA6YWZZBr2UWdxvgVpiDaYIVEXkRo5mPiAuGav+26upmaLlQM5Zew9UMzeA8WO+dlPFxfux+CGIO8RE9hpz5PbbDaySw7w2bH/8nvZIcdytVLN8JhUxiaMJFwf5vScP/qsVtVXs6cwi10FezhSccxpnUalYXDEhQyLHkK/EDk/frqu/Pcv4d0CCW/PJjU9u3pLPb+U/Maewkyyivc7nXsFiNCHkRI1mJTIJGL9o53WtaamNpuNE4Zq9h4wsPdAMccKmv6dBPlpHAPeLuge0inulma1Wckw7OPzY//leGWuY7lOpeXyuEsYHT+CIG3gWZ/b2s9qUU0xPxXsYVfBHorrSpzWBWkCGRqdzLDoIXJ+nK799y/h3QIJb88mNW1ZndnIvpO/kl6Yyf6T2ZhP654FiPWLJiUqiSGRSUT6hp9TTUsq6sg8WMzeg8VkHyvFbHH+34XGR8mFPcIazpOHed2d0sxWM7sK9vD58a8pqjk146O/jx+j4y/nirhL8PX54yk621pXm83G4fJj7CpIJ70oq8mI9cbz40Ojkpv9wtDZdeW/fwnvFkh4ezapadvU1NeSWfwLewozyS494DSwCiAhII6hMclc2fcSVCbtOdW01mhm35ESMg4YyDp0kuo65/PkCgX0iQticJ8IkvuEExXquefJ68xGduTtdNwhrlGoLoSrEkZySUwqmlaejz6fz2q91cy+4l/ZVbCHX05my/nxBl3571/CuwUS3p5NanruqkzV7DX8THphBgfLjmDD+U881i+KviG96R/ah97BPdGrdc28UvPMFisHTpSTcaCYvQcMFJc3PU8eE+bL4D7hJPeJoGesZ5wnr6qv5uucHWw/scNpRHiMXxRXJ44mJTKp1ZfjNXLVZ9V+fjyz4fz4cad1WpWGwRGDGBY9hL4hvTr9+fGu/Pcv4d0CCW/PJjV1jTJjOXuLfia9MLPJgCkApUJJ98B4+oX0pl9IH3oEJaBWtm06VZvNRm5xNXsPFJNxwMCR/KZ/W4G+Pg3nySO4oHsIGp+OPU9eWlfGlznfsCN3J6bTr9EOTODqxNFcGD7gnAOxPT6rBVUGdhaks7twLyXGUqd1AT4BXBA0iAGBFxLqE4HZYsVitWG22LBYrVgsNswNPy1Wm31+cKsNi8WG1WbDR61E66NCo1ai8VGh8VGiUavsyxoea3zs63zUSrd86erKf/8S3i2Q8PZsUlPXK6sv42DVQfac+IXfSg5RZ2l6tKxR+tA7uCf9QnvTP6QPsf7RbQ610kqj/Tz5gWJ+PVbS9Dy5WsnAHqEM7hNOUq9wAv3arzu4sLqIz45/zU8Fe526pAeE9uWaxNH0Du75h3dlM1us1BrN1BjN9p915lO/19l/GustKFUqampNmOqtjgC1WE8L0TMC1hG0jctOC1iL1cqp/yvbUPqXoQrPRRVagELtfKrCWh2AuTgOS0kM1LfPtKxnhnxjsGsdy52/CNi/GKjQ+jSz/rTXaHy9M+9b35X//iW8WyDh7dmkpq53ek2NpnqOV+byW+kBsksOcKT8WJMBb2AfuNUvpLcjzMP0oW3aZ53JzL7DJWQcLCbzYHHT8+RAr25BJPe2X4YWE9a62yG25FhFDp8d+y+Zhl+cThv09O1Hf91QdJZQp1A+M5gbl5vqPeizp7CgDDagDs9DGWRAoTz1vmw2sJaHYzkZi6U0CqxNezaUCgVKJU2+THkClVLh9OVA66MiwE+Lv15NkJ+GEH8twf5agv01BAfYH3fWG+5IeLdAwtuzSU1d749qarKYOFR2lOzSA/xWepATlXlNzpUDhOtC6Rfah/6hfegb3At/TevD1mK1cvBEeUP3ejFFZbVNtokObTxPHk5CVAB1Z4RprdFCTV29/aexnto6i2NdtbGeCvKoCszG6ndqqlKbVYGlOA5zfg9sRtd8OQD7Fw+dVo2vVo1Oq0aBDZVSiUqlQKVUoFYpUSlPe6xSONarlQpUjetVCtSNy898TuN6x2MlaqUCk62WQzXZZFfuo6Auz6ldGqWGC0MHkhqVTJ/gnvio1KhUCkf3t9Vmo77eitFswVRvwVRvxWRu+Flvwej4/dQ6Y8M6k7nh52mPz7q9yYq1naNFq1ER7K8l5LRAbwz4kIBTj73tUkYJ7xZIeHs2qanrtaWmVaZqfi87RHaJPcyLz7h5CthHQ3fzj7GHeUgfegV3b/UIbZvNRl5xNRkN3euH8yrO6T01vBrKkCJ8Yg6j9D81ctxmUWEuisdc0B3qmw7K06iV6LVqfHVq+0+t/eeZy3y1avS6U+sbf+q0KpQKhds/q4U1BnYV7OGngj2crHM+Px6sDWJolP368TOv+29vZsuZ4X4q+I2OLwFnfHFoeGy2WDFZbBSV1FBWWUdZlanJdL6t5adTExygPXX0HqBpCH2tI/QD/XxQKT1jEKCEdwskvD2b1NT1zqemxbUl/FZ6gN9KDvJb6UGq6pveglStUNEjKJH+oX3oF9KbhIBurR65XVZ16jz5/qOlzd4RrZFSoUCnU6CJyMccdhCLz6m/Z7VNS6LqIvrqkwjS+TuH8WmhfOZ51nPlKZ9Vq83quH58T1EWtWbnMQ3x/rEMjR5Cr+DuRPtGefStS8+sqdVmo6qmnrIqY8M/E6WVDY8rG36vMlJZbTpLf1HLFAoI9Dst1P01DUHv3F0foPf5wzESriDh3QIJb88mNXU9V9XUarOSV1Xg6GI/WHrYaQR3I51KR9+QXvQL6U3/0N5E+Ua26n98RpOF/UdLKKs2odeq8NX6OB39KtUW0g3pfJXzrdNd2EK0wVyZcAWXxg7r0GuiPfGzWm+p5+eTv7KrIJ1fTv7W5Lp/sJ8CifGPJs4vmlj/aGL8oonyjWjzpXLt4VxrarZYqag2OYf7aYFf1rDszLEXrW6XSkGQ36mj92B/LSEBWpJdOF5DwrsFEt6eTWrqeu1VU7PVzJHy4/Yj89KDHK3IOWtYBGkCHUfl/UJ7t/nuWtX1NWw/sYOvT+yguv7UjVqifCMZmziKoVGD23yZmyt4+me10lRFepH9+vFjFTl/uK1KoSLKN4JY/2ji/GKI8Y8i1i+GUF1wux9xnq69a2qqt1BWfSrMTz96b1xWWmVs9YBFP52av88e4ZKudwnvFkh4ezapqet1VE1rzXUcLDvMbyUHyS49QH514Vm3i/aNpF+o/fryviE90avPPg1pmbGcr45/y3d5P2K0mBzLEwPiubr7aC4Kv8Ctk5Z402e10lRFXlUBedUF5FUVkF9tf3x6Xc9Gp9IS4xdNbEOYx/pHE+sX3aYBi23hCTW12WzUmSxnHMGfcURfaaKixkRSrzDunXihS77gSHi3QMLbs0lNXc9dNS03VvBb6UFHmJ8+HWkjBQoSA+PpH9KbfqF96BGUSGldKZ8f286ugnSny9j6h/Th6sTR9A3p1aFHg83x9s+q1WalpK6M/OoCchsDvaqAgpqis/agnC5QE0BsQ7d7489ov6jzPm3hTTW12Wwu/RxKeLdAwtuzSU1dzxNqarPZKKox8FvpQbJLD/J76cEmA6sAfJQ+mK1mx+VqChQkRVzI1YmjSAyM7+hm/yFPqGt7MFvNFNUUk1eVT151IXnV+eRVFXLyjDuinUmBgjB9KHF+0cQ0hHqcfzQR+vBWn0/31JrabDaMFhN1ljrqzEbqLHUYzSbiA2Lx9XHNXP6tDe/OeZW7EMIjKRQKovwiifKL5Ipul2K1WcmpzCW75ADZpQc5XH4Us9XsuNWpUqFkWPQQxiaMItov0s2t71rUSrX9iPqMS8zqzHXkVxc6ut7zqgvJq8p3XIFgw0Zx7UmKa0+SWfzLqddTqIjyi2xypB6ibf/z6VabFaPFSJ3ZSK25jjqLkbrTf5rrqD3rY6NTUNeZjWed/yBEG8ziSx7t0MF+Et5CCLdRKpQkBsaTGBjPNd3HYLLUc7j8KL+VHkStVHNpzFBCdMHubqY4jU6to0dQIj2CEp2WV5gqHefT86sKyK0uIL+6EFPD+XSzzUJuVT65Vflw2jAInUrXcC49mlj/GGL9okgIjiMEPyxWC9X1tY6grW0I16bha6S2MWSbbFvX4jn98xWg8e/wcRcS3kIIj6FR+dC/YQY34V0CNQEEhgY4/bezn08vdTqXnlddQGGNwXE+vc5Sx+HyYxwud75xjo/Kh3pL00sQ24MCBTq1Fp1K5/xTrUOvsv/UNf50rNehb3gc5RvR4eMvJLyFEEK0C6VCSbg+jHB9GEkRAx3L661mimoMTiPf86oLKDlthrjWBLdSoUR/WtDqzhK0+oagdVqm1jkFtUbl43W3WZXwFkII0aF8lGri/GOI849xWl7bcD69oLYQi8oE9Uo0Cm3TwG147KNUe8QVB+4g4S2EEMIj6NU6egYl0jesh0eONvck3tVPIIQQQggJbyGEEMLbSHgLIYQQXkbCWwghhPAyEt5CCCGEl5HwFkIIIbyMhLcQQgjhZSS8hRBCCC8j4S2EEEJ4GQlvIYQQwstIeAshhBBeRmGz2ZreWVwIIYQQHkuOvIUQQggvI+EthBBCeBkJbyGEEMLLSHgLIYQQXkbCWwghhPAyEt5CCCGEl5HwFkIIIbyMhLcQQgjhZSS8hRBCCC8j4X0ejEYjjz/+OKmpqYwYMYK1a9e6u0mdQmFhIbNnz2bYsGFcfvnlLFu2DKPR6O5mdRrTp0/nsccec3czvJ7JZGLx4sUMHTqUSy+9lOeffx6ZsPL85efnM2PGDIYMGcKYMWN488033d0kj6R2dwO8WVpaGvv27WP9+vXk5eXx6KOPEhsby7hx49zdNK9ls9mYPXs2gYGBbNiwgfLych5//HGUSiWPPvqou5vn9T755BO2b9/OX/7yF3c3xes99dRT7Ny5kzfeeIPq6moeeughYmNjuemmm9zdNK/24IMPEhsby+bNmzl48CBz5swhLi6OsWPHurtpHkWOvM9RTU0NmzZtYv78+QwcOJCxY8dy1113sWHDBnc3zasdPnyYjIwMli1bRp8+fUhNTWX27Nls2bLF3U3zemVlZaSlpTFo0CB3N8XrlZWV8f7777NkyRIuuugiLrnkEqZOnUpmZqa7m+bVysvLycjI4N5776V79+5cddVVXH755fzwww/ubprHkfA+R9nZ2ZjNZpKTkx3LUlJSyMzMxGq1urFl3i0iIoLXX3+d8PBwp+VVVVVualHn8eyzz3L99dfTu3dvdzfF66Wnp+Pv78+wYcMcy6ZPn86yZcvc2Crvp9Pp0Ov1bN68mfr6eg4fPsyePXsYMGCAu5vmcSS8z5HBYCAkJASNRuNYFh4ejtFopKyszH0N83KBgYFcfvnljt+tVitvv/02w4cPd2OrvN8PP/zA7t27ue+++9zdlE4hJyeHuLg4PvjgA8aNG8eVV17JSy+9JF/cz5NWq2XBggVs3LiRpKQkxo8fzxVXXMGNN97o7qZ5HDnnfY5qa2udghtw/G4ymdzRpE7pueeeY//+/bz33nvuborXMhqNLFy4kAULFqDT6dzdnE6hpqaGY8eO8e6777Js2TIMBgMLFixAr9czdepUdzfPqx06dIjRo0dz5513cuDAAZYsWcIll1zChAkT3N00jyLhfY60Wm2TkG78Xf4H6RrPPfcc69evZ+XKlfTt29fdzfFaq1at4sILL3Tq0RDnR61WU1VVxYoVK4iLiwMgLy+Pd955R8L7PPzwww+89957bN++HZ1Ox6BBgygsLOSVV16R8D6DhPc5ioqKorS0FLPZjFptL6PBYECn0xEYGOjm1nm/JUuW8M477/Dcc89xzTXXuLs5Xu2TTz6huLjYMT6j8Uvmp59+yt69e93ZNK8VERGBVqt1BDdAjx49yM/Pd2OrvN++fftITEx0OgC64IILePXVV93YKs8k4X2OBgwYgFqtJiMjg9TUVMA+iGXQoEEolTKU4HysWrWKd999l+eff14uu3OBf/7zn5jNZsfvy5cvB2DOnDnuapLXS0pKwmg0cuTIEXr06AHYr5Q4PcxF20VGRnLs2DFMJpPjNOThw4fp1q2bm1vmeSRlzpFer2fixIksWrSIrKwsvvjiC9auXcttt93m7qZ5tUOHDvHyyy9z9913k5KSgsFgcPwT5yYuLo7ExETHPz8/P/z8/EhMTHR307xWz549GTVqFPPmzSM7O5tvv/2W1atXM2XKFHc3zauNGTMGHx8fnnjiCY4cOcJXX33Fq6++yq233urupnkchU2mBDpntbW1LFq0iM8++wx/f3+mTZvGHXfc4e5mebXVq1ezYsWKs6777bffOrg1nVPj7GrPPPOMm1vi3SorK1myZAmff/45er2em2++mZkzZ6JQKNzdNK928OBBli5dSlZWFqGhodxyyy3cfvvtUtczSHgLIYQQXka6zYUQQggvI+EthBBCeBkJbyGEEMLLSHgLIYQQXkbCWwghhPAyEt5CCCGEl5HwFkIIIbyMhLcQQgjhZSS8hRBCCC8j4S2EEEJ4GQlvIYQQwstIeAshhBBe5v8DrfJpLezPMr0AAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1200x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# construct model\n",
    "training_model = tf.keras.Sequential([\n",
    "    tf.keras.layers.LSTM(64, return_sequences=True),\n",
    "    tf.keras.layers.LSTM(64),\n",
    "    tf.keras.layers.Dense(len(prices))    # output layer with the same number of columns as the input\n",
    "])\n",
    "\n",
    "# compile the model\n",
    "training_model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.002), loss='mean_squared_error')\n",
    "\n",
    "# train the model\n",
    "history = training_model.fit(X_train, y_train, epochs=10, batch_size=16, validation_data=(X_test, y_test))\n",
    "\n",
    "# plot the training and validation loss\n",
    "plt.figure(figsize=(12, 4))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(history.history['loss'], label='Train Loss')\n",
    "plt.plot(history.history['val_loss'], label='Test Loss')\n",
    "plt.legend()\n",
    "plt.title('Loss Evolution')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tokens:  {'WBTC': 0, 'WETH': 1, 'USDC': 2, 'USDT': 3, 'DAI': 4, 'LINK': 5, 'AAVE': 6, 'STETH': 7, 'WSTETH': 8, 'ETH': 9, 'FRAX': 10, 'RETH': 11, 'YFI': 12, 'MIM': 13, '3CRV': 14, 'ALCX': 15, 'MKR': 16, 'STMATIC': 17, 'WAVAX': 18, 'UNI': 19, 'COMP': 20, 'GNO': 21, 'COW': 22, 'ALUSD': 23, 'SAVAX': 24, 'WMATIC': 25, 'CVX': 26, 'WOO': 27, 'TUSD': 28, 'FRXETH': 29}\n"
     ]
    }
   ],
   "source": [
    "class InferenceModel(tf.keras.Model):\n",
    "    def __init__(self, training_model, latest_sequence_per_token, min_price: float, max_price: float):\n",
    "        super(InferenceModel, self).__init__()\n",
    "        self.training_model = training_model\n",
    "        self.min_price = min_price\n",
    "        self.max_price = max_price\n",
    "        self.latest_sequence_per_token = tf.constant(latest_sequence_per_token, dtype=tf.float32)  # convert to tensor\n",
    "\n",
    "    @tf.function\n",
    "    def call(self, inputs):\n",
    "        num_intervals = inputs[0]\n",
    "        num_intervals = tf.cast(num_intervals, dtype=tf.int32)  # convert to int32\n",
    "        token = inputs[1]\n",
    "        token = tf.cast(token, dtype=tf.int32)  # convert to int32\n",
    "        price = inputs[2]\n",
    "        price = tf.cast(price, dtype=tf.int32)\n",
    "        sequence = tf.gather(self.latest_sequence_per_token, token)    # get the latest sequence for the token\n",
    "        len_sequence = tf.constant(sequence.shape[0], dtype=tf.int32)  # get the length of the sequence\n",
    "\n",
    "        # Define the loop condition\n",
    "        def condition(i, sequence):\n",
    "            return i < num_intervals\n",
    "\n",
    "        # Define the loop body\n",
    "        def body(i, sequence):\n",
    "            inputs = sequence[-len_sequence:, :]        # get the last sequence\n",
    "            inputs = tf.expand_dims(inputs, axis=0)     # add the batch dimension\n",
    "            prediction = self.training_model(inputs)    # predict the closing price for the next interval\n",
    "            sequence = tf.concat([sequence, prediction], axis=0)            # append the prediction to the sequence\n",
    "            return i + 1, sequence\n",
    "\n",
    "        # Use tf.while_loop to construct the loop\n",
    "        _, final_sequence = tf.while_loop(condition, body, [0, sequence], shape_invariants=[tf.TensorShape([]), tf.TensorShape([None, sequence.shape[1]])])\n",
    "\n",
    "        # denormalize the result\n",
    "        final_sequence = final_sequence * (self.max_price - self.min_price) + self.min_price\n",
    "\n",
    "        return final_sequence[-num_intervals:, price]  # return the last value in the sequence\n",
    "    \n",
    "\n",
    "# create a dictionary with the latest sequence for each token\n",
    "latest_sequence_per_token = [data[data['token'] == token][prices].values[-len_sequence:] for token in data['token'].unique()]\n",
    "tokens = {element:idx for idx, element in enumerate(data['token'].unique())}\n",
    "print(\"Tokens: \", tokens)\n",
    "\n",
    "# construct and save the inference model\n",
    "inference_model = InferenceModel(training_model, latest_sequence_per_token, min_price, max_price)\n",
    "inference_model.build(input_shape=(3,))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(\n",
      "[2536.0488 2567.1807 2603.615  2642.2266 2682.7385 2725.132  2769.457\n",
      " 2815.7961 2864.2466 2914.9167 2967.9175 3023.366  3081.382  3142.0906\n",
      " 3205.6208 3272.1108 3341.7043 3414.5544 3490.8228 3570.68   3654.3088\n",
      " 3741.9014 3833.6626 3929.8118 4030.5781 4136.21   4246.969  4363.1333\n",
      " 4485.0005 4612.887  4747.132  4888.095  5036.1636 5191.7485 5355.291\n",
      " 5527.266  5708.178  5898.578 ], shape=(38,), dtype=float32)\n",
      "Predicted Close price for token 'BTC' on 2024-06-24: 5898.578125\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Run the model to predict the closing price on 2024-02-04 for token 'FRXETH'\n",
    "\n",
    "# get the number of intervals between the most recent date and the target date (interval = 4 days)\n",
    "target_date = '2024-06-24'\n",
    "num_intervals = (pd.Timestamp(target_date) - most_recent_date).days // 4\n",
    "token = tokens['FRXETH']\n",
    "price = prices.index('Close')\n",
    "\n",
    "# predict the closing price for the target date\n",
    "predicted_sequence = inference_model(inputs=[num_intervals, token, price])\n",
    "print(predicted_sequence)\n",
    "print(f\"Predicted Close price for token 'BTC' on {target_date}: {predicted_sequence[-1]}\")\n",
    "\n",
    "# append the predicted sequence to the latest sequence in the dataset\n",
    "final_sequence = latest_sequence_per_token[token][:, price] # shape: (len_sequence,)\n",
    "final_sequence = final_sequence * (max_price - min_price) + min_price   # unnormalize the sequence\n",
    "final_sequence = np.concatenate([final_sequence, predicted_sequence.numpy()]) # shape: (len_sequence + num_intervals,)\n",
    "\n",
    "# plot the final sequence\n",
    "plt.figure(figsize=(12, 4))\n",
    "plt.plot(final_sequence[:len_sequence], label='Final', color='blue')\n",
    "\n",
    "# plot the predicted sequence\n",
    "plt.plot(range(len_sequence-1, len(final_sequence)), final_sequence[len_sequence-1:], label='Predicted', color='red')\n",
    "\n",
    "plt.title(f'Predicted Close Price for Token \"BTC\" from {most_recent_date} to {target_date}')\n",
    "plt.xlabel('Interval')\n",
    "plt.ylabel('Closing Price')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: TokenPricePredictionModel/0000001\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: TokenPricePredictionModel/0000001\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved artifact at 'TokenPricePredictionModel/0000001'. The following endpoints are available:\n",
      "\n",
      "* Endpoint 'serve'\n",
      "  args_0 (POSITIONAL_ONLY): TensorSpec(shape=(None,), dtype=tf.float32, name=None)\n",
      "Output Type:\n",
      "  TensorSpec(shape=(None,), dtype=tf.float32, name=None)\n",
      "Captures:\n",
      "  1993030317184: TensorSpec(shape=(30, 101, 4), dtype=tf.float32, name=None)\n",
      "  1993025029744: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025030272: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025030800: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025030096: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025030976: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025032032: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025031680: TensorSpec(shape=(), dtype=tf.resource, name=None)\n",
      "  1993025032208: TensorSpec(shape=(), dtype=tf.resource, name=None)\n"
     ]
    }
   ],
   "source": [
    "    # Save the model\n",
    "inference_model.export('TokenPricePredictionModel/0000001')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12269faaf521f9b021cbce562702705f9895359f723122663aa8676531abbb8a\n"
     ]
    }
   ],
   "source": [
    "# deploy the model to a TensorFlow Serving container\n",
    "!docker run -d -p 8501:8501 --name tfserving_token_price_prediction --mount type=bind,source=/c/Users/Moses/Desktop/TokenPricePredictionModel/TokenPricePredictionModel,target=/models/TokenPricePredictionModel -e MODEL_NAME=TokenPricePredictionModel -t tensorflow/serving"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'predictions': [2536.04883, 2567.18066, 2603.61499]}\n"
     ]
    }
   ],
   "source": [
    "# make a request to the container to predict the closing price on 2024-02-04 for token 'FRXETH'\n",
    "\n",
    "# get the number of intervals between the most recent date and the target date (interval = 4 days)\n",
    "target_date = '2024-02-04'\n",
    "num_intervals = (pd.Timestamp(target_date) - most_recent_date).days // 4\n",
    "\n",
    "# make a request to the TensorFlow Serving container\n",
    "import requests\n",
    "import json\n",
    "\n",
    "url = 'http://localhost:8501/v1/models/TokenPricePredictionModel:predict'\n",
    "headers = {\"content-type\": \"application/json\"}\n",
    "data = json.dumps({\"signature_name\": \"serving_default\", \"instances\": [num_intervals, tokens[\"FRXETH\"], prices.index(\"Close\")]})\n",
    "response = requests.post(url, data=data, headers=headers)\n",
    "print(response.json())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}