17-train_RNN_RNN_Wikipedia_geo_high_25.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "606fa719",
   "metadata": {},
   "source": [
    "# Train 17-train_RNN_RNN_Wikipedia_geo_high_25"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a8a5bd39",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.nn.utils import clip_grad_norm_\n",
    "\n",
    "from tqdm import tqdm\n",
    "\n",
    "from rouge_score import rouge_scorer\n",
    "\n",
    "from utils_french.EmbMgr import EmbMgr\n",
    "from utils_french.Dataset import Dataset\n",
    "from utils_french.DataLoader import DataLoader\n",
    "from utils_french.preprocess_df import preprocess_df\n",
    "from utils_french.accuracy_nb_sent_per_doc import accuracy_nb_sent_per_doc_fn\n",
    "from utils_french.accuracy_prop_sent_per_doc import accuracy_prop_sent_per_doc_fn\n",
    "from utils_french.split_all_docs import split_all_docs\n",
    "\n",
    "from models.RNN_RNN import RNN_RNN\n",
    "\n",
    "from time import time\n",
    "\n",
    "import json\n",
    "import argparse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "818410ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "def is_notebook() -> bool:\n",
    "    try:\n",
    "        shell = get_ipython().__class__.__name__\n",
    "        if shell == 'ZMQInteractiveShell':\n",
    "            return True   # Jupyter notebook or qtconsole\n",
    "        elif shell == 'TerminalInteractiveShell':\n",
    "            return False  # Terminal running IPython\n",
    "        else:\n",
    "            return False  # Other type (?)\n",
    "    except NameError:\n",
    "        return False      # Probably standard Python interpreter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "690fa032",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', None)\n",
    "pd.set_option('display.max_rows', None)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "b2636e0b",
   "metadata": {},
   "source": [
    "## (Hyper-)parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6e7aad39",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parse args if script mode\n",
    "parser = argparse.ArgumentParser(description='extractive summary')\n",
    "\n",
    "parser.add_argument('-is_graphic',type=int,default=1,choices=[0,1])\n",
    "parser.add_argument('-gpu_num',type=int,default=0)\n",
    "parser.add_argument('-batch_size',type=int,default=32)\n",
    "parser.add_argument('-epochs',type=int,default=100)\n",
    "\n",
    "args = None\n",
    "\n",
    "if is_notebook():\n",
    "    args = parser.parse_args(\"\")\n",
    "else:\n",
    "    args = parser.parse_args()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "824eacf5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "parse:\n",
      "is_graphic: True\n",
      "cuda_num: 0\n",
      "epochs 100\n",
      "batch_size 32\n"
     ]
    }
   ],
   "source": [
    "# parameters\n",
    "is_graphic = args.is_graphic != 0\n",
    "cuda_num = args.gpu_num\n",
    "embed_name = \"cc.fr.300.vec\"\n",
    "\n",
    "# hyper-parameters\n",
    "vocab_size = 150000\n",
    "batch_size = args.batch_size\n",
    "epochs = args.epochs\n",
    "learning_rate = 1e-3\n",
    "early_stopping = 3\n",
    "model_name = \"17-train_RNN_RNN_Wikipedia_geo_high_25\"\n",
    "sub_folder_name = \"model_name__{}__time__{}__embed_name__{}__lr__{}__batch_size__{}__vocab_size__{}__cuda_num__{}\".format(model_name, time(), embed_name, learning_rate, batch_size, vocab_size, cuda_num)\n",
    "checkpoints_folder = \"./checkpoints/\" + sub_folder_name\n",
    "average_number_of_sentences_per_document = 3\n",
    "\n",
    "# print\n",
    "print(\"parse:\")\n",
    "print(\"is_graphic:\", is_graphic)\n",
    "print(\"cuda_num:\", cuda_num)\n",
    "print(\"epochs\", epochs)\n",
    "print(\"batch_size\", batch_size)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "c8ddd4d5",
   "metadata": {},
   "source": [
    "## PyTorch initialisation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d09bc3ed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of available GPUs: 6\n",
      "GPU 0: NVIDIA GeForce GTX 1080 Ti\n",
      "GPU 1: NVIDIA GeForce GTX 1080 Ti\n",
      "GPU 2: NVIDIA GeForce GTX 1080 Ti\n",
      "GPU 3: NVIDIA GeForce GTX 1080\n",
      "GPU 4: NVIDIA GeForce GTX 1080\n",
      "GPU 5: NVIDIA GeForce GTX 1080\n"
     ]
    }
   ],
   "source": [
    "# Check if a GPU is available\n",
    "if torch.cuda.is_available():\n",
    "    # Display the number of available GPUs\n",
    "    print(f\"Number of available GPUs: {torch.cuda.device_count()}\")\n",
    "    # Display the name of each GPU\n",
    "    for i in range(torch.cuda.device_count()):\n",
    "        print(f\"GPU {i}: {torch.cuda.get_device_name(i)}\")\n",
    "else:\n",
    "    print(\"No GPU available.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b0c5786e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cuda', index=0)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "if torch.cuda.is_available():  \n",
    "  dev = \"cuda:\" + str(cuda_num) \n",
    "else:  \n",
    "  dev = \"cpu\" \n",
    "\n",
    "device = torch.device(dev)\n",
    "device"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "3393fe4e",
   "metadata": {},
   "source": [
    "## Chargement des données"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "97644745",
   "metadata": {},
   "outputs": [],
   "source": [
    "embmgr = EmbMgr(\"./data/cc.fr.300.vec\", vocab_size=vocab_size, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fe745c8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "df_train = pd.read_json(\"./data/wiki_geo_high_25/train.json\")\n",
    "train_dataset = Dataset(preprocess_df(df_train, glovemgr=embmgr, is_sep_n=False, doc_column_name=\"flat_contents\", labels_column_name=\"labels_sentences\", remove_stop_word=True, stemming=False, trunc_sent=50, padding_sent=50, trunc_doc=100))\n",
    "train_iter = DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f8b39f3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_val = pd.read_json(\"./data/wiki_geo_high_25/val.json\")\n",
    "val_dataset = Dataset(preprocess_df(df_val, glovemgr=embmgr, is_sep_n=False, doc_column_name=\"flat_contents\", labels_column_name=\"labels_sentences\", remove_stop_word=True, stemming=False, trunc_sent=50, padding_sent=50, trunc_doc=100))\n",
    "val_iter = DataLoader(dataset=val_dataset, batch_size=1, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fb052a32",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_test = pd.read_json(\"./data/wiki_geo_high_25/test.json\")\n",
    "test_dataset = Dataset(preprocess_df(df_test, glovemgr=embmgr, is_sep_n=False, doc_column_name=\"flat_contents\", labels_column_name=\"labels_sentences\", remove_stop_word=True, stemming=False, trunc_sent=50, padding_sent=50, trunc_doc=100))\n",
    "test_iter = DataLoader(dataset=test_dataset, batch_size=1, shuffle=False)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "fc41eb07",
   "metadata": {},
   "source": [
    "## Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9683721f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "380f8ba1",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = RNN_RNN(device=device, vocab_size=vocab_size, word_embed=embmgr.getEmbeddings(), dim_word_embed=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2ae1a236",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RNN_RNN(\n",
       "  (word_embedding): Embedding(150001, 300)\n",
       "  (word_GRU): GRU(300, 200, batch_first=True, bidirectional=True)\n",
       "  (sent_GRU): GRU(400, 200, batch_first=True, bidirectional=True)\n",
       "  (rel_pos_emb): Embedding(11, 100)\n",
       "  (abs_pos_emb): Embedding(100, 100)\n",
       "  (Wdoc): Linear(in_features=400, out_features=400, bias=True)\n",
       "  (Wcontent): Linear(in_features=400, out_features=1, bias=False)\n",
       "  (Wsalience): Bilinear(in1_features=400, in2_features=400, out_features=1, bias=False)\n",
       "  (Wnovelty): Bilinear(in1_features=400, in2_features=400, out_features=1, bias=False)\n",
       "  (Wabs_pos): Linear(in_features=100, out_features=1, bias=False)\n",
       "  (Wrel_pos): Linear(in_features=100, out_features=1, bias=False)\n",
       ")"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4bc411ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_fn = nn.BCELoss()\n",
    "optimizer = torch.optim.Adamax(model.parameters(),lr=learning_rate)\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1, gamma=0.5)\n",
    "mae_fn = nn.L1Loss()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1852182a",
   "metadata": {},
   "source": [
    "## Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "fef6d114",
   "metadata": {},
   "outputs": [],
   "source": [
    "if not os.path.exists(checkpoints_folder):\n",
    "    os.makedirs(checkpoints_folder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5c803c4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_val[\"doc_splitted\"] = split_all_docs(df_val[\"flat_contents\"], is_sep_n = False)\n",
    "val_set = df_val\n",
    "scorer = rouge_scorer.RougeScorer(['rouge1', 'rouge2', 'rougeL'], use_stemmer=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0991d3c5",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1:   0%|          | 0/4 [00:00<?, ?batch/s]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1: 100%|██████████| 4/4 [00:00<00:00,  5.62batch/s, accuracy=0.399, loss=0.827, mae=0.466]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1 : val loss = 0.562, val mae = 0.402, val accuracy = 0.795, r1 = 0.248, r2 = 0.076, rL = 0.161\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2: 100%|██████████| 4/4 [00:00<00:00,  8.51batch/s, accuracy=0.516, loss=0.639, mae=0.384]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2 : val loss = 0.533, val mae = 0.322, val accuracy = 0.731, r1 = 0.246, r2 = 0.073, rL = 0.156\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 3: 100%|██████████| 4/4 [00:00<00:00,  8.52batch/s, accuracy=0.464, loss=0.625, mae=0.307]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3 : val loss = 0.513, val mae = 0.326, val accuracy = 0.746, r1 = 0.247, r2 = 0.075, rL = 0.157\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 4: 100%|██████████| 4/4 [00:00<00:00,  7.63batch/s, accuracy=0.499, loss=0.518, mae=0.311]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4 : val loss = 0.496, val mae = 0.338, val accuracy = 0.766, r1 = 0.244, r2 = 0.073, rL = 0.156\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 5: 100%|██████████| 4/4 [00:00<00:00,  8.05batch/s, accuracy=0.51, loss=0.522, mae=0.33]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5 : val loss = 0.493, val mae = 0.340, val accuracy = 0.777, r1 = 0.247, r2 = 0.074, rL = 0.160\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 6: 100%|██████████| 4/4 [00:00<00:00,  8.02batch/s, accuracy=0.512, loss=0.526, mae=0.333]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6 : val loss = 0.492, val mae = 0.339, val accuracy = 0.777, r1 = 0.247, r2 = 0.074, rL = 0.160\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 7: 100%|██████████| 4/4 [00:00<00:00,  8.63batch/s, accuracy=0.492, loss=0.505, mae=0.323]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7 : val loss = 0.492, val mae = 0.337, val accuracy = 0.777, r1 = 0.247, r2 = 0.074, rL = 0.160\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 8: 100%|██████████| 4/4 [00:00<00:00,  8.51batch/s, accuracy=0.498, loss=0.495, mae=0.317]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8 : val loss = 0.492, val mae = 0.337, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 9: 100%|██████████| 4/4 [00:00<00:00,  7.79batch/s, accuracy=0.499, loss=0.523, mae=0.328]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 10: 100%|██████████| 4/4 [00:00<00:00,  8.47batch/s, accuracy=0.499, loss=0.508, mae=0.321]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 11: 100%|██████████| 4/4 [00:00<00:00,  8.27batch/s, accuracy=0.498, loss=0.508, mae=0.323]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 12: 100%|██████████| 4/4 [00:00<00:00,  8.33batch/s, accuracy=0.499, loss=0.505, mae=0.322]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 12 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 13: 100%|██████████| 4/4 [00:00<00:00,  8.80batch/s, accuracy=0.495, loss=0.503, mae=0.322]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 14: 100%|██████████| 4/4 [00:00<00:00,  8.37batch/s, accuracy=0.486, loss=0.513, mae=0.322]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 14 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 15: 100%|██████████| 4/4 [00:00<00:00,  8.62batch/s, accuracy=0.515, loss=0.524, mae=0.329]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 15 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 16: 100%|██████████| 4/4 [00:00<00:00,  8.37batch/s, accuracy=0.496, loss=0.518, mae=0.326]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 17: 100%|██████████| 4/4 [00:00<00:00,  8.57batch/s, accuracy=0.484, loss=0.505, mae=0.32] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 18: 100%|██████████| 4/4 [00:00<00:00,  8.32batch/s, accuracy=0.522, loss=0.515, mae=0.326]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 18 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 19: 100%|██████████| 4/4 [00:00<00:00,  8.71batch/s, accuracy=0.491, loss=0.517, mae=0.322]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 19 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 20: 100%|██████████| 4/4 [00:00<00:00,  8.54batch/s, accuracy=0.487, loss=0.496, mae=0.317]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 21: 100%|██████████| 4/4 [00:00<00:00,  8.12batch/s, accuracy=0.505, loss=0.498, mae=0.318]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 21 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 22: 100%|██████████| 4/4 [00:00<00:00,  8.59batch/s, accuracy=0.487, loss=0.51, mae=0.321] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 22 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 23: 100%|██████████| 4/4 [00:00<00:00,  8.26batch/s, accuracy=0.495, loss=0.491, mae=0.316]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 23 : val loss = 0.492, val mae = 0.336, val accuracy = 0.773, r1 = 0.247, r2 = 0.074, rL = 0.159\n",
      "Training duration = 40.79742646217346\n"
     ]
    }
   ],
   "source": [
    "arr_train_loss = []\n",
    "arr_train_mae = []\n",
    "arr_train_acc = []\n",
    "arr_val_loss = []\n",
    "arr_val_mae = []\n",
    "arr_val_acc = []\n",
    "\n",
    "t1 = time()\n",
    "for epoch in range(1, epochs+1):\n",
    "    # train 1 epoch\n",
    "    model.train()\n",
    "    nb_batch_train = 0\n",
    "    total_train_loss = 0\n",
    "    total_train_mae = 0\n",
    "    total_train_acc = 0\n",
    "    with tqdm(train_iter, unit=\"batch\", total=len(train_iter)) as tepoch:\n",
    "        for batch in tepoch:\n",
    "            tepoch.set_description(f\"Epoch {epoch}\")\n",
    "            features = []\n",
    "            doc_lens = []\n",
    "\n",
    "            for j in range(batch_size):\n",
    "                doc_lens.append(len(batch[j][\"docs\"]))\n",
    "                features = features + batch[j][\"docs\"]\n",
    "            \n",
    "            features = torch.tensor(features, dtype=torch.long).to(device)\n",
    "\n",
    "            targets = [torch.tensor(batch[j][\"labels\"], dtype=torch.float) for j in range(batch_size)]\n",
    "            targets = torch.cat(targets)\n",
    "            targets = targets.to(device)\n",
    "            \n",
    "            probs = model(features, doc_lens)\n",
    "            \n",
    "            loss = loss_fn(probs, targets)\n",
    "            optimizer.zero_grad()\n",
    "            loss.backward()\n",
    "            clip_grad_norm_(model.parameters(), 1.0)\n",
    "            optimizer.step()\n",
    "            nb_batch_train += 1\n",
    "            total_train_loss += loss.item()\n",
    "            total_train_mae += mae_fn(probs, targets).item()\n",
    "            total_train_acc += accuracy_prop_sent_per_doc_fn(probs=probs.cpu().detach().numpy(), targets=targets.cpu().detach().numpy(), doc_lens=doc_lens)\n",
    "            tepoch.set_postfix(loss=total_train_loss/nb_batch_train, mae=total_train_mae/nb_batch_train, accuracy=total_train_acc/nb_batch_train)\n",
    "\n",
    "    # Save model\n",
    "    model.save(checkpoints_folder + \"/\" + model_name + \"-\" + str(epoch) + \".pt\")\n",
    "\n",
    "    # Show train and val score\n",
    "    model.eval()\n",
    "    nb_batch_val = 0\n",
    "    total_val_loss = 0\n",
    "    total_val_mae = 0\n",
    "    total_val_acc = 0\n",
    "\n",
    "    total_r1 = 0\n",
    "    total_r2 = 0\n",
    "    total_rl = 0\n",
    "\n",
    "    for i, batch in enumerate(val_iter):\n",
    "        features = []\n",
    "        doc_lens = []\n",
    "\n",
    "        for j in range(1):\n",
    "            doc_lens.append(len(batch[j][\"docs\"]))\n",
    "            features = features + batch[j][\"docs\"]\n",
    "            \n",
    "        features = torch.tensor(features, dtype=torch.long).to(device)\n",
    "\n",
    "        targets = [torch.tensor(batch[j][\"labels\"], dtype=torch.float) for j in range(1)]\n",
    "        targets = torch.cat(targets)\n",
    "        targets = targets.to(device)\n",
    "        \n",
    "        y_pred = model(features, doc_lens)\n",
    "\n",
    "        loss = loss_fn(y_pred, targets)\n",
    "\n",
    "        nb_batch_val += 1\n",
    "        total_val_loss += loss.item()\n",
    "\n",
    "        doc = val_set[\"doc_splitted\"].iloc[i]\n",
    "        summaries = val_set[\"summaries\"].iloc[i]\n",
    "\n",
    "        indices = torch.argsort(y_pred, descending=True)\n",
    "\n",
    "        y_pred_thresh = []\n",
    "\n",
    "        txt = \"\"\n",
    "\n",
    "        for j in range(min(len(doc), y_pred.shape[0])):\n",
    "            txt = txt + \". \" + doc[indices[j]]\n",
    "            y_pred_thresh.append(indices[j])\n",
    "            if len(txt) >= len(summaries):\n",
    "                break\n",
    "\n",
    "        y_pred_thresh.sort()\n",
    "\n",
    "        txt = \"\"\n",
    "\n",
    "        for j in y_pred_thresh:\n",
    "            txt = txt + \". \" + doc[j]\n",
    "\n",
    "        n = min(len(txt), len(summaries))\n",
    "\n",
    "        while n < len(txt) and txt[n].isalnum():\n",
    "            n += 1\n",
    "\n",
    "        txt = txt[:n]\n",
    "\n",
    "        #assert len(txt) - len(summaries) <= 20\n",
    "\n",
    "        scores = scorer.score(summaries, txt)\n",
    "        total_r1 += scores[\"rouge1\"].recall\n",
    "        total_r2 += scores[\"rouge2\"].recall\n",
    "        total_rl += scores[\"rougeL\"].recall\n",
    "\n",
    "        probs = y_pred.tolist() # compute_probs(y_pred)\n",
    "        probs_tensor = y_pred #torch.tensor(probs, dtype=torch.float).to(device)\n",
    "\n",
    "        total_val_mae += mae_fn(probs_tensor, targets).item()\n",
    "        total_val_acc += accuracy_nb_sent_per_doc_fn(probs=probs, targets=targets.cpu().detach().numpy(), doc_lens=doc_lens, average_number_of_sentences_per_document=average_number_of_sentences_per_document)\n",
    "\n",
    "    arr_train_loss.append(total_train_loss / nb_batch_train)\n",
    "    arr_train_mae.append(total_train_mae / nb_batch_train)\n",
    "    arr_train_acc.append(total_train_acc / nb_batch_train)\n",
    "\n",
    "    arr_val_loss.append(total_val_loss / nb_batch_val)\n",
    "    arr_val_mae.append(total_val_mae / nb_batch_val)\n",
    "    arr_val_acc.append(total_val_acc / nb_batch_val)\n",
    "    \n",
    "    print(\"Epoch {} : val loss = {:.3f}, val mae = {:.3f}, val accuracy = {:.3f}, r1 = {:.3f}, r2 = {:.3f}, rL = {:.3f}\".format(epoch, total_val_loss / nb_batch_val, total_val_mae / nb_batch_val, total_val_acc / nb_batch_val, total_r1 / nb_batch_val, total_r2 / nb_batch_val, total_rl / nb_batch_val))\n",
    "\n",
    "    if len(arr_val_loss) >= early_stopping+1:\n",
    "        if min(arr_val_loss[-early_stopping:]) >= arr_val_loss[-(early_stopping+1)]:\n",
    "            break\n",
    "\n",
    "    # update lr\n",
    "    scheduler.step()\n",
    "\n",
    "t2 = time()\n",
    "print(\"Training duration =\", t2-t1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "2f39ceed",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_metrics = {}\n",
    "training_metrics[\"duration\"]   = t2 - t1\n",
    "training_metrics[\"train_loss\"] = arr_train_loss\n",
    "training_metrics[\"train_mae\"]  = arr_train_mae\n",
    "training_metrics[\"train_acc\"]  = arr_train_acc\n",
    "training_metrics[\"val_loss\"]   = arr_val_loss\n",
    "training_metrics[\"val_mae\"]    = arr_val_mae\n",
    "training_metrics[\"val_acc\"]    = arr_val_acc\n",
    "\n",
    "# Save to file in JSON format\n",
    "\n",
    "with open(checkpoints_folder + \"/training_metrics.json\", 'w') as fp:\n",
    "    json.dump(training_metrics, fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8cd8fbee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAHHCAYAAABXx+fLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAA9hAAAPYQGoP6dpAABZWUlEQVR4nO3dd3hUZd7G8e/MJJkkQBJCIAUCoUOQoiBZiru6BkFdBCsoSlHxXQQboiu6gG3FsrKsiuK6suiuBUWsICtGYS0URZQWOkhLgUAy6WXmvH8MGQgECGlnJnN/rutcmZw5c+Y3GcLcec5TLIZhGIiIiIj4EavZBYiIiIjUNwUgERER8TsKQCIiIuJ3FIBERETE7ygAiYiIiN9RABIRERG/owAkIiIifkcBSERERPyOApCIiIj4HQUgERER8TsKQCLiU+bPn4/FYuHHH380uxQR8WEKQCIiIuJ3FIBERETE7ygAiUiDs27dOi6//HLCwsJo3Lgxl156KatWrapwTGlpKY899hgdO3YkODiYZs2aMXDgQJYtW+Y5Jj09nXHjxtGqVSvsdjuxsbEMGzaMPXv21PMrEpHaFmB2ASIitWnTpk1cdNFFhIWF8eCDDxIYGMirr77KxRdfzIoVK0hKSgLg0UcfZebMmdx+++307dsXh8PBjz/+yE8//cSgQYMAuPbaa9m0aRN33XUXCQkJZGZmsmzZMvbu3UtCQoKJr1JEaspiGIZhdhEiIlU1f/58xo0bxw8//ECfPn1Ouf/qq69myZIlpKam0q5dOwDS0tLo3Lkz559/PitWrACgV69etGrVis8++6zS58nOzqZp06Y899xzTJkype5ekIiYQpfARKTBcDqdfPHFFwwfPtwTfgBiY2O56aab+Pbbb3E4HABERESwadMmtm/fXum5QkJCCAoKYvny5Rw9erRe6heR+qMAJCINxqFDhygoKKBz586n3Ne1a1dcLhf79u0D4PHHHyc7O5tOnTrRvXt3HnjgAdavX+853m6388wzz/D5558THR3Nb3/7W5599lnS09Pr7fWISN1RABIRv/Tb3/6WnTt3Mm/ePM477zz++c9/csEFF/DPf/7Tc8y9997Ltm3bmDlzJsHBwUybNo2uXbuybt06EysXkdqgACQiDUbz5s0JDQ1l69atp9y3ZcsWrFYr8fHxnn2RkZGMGzeOd955h3379tGjRw8effTRCo9r3749999/P1988QUbN26kpKSE559/vq5fiojUMQUgEWkwbDYbl112GR9//HGFoeoZGRm8/fbbDBw4kLCwMACysrIqPLZx48Z06NCB4uJiAAoKCigqKqpwTPv27WnSpInnGBHxXRoGLyI+ad68eSxduvSU/Y8++ijLli1j4MCB3HnnnQQEBPDqq69SXFzMs88+6zkuMTGRiy++mN69exMZGcmPP/7IwoULmTRpEgDbtm3j0ksv5YYbbiAxMZGAgAA+/PBDMjIyGDlyZL29ThGpGxoGLyI+pXwY/Ons27ePQ4cOMXXqVL777jtcLhdJSUn85S9/oV+/fp7j/vKXv/DJJ5+wbds2iouLadOmDbfccgsPPPAAgYGBZGVlMWPGDFJSUti3bx8BAQF06dKF+++/n+uvv74+XqqI1CEFIBEREfE76gMkIiIifkcBSERERPyOApCIiIj4HQUgERER8TsKQCIiIuJ3FIBERETE72gixEq4XC4OHjxIkyZNsFgsZpcjIiIiVWAYBrm5ucTFxWG1nrmNRwGoEgcPHqywXpCIiIj4jn379tGqVaszHqMAVIkmTZoA7h9g+bpBIiIi4t0cDgfx8fGez/EzUQCqRPllr7CwMAUgERERH1OV7ivqBC0iIiJ+RwFIRERE/I4CkIiIiPgd9QGqAafTSWlpqdll+KSgoKCzDlEUERGpKwpA1WAYBunp6WRnZ5tdis+yWq20bduWoKAgs0sRERE/pABUDeXhp0WLFoSGhmqyxHNUPtFkWloarVu31s9PRETqnQLQOXI6nZ7w06xZM7PL8VnNmzfn4MGDlJWVERgYaHY5IiLiZ9QJ4xyV9/kJDQ01uRLfVn7py+l0mlyJiIj4IwWgatJlm5rRz09ERMykACQiIiJ+RwFIqiUhIYHZs2ebXYaIiEi1qBO0H7n44ovp1atXrQSXH374gUaNGtW8KBERERMoANUjl2FQ5nRhwUJggPc1vhmGgdPpJCDg7P8smjdvXg8ViYiI1A3v+xRuwDIdRWxJzyUzr7jen3vs2LGsWLGCv//971gsFiwWC/Pnz8disfD555/Tu3dv7HY73377LTt37mTYsGFER0fTuHFjLrzwQr788ssK5zv5EpjFYuGf//wnV199NaGhoXTs2JFPPvmknl+liIhI1SgA1QLDMCgoKTvrVuY0KCp14igordLxZ9sMw6hyjX//+9/p168f48ePJy0tjbS0NOLj4wF46KGHePrpp0lNTaVHjx7k5eVxxRVXkJKSwrp16xgyZAhDhw5l7969Z3yOxx57jBtuuIH169dzxRVXMGrUKI4cOVKjn62IiEhd0CWwWlBY6iRx+n/r/Xk3Pz6Y0KCqvYXh4eEEBQURGhpKTEwMAFu2bAHg8ccfZ9CgQZ5jIyMj6dmzp+f7J554gg8//JBPPvmESZMmnfY5xo4dy4033gjAU089xQsvvMCaNWsYMmTIOb82ERGRuqQWIKFPnz4Vvs/Ly2PKlCl07dqViIgIGjduTGpq6llbgHr06OG53ahRI8LCwsjMzKyTmkVERGpCLUC1ICTQxubHB5/1uFKni63puQAkxoVhreFkgCGBtho9vtzJo7mmTJnCsmXL+Otf/0qHDh0ICQnhuuuuo6Sk5IznOXlJC4vFgsvlqpUaRUREapMCUC2wWCxVuhRlGAYhQQEYhkGQzUZQPY8ECwoKqtLSE9999x1jx47l6quvBtwtQnv27Knj6kREROqPLoHVI4vFQqDN3epT6qz/lpGEhARWr17Nnj17OHz48GlbZzp27MiiRYv4+eef+eWXX7jpppvUkiMiIg2KAlA9C7S6f+RmBKApU6Zgs9lITEykefPmp+3TM2vWLJo2bUr//v0ZOnQogwcP5oILLqjnakVEROqOxTiXsdR+wuFwEB4eTk5ODmFhYRXuKyoqYvfu3bRt25bg4OBzPvferAKyC0uIDQ+heRN7bZXsc2r6cxQRETnZmT6/T6YWoHoWGGDeJTARERFxUwCqZ2ZeAhMRERE3BaB6drwTtK48ioiImEUBqJ4F2tw/8jK1AImIiJhGAaielQegUqdxTmt5iYiISO1RAKpnATYLFiwYGJS5FIBERETMoABUzywWCwEmToYoIiIiCkCmOPEymIiIiNQ/BSATmLkchoiIiCgAmeJ4C5BvBaCEhARmz55tdhkiIiI1pgBkAs0FJCIiYi4FIBP4aguQiIhIQ6EAZAIzAtA//vEP4uLicLkqPuewYcO49dZb2blzJ8OGDSM6OprGjRtz4YUX8uWXX9ZbfSIiIvVJAag2GAaU5Fd5C3QWYCktoKwoH6M475weW2E7h4kUr7/+erKysvj66689+44cOcLSpUsZNWoUeXl5XHHFFaSkpLBu3TqGDBnC0KFD2bt3b138xEREREwVYHYBDUJpATwVV+XDg4DutfG8Dx+EoEZVOrRp06ZcfvnlvP3221x66aUALFy4kKioKC655BKsVis9e/b0HP/EE0/w4Ycf8sknnzBp0qTaqFZERMRrmN4CNGfOHBISEggODiYpKYk1a9ac8fjZs2fTuXNnQkJCiI+P57777qOoqKhG5/QXo0aN4oMPPqC4uBiAt956i5EjR2K1WsnLy2PKlCl07dqViIgIGjduTGpqqlqARESkQTK1BWjBggVMnjyZuXPnkpSUxOzZsxk8eDBbt26lRYsWpxz/9ttv89BDDzFv3jz69+/Ptm3bGDt2LBaLhVmzZlXrnLUiMNTdGnMOdmTmUVjqpE1kKGEhgdV/3nMwdOhQDMNg8eLFXHjhhXzzzTf87W9/A2DKlCksW7aMv/71r3To0IGQkBCuu+46SkpKqlebiIiIFzM1AM2aNYvx48czbtw4AObOncvixYuZN28eDz300CnHf//99wwYMICbbroJcM9Lc+ONN7J69epqn7NWWCxVvhRVLiAYDEoptYVAkL1u6jpJcHAw11xzDW+99RY7duygc+fOXHDBBQB89913jB07lquvvhqAvLw89uzZUy91iYiI1DfTLoGVlJSwdu1akpOTjxdjtZKcnMzKlSsrfUz//v1Zu3at55LWrl27WLJkCVdccUW1zwlQXFyMw+GosNU1s5bDGDVqlCcQjho1yrO/Y8eOLFq0iJ9//plffvmFm2666ZQRYyIiIg2FaQHo8OHDOJ1OoqOjK+yPjo4mPT290sfcdNNNPP744wwcOJDAwEDat2/PxRdfzMMPP1ztcwLMnDmT8PBwzxYfH1/DV3d2Zi2H8fvf/57IyEi2bt3qaUkDd8tZ06ZN6d+/P0OHDmXw4MGe1iEREZGGxqdGgS1fvpynnnqKl19+maSkJHbs2ME999zDE088wbRp06p93qlTpzJ58mTP9w6Ho85DkFmTIVqtVg4ePLW/UkJCAl999VWFfRMnTqzwvS6JiYhIQ2FaAIqKisJms5GRkVFhf0ZGBjExMZU+Ztq0adxyyy3cfvvtAHTv3p38/HzuuOMOHnnkkWqdE8But2O3108/nHJaDkNERMQ8pl0CCwoKonfv3qSkpHj2uVwuUlJS6NevX6WPKSgowGqtWLLNZgPAMIxqndMsJ7YAGecwoaGIiIjUnKmXwCZPnsyYMWPo06cPffv2Zfbs2eTn53tGcI0ePZqWLVsyc+ZMwD2Me9asWZx//vmeS2DTpk1j6NChniB0tnN6i/IA5DIMXIaBzWIxuSIRERH/YWoAGjFiBIcOHWL69Omkp6fTq1cvli5d6unEvHfv3gotPn/+85+xWCz8+c9/5sCBAzRv3pyhQ4fyl7/8pcrn9BZWqwWb1YLTZVDqNLCZPiWliIiI/7AYuv5yCofDQXh4ODk5OYSFhVW4r6ioiN27d5OQkEBISEiNnmdbRi5FpU7aRjWiSXA1J0P0UYWFhezZs4e2bdsSHBxsdjkiItIAnOnz+2RqdzhHgYHuoFJQUFDzc5k0EswblM8wXX7pUkREpD751DB4b2Cz2YiIiCAzMxOA0NBQLNXsv2NxlmCUlVJQaCHU5j8NcS6Xi0OHDhEaGkpAgP4JiohI/dOnTzWUD6kvD0HV5SgqxVFYRr7dRm5oUG2U5jOsViutW7eudngUERGpCQWgarBYLMTGxtKiRQtKS0urfZ7PN6Tx16+30rdtJDOv6VyLFXq/oKCgU6Y0EBERqS8KQDVgs9lq1IelWXhjDuQ6Sc0sUkdgERGReqQ/wU0UG+4OPWk5RSZXIiIi4l8UgEwUcywA5RSWUlBSZnI1IiIi/kMByERNggNpFOS+hJauViAREZF6owBksvJWIAUgERGR+qMAZLLYcPds0uoHJCIiUn8UgEzmaQFyKACJiIjUFwUgk8XqEpiIiEi9UwAyWYyGwouIiNQ7BSCTeVqAHIUmVyIiIuI/FIBMFhPm7gStS2AiIiL1RwHIZOWXwA7nlVBc5jS5GhEREf+gAGSypqGBBAW434ZMR7HJ1YiIiPgHBSCTWSwWrQkmIiJSzxSAvEBMWHkAUkdoERGR+qAA5AXKW4AyNBmiiIhIvVAA8gIxWg5DRESkXikAeQHNBi0iIlK/FIC8gGaDFhERqV8KQF5ALUAiIiL1SwHIC5SPAsvMLaLM6TK5GhERkYZPAcgLNGtsJ8BqwWXAoTxNhigiIlLXFIC8gM1qITpM/YBERETqiwKQl4hRPyAREZF6owDkJRSARERE6o8CkJeIPXYJLF2zQYuIiNQ5BSAvobmARERE6o8CkJeIPbYcRroWRBUREalzCkBeIibcDqgFSEREpD4oAHmJ8gVRMxxFuFyGydWIiIg0bApAXqJFEzsWC5Q6DbLyS8wuR0REpEFTAPISgTYrzRu7L4NpKLyIiEjdUgDyIp5FUTUUXkREpE4pAHmR45MhaiSYiIhIXVIA8iLlQ+E1EkxERKRueUUAmjNnDgkJCQQHB5OUlMSaNWtOe+zFF1+MxWI5Zbvyyis9x4wdO/aU+4cMGVIfL6VGtByGiIhI/Qgwu4AFCxYwefJk5s6dS1JSErNnz2bw4MFs3bqVFi1anHL8okWLKCk5PkoqKyuLnj17cv3111c4bsiQIfzrX//yfG+32+vuRdSSWM0GLSIiUi9MbwGaNWsW48ePZ9y4cSQmJjJ37lxCQ0OZN29epcdHRkYSExPj2ZYtW0ZoaOgpAchut1c4rmnTpvXxcmokWuuBiYiI1AtTA1BJSQlr164lOTnZs89qtZKcnMzKlSurdI7XX3+dkSNH0qhRowr7ly9fTosWLejcuTMTJkwgKyvrtOcoLi7G4XBU2MxwvAWoEMPQZIgiIiJ1xdQAdPjwYZxOJ9HR0RX2R0dHk56eftbHr1mzho0bN3L77bdX2D9kyBDefPNNUlJSeOaZZ1ixYgWXX345Tqez0vPMnDmT8PBwzxYfH1/9F1UD5S1ARaUucgpLTalBRETEH5jeB6gmXn/9dbp3707fvn0r7B85cqTndvfu3enRowft27dn+fLlXHrppaecZ+rUqUyePNnzvcPhMCUEBQfaiGwUxJH8EtJyiogIDar3GkRERPyBqS1AUVFR2Gw2MjIyKuzPyMggJibmjI/Nz8/n3Xff5bbbbjvr87Rr146oqCh27NhR6f12u52wsLAKm1li1A9IRESkzpkagIKCgujduzcpKSmefS6Xi5SUFPr163fGx77//vsUFxdz8803n/V59u/fT1ZWFrGxsTWuua7Faii8iIhInTN9FNjkyZN57bXXeOONN0hNTWXChAnk5+czbtw4AEaPHs3UqVNPedzrr7/O8OHDadasWYX9eXl5PPDAA6xatYo9e/aQkpLCsGHD6NChA4MHD66X11QTMRoKLyIiUudM7wM0YsQIDh06xPTp00lPT6dXr14sXbrU0zF67969WK0Vc9rWrVv59ttv+eKLL045n81mY/369bzxxhtkZ2cTFxfHZZddxhNPPOFTcwFpOQwREZG6YzE03voUDoeD8PBwcnJy6r0/0Ps/7uOBheu5qGMU/74tqV6fW0RExJedy+e36ZfApKLy9cDUB0hERKTuKAB5Ga0HJiIiUvcUgLxMeQDKLS4jr7jM5GpEREQaJgUgL9PYHkCTYHffdLUCiYiI1A0FIC+kuYBERETqlgKQF4o51hE6TUPhRURE6oQCkBeKDVMLkIiISF1SAPJCntmgtR6YiIhInVAA8kIaCi8iIlK3FIC8kNYDExERqVsKQF5I64GJiIjULQUgLxQb5h4FdrSglKJSp8nViIiINDwKQF4oLCSAkEAbABnqCC0iIlLrFIC8kMVi8VwGUz8gERGR2qcA5KU0EkxERKTuKAB5KY0EExERqTsKQF4qJkwjwUREROqKApCXUh8gERGRuqMA5KXKF0RN1ygwERGRWqcA5KXUAiQiIlJ3FIC8VHkn6MN5xZQ6XSZXIyIi0rAoAHmpyNAggmxWDAMyc4vNLkdERKRBUQDyUlarhehwO6CRYCIiIrVNAciLla8Jpn5AIiIitUsByItpNmgREZG6oQDkxTQbtIiISN1QAPJix2eDVgASERGpTQpAXuz4XEDqBC0iIlKbFIC8mPoAiYiI1A0FIC8We2w5jMzcYpwuw+RqREREGg4FIC/WvIkdm9VCmcsgK0+TIYqIiNQWBSAvZrNaaNHEPRmiRoKJiIjUHgUgL6eh8CIiIrVPAcjLHR8Kr5FgIiIitUUByMt5WoAcagESERGpLQpAXi5WQ+FFRERqnQKQl4sJ14KoIiIitU0ByMuVtwBl6BKYiIhIrVEA8nLlnaDTcoowDE2GKCIiUhu8IgDNmTOHhIQEgoODSUpKYs2aNac99uKLL8ZisZyyXXnllZ5jDMNg+vTpxMbGEhISQnJyMtu3b6+Pl1Lroo8FoJIyF0cLSk2uRkREpGEwPQAtWLCAyZMnM2PGDH766Sd69uzJ4MGDyczMrPT4RYsWkZaW5tk2btyIzWbj+uuv9xzz7LPP8sILLzB37lxWr15No0aNGDx4MEVFvncZKSjASlTj8skQNRReRESkNpgegGbNmsX48eMZN24ciYmJzJ07l9DQUObNm1fp8ZGRkcTExHi2ZcuWERoa6glAhmEwe/Zs/vznPzNs2DB69OjBm2++ycGDB/noo4/q8ZXVHo0EExERqV2mBqCSkhLWrl1LcnKyZ5/VaiU5OZmVK1dW6Ryvv/46I0eOpFGjRgDs3r2b9PT0CucMDw8nKSnptOcsLi7G4XBU2LxJdJhmgxYREalNpgagw4cP43Q6iY6OrrA/Ojqa9PT0sz5+zZo1bNy4kdtvv92zr/xx53LOmTNnEh4e7tni4+PP9aXUKbUAiYiI1C7TL4HVxOuvv0737t3p27dvjc4zdepUcnJyPNu+fftqqcLaofXAREREapepASgqKgqbzUZGRkaF/RkZGcTExJzxsfn5+bz77rvcdtttFfaXP+5czmm32wkLC6uweRNPC5BDnaBFRERqg6kBKCgoiN69e5OSkuLZ53K5SElJoV+/fmd87Pvvv09xcTE333xzhf1t27YlJiamwjkdDgerV68+6zm9VYwugYmIiNSqALMLmDx5MmPGjKFPnz707duX2bNnk5+fz7hx4wAYPXo0LVu2ZObMmRUe9/rrrzN8+HCaNWtWYb/FYuHee+/lySefpGPHjrRt25Zp06YRFxfH8OHD6+tl1arYE5bDMAwDi8VickUiIiK+zfQANGLECA4dOsT06dNJT0+nV69eLF261NOJee/evVitFRuqtm7dyrfffssXX3xR6TkffPBB8vPzueOOO8jOzmbgwIEsXbqU4ODgOn89daF8NuiCEie5xWWEBQeaXJGIiIhvsxhaX+EUDoeD8PBwcnJyvKY/UK/HvyC7oJQv7vstnaKbmF2OiIiI1zmXz2+fHgXmT2I0F5CIiEitUQDyEcc7QmskmIiISE0pAPmIWM0FJCIiUmsUgHxETJh7JJiGwouIiNScApCPOD4ZogKQiIhITSkA+QhNhigiIlJ7FIB8hPoAiYiI1B4FIB9R3gKUU1hKQUmZydWIiIj4NgUgH9EkOJBGQTZAl8FERERqSgHIh6gfkIiISO1QAPIhJy6KKiIiItWnAORDYjQUXkREpFYoAPmQ4yPBtByGiIhITSgA+ZDjfYCKTa5ERETEtykA+ZDjs0GrBUhERKQmFIB8iNYDExERqR0KQD6kvAXocF4JxWVOk6sRERHxXQpAPiQiNJCgAPdblulQPyAREZHqUgDyIRaLRWuCiYiI1AIFIB8TE6ah8CIiIjWlAORjYrUchoiISI0pAPmYmGPLYWg2aBERkepTAPIxagESERGpOQUgHxOjTtAiIiI1Vq0A9MYbb7B48WLP9w8++CARERH079+fX3/9tdaKk1OpBUhERKTmqhWAnnrqKUJC3H1RVq5cyZw5c3j22WeJiorivvvuq9UCpaLyUWCZuUWUOV0mVyMiIuKbAqrzoH379tGhQwcAPvroI6699lruuOMOBgwYwMUXX1yb9clJmjW2E2C1UOYyOJRXTOyxTtEiIiJSddVqAWrcuDFZWVkAfPHFFwwaNAiA4OBgCgs1P01dslktRIepH5CIiEhNVKsFaNCgQdx+++2cf/75bNu2jSuuuAKATZs2kZCQUJv1SSViwoM5kF2ofkAiIiLVVK0WoDlz5tCvXz8OHTrEBx98QLNmzQBYu3YtN954Y60WKKfSSDAREZGaqVYLUEREBC+99NIp+x977LEaFyRnF3vsEliGJkMUERGplmq1AC1dupRvv/3W8/2cOXPo1asXN910E0ePHq214qRyagESERGpmWoFoAceeACHwwHAhg0buP/++7niiivYvXs3kydPrtUC5VTlI7/StSCqiIhItVTrEtju3btJTEwE4IMPPuAPf/gDTz31FD/99JOnQ7TUHbUAiYiI1Ey1WoCCgoIoKCgA4Msvv+Syyy4DIDIy0tMyJHWnPABlOIpwuQyTqxEREfE91WoBGjhwIJMnT2bAgAGsWbOGBQsWALBt2zZatWpVqwXKqVo0sWOxQKnTICu/hOZN7GaXJCIi4lOq1QL00ksvERAQwMKFC3nllVdo2bIlAJ9//jlDhgyp1QLlVIE2K80bu0OP5gISERE5d9VqAWrdujWfffbZKfv/9re/1bggqZrY8GAyc4tJyymke6tws8sRERHxKdUKQABOp5OPPvqI1NRUALp168ZVV12FzWarteLk9GLCg/llf47mAhIREamGal0C27FjB127dmX06NEsWrSIRYsWcfPNN9OtWzd27tx5TueaM2cOCQkJBAcHk5SUxJo1a854fHZ2NhMnTiQ2Nha73U6nTp1YsmSJ5/5HH30Ui8VSYevSpUt1XqZXKx8Kr5FgIiIi565aLUB333037du3Z9WqVURGRgKQlZXFzTffzN13383ixYurdJ4FCxYwefJk5s6dS1JSErNnz2bw4MFs3bqVFi1anHJ8SUkJgwYNokWLFixcuJCWLVvy66+/EhERUeG4bt268eWXXx5/kQHVbujyWuUjwdQHSERE5NxVKxmsWLGiQvgBaNasGU8//TQDBgyo8nlmzZrF+PHjGTduHABz585l8eLFzJs3j4ceeuiU4+fNm8eRI0f4/vvvCQwMBKh08dWAgABiYmLO8VX5lljNBSQiIlJt1boEZrfbyc3NPWV/Xl4eQUFBVTpHSUkJa9euJTk5+XgxVivJycmsXLmy0sd88skn9OvXj4kTJxIdHc15553HU089hdPprHDc9u3biYuLo127dowaNYq9e/eesZbi4mIcDkeFzdtFH1sPLF19gERERM5ZtQLQH/7wB+644w5Wr16NYRgYhsGqVav44x//yFVXXVWlcxw+fBin00l0dHSF/dHR0aSnp1f6mF27drFw4UKcTidLlixh2rRpPP/88zz55JOeY5KSkpg/fz5Lly7llVdeYffu3Vx00UWVBrZyM2fOJDw83LPFx8dX6TWY6XgLUCGGockQRUREzkW1AtALL7xA+/bt6devH8HBwQQHB9O/f386dOjA7Nmza7nE41wuFy1atOAf//gHvXv3ZsSIETzyyCPMnTvXc8zll1/O9ddfT48ePRg8eDBLliwhOzub995777TnnTp1Kjk5OZ5t3759dfYaakt5C1BRqYucwlKTqxEREfEt1eoDFBERwccff8yOHTs8w+C7du1Khw4dqnyOqKgobDYbGRkZFfZnZGSctv9ObGwsgYGBFYbad+3alfT0dEpKSiq9/BYREUGnTp3YsWPHaWux2+3Y7b41m3JwoI3IRkEcyS8hLaeIiNCqXXoUERGRcwhAZ1vl/euvv/bcnjVr1lnPFxQURO/evUlJSWH48OGAu4UnJSWFSZMmVfqYAQMG8Pbbb+NyubBa3Y1X27ZtIzY29rR9j/Ly8ti5cye33HLLWWvyNTFhwRzJLyE9p4iusWFmlyMiIuIzqhyA1q1bV6XjLBZLlZ988uTJjBkzhj59+tC3b19mz55Nfn6+Z1TY6NGjadmyJTNnzgRgwoQJvPTSS9xzzz3cddddbN++naeeeoq7777bc84pU6YwdOhQ2rRpw8GDB5kxYwY2m40bb7yxynX5itjwYDanOdQRWkRE5BxVOQCd2MJTW0aMGMGhQ4eYPn066enp9OrVi6VLl3o6Ru/du9fT0gMQHx/Pf//7X+677z569OhBy5Ytueeee/jTn/7kOWb//v3ceOONZGVl0bx5cwYOHMiqVato3rx5rddvthgNhRcREakWi6EhRKdwOByEh4eTk5NDWJj3Xlp66avt/PWLbdzQpxXPXtfT7HJERERMdS6f39UaBSbeIUbLYYiIiFSLApAPiwnTchgiIiLVoQDkw7QemIiISPUoAPmw8gCUW1zGhv05JlcjIiLiOxSAfFhjewA9WoUDcM0r3/Hqip04XerTLiIicjYKQD7uX2MvZFBiNKVOg5mfb+Gm11ZxILvQ7LJERES8mgKQj2vW2M4/bunN09d0JzTIxurdRxgy+398/PMBs0sTERHxWgpADYDFYmFk39YsufsiesVHkFtUxj3v/sxd76wjp0ALpYqIiJxMAagBSYhqxMI/9uPe5I7YrBY+/eUgQ/7+P77fcdjs0kRERLyKAlADE2Czcm9yJxb+sR8JzUJJyynipn+u5i+LN1Nc5jS7PBEREa+gANRAnd+6KYvvvogb+7YG4LVvdjPspe/Yku4wuTIRERHzKQA1YI3sAcy8pjuvje5Ds0ZBbEnP5aoXv+Of3+zCpeHyIiLixxSA/MCgxGiW3vtbft+lBSVOF08uTuXm11eTlqPh8iIi4p8UgPxE8yZ2Xh/Th79cfR4hgTa+35nF4L/9j09/OWh2aSIiIvVOAciPWCwWRiW1YfHdA+nZKhxHURl3vbOO+xb8jKNIw+VFRMR/KAD5oXbNG7NwQn/uvrQjVgt8uO4Al8/+hlW7sswuTUREpF4oAPmpQJuVyYM68f4f+9M6MpQD2YXc+NoqZn6equHyIiLS4CkA+bnebZqy5J6LGNEnHsOAV1fsYvic7zmo9cRERKQBUwASGtsDeOa6Hrx6S2+ahgaSmuZg9pfbzC5LRESkzgSYXYB4j8HdYrAAd/x7Lev355hdjoiISJ1RC5BUcF7LcAC2Z+ZRVKq+QCIi0jApAEkFseHBNA0NxOky2Jqea3Y5IiIidUIBSCqwWCx0i3O3Am06qHXDRESkYVIAklN0iwsDYNNB9QMSEZGGSQFITpHoCUBqARIRkYZJAUhOUX4JbEu6A6dWjRcRkQZIAUhO0TaqESGBNopKXew6lGd2OSIiIrVOAUhOYbNa6BrbBNBlMBERaZgUgKRSx0eCqSO0iIg0PApAUqnzWqojtIiINFwKQFKpE+cCMgx1hBYRkYZFAUgq1TG6MQFWCzmFpRzQyvAiItLAKABJpewBNjpGqyO0iIg0TApAclrdNCGiiIg0UApAclrlAWizRoKJiEgDowAkp1XeEXrjAbUAiYhIw6IAJKdVPhliuqOIrLxik6sRERGpPQpAclpNggNJaBYKqB+QiIg0LApAckYnzgckIiLSUJgegObMmUNCQgLBwcEkJSWxZs2aMx6fnZ3NxIkTiY2NxW6306lTJ5YsWVKjc8rpJXpGgqkjtIiINBymBqAFCxYwefJkZsyYwU8//UTPnj0ZPHgwmZmZlR5fUlLCoEGD2LNnDwsXLmTr1q289tprtGzZstrnlDM7PhJMLUAiItJwWAwT1zlISkriwgsv5KWXXgLA5XIRHx/PXXfdxUMPPXTK8XPnzuW5555jy5YtBAYG1so5K+NwOAgPDycnJ4ewsLBqvrqG4VBuMRf+5UssFtj46GAa2QPMLklERKRS5/L5bVoLUElJCWvXriU5Ofl4MVYrycnJrFy5stLHfPLJJ/Tr14+JEycSHR3Neeedx1NPPYXT6az2OQGKi4txOBwVNnFr3sROdJgdw4DUNP1cRESkYTAtAB0+fBin00l0dHSF/dHR0aSnp1f6mF27drFw4UKcTidLlixh2rRpPP/88zz55JPVPifAzJkzCQ8P92zx8fE1fHUNizpCi4hIQ2N6J+hz4XK5aNGiBf/4xz/o3bs3I0aM4JFHHmHu3Lk1Ou/UqVPJycnxbPv27aulihuGbuoILSIiDYxpHTqioqKw2WxkZGRU2J+RkUFMTEylj4mNjSUwMBCbzebZ17VrV9LT0ykpKanWOQHsdjt2u70Gr6Zh05pgIiLS0JjWAhQUFETv3r1JSUnx7HO5XKSkpNCvX79KHzNgwAB27NiBy+Xy7Nu2bRuxsbEEBQVV65xyduWXwLZl5FJS5jrL0SIiIt7P1EtgkydP5rXXXuONN94gNTWVCRMmkJ+fz7hx4wAYPXo0U6dO9Rw/YcIEjhw5wj333MO2bdtYvHgxTz31FBMnTqzyOeXctWoaQlhwAKVOg+2ZuWaXIyIiUmOmjmkeMWIEhw4dYvr06aSnp9OrVy+WLl3q6cS8d+9erNbjGS0+Pp7//ve/3HffffTo0YOWLVtyzz338Kc//anK55RzZ7FYSIwLY9WuI2w66PC0CImIiPgqU+cB8laaB+hUT3y2mde/3c2Yfm14bNh5ZpcjIiJyCp+YB0h8izpCi4hIQ6IAJFVSftkrNc2By6VGQxER8W0KQFIl7Zs3wh5gJb/EyZ6sfLPLERERqREFIKmSAJuVLjFNAF0GExER36cAJFWWqCUxRESkgVAAkio7r6WWxBARkYZBAUiqrLwj9OaDDjR7goiI+DIFIKmyLjFNsFktZOWXkOEoNrscERGRalMAkioLDrTRvnkjQJfBRETEtykAyTnppo7QIiLSACgAyTk5PiO0WoBERMR3KQDJOUnUkhgiItIAKADJOekW674Etv9oITkFpSZXIyIiUj0KQHJOwkMDadU0BNBlMBER8V0KQHLOtDK8iIj4OgUgOWfHR4KpBUhERHyTApCcM7UAiYiIr1MAknNW3gK081AehSVOk6sRERE5dwpAcs6iw+xENQ7CZcCWdLUCiYiI71EAknNmsVhI1IzQIiLiwxSApFrUD0hERHyZApBUS3kA2qyRYCIi4oMUgKRayjtCb0nPpczpMrkaERGRc6MAJNXSJjKUxvYAistc7DyUb3Y5IiIi50QBSKrFarXQNbYJoAkRRUTE9ygASbV100gwERHxUQpAUm2JnpFgagESERHfogAk1XbiUHjDMEyuRkREpOoUgKTaOrZoQqDNQm5RGfuOFJpdjoiISJUpANW3o79CA2ktCQqw0ilaHaFFRMT3KADVpx//BXP6wprXzK6k1mhGaBER8UUKQPXJVQZlRbBsGhzaZnY1teK8luUjwdQCJCIivkMBqD5deDu0v9QdghaNB2ep2RXVmFqARETEFykA1SeLBYbNgZCmkPYzrHjW7IpqrEtMGBYLZOYWcyi32OxyREREqkQBqL6FxcIfZrtvf/NX2LfG1HJqqpE9gLZRjQBdBhMREd+hAGSGbsOhx0gwXLDoDijOM7uiGtGM0CIi4msUgMxyxbMQHg9Hd8MXj5hdTY2U9wParAAkIiI+QgHILMHhMPwVwAJr58PWpWZXVG3dtCSGiIj4GAUgM7W9CPpNdN/+ZBLkHTK3nmoqvwS2J6uA3CLfH9kmIiINn1cEoDlz5pCQkEBwcDBJSUmsWXP6jsHz58/HYrFU2IKDgyscM3bs2FOOGTJkSF2/jOq5dDq06Ab5h+DTe3xylujIRkHEhrvfg9S0XJOrEREROTvTA9CCBQuYPHkyM2bM4KeffqJnz54MHjyYzMzM0z4mLCyMtLQ0z/brr7+ecsyQIUMqHPPOO+/U5cuovgA7XPMPsAXB1sWw7j9mV1QtugwmIiK+xPQANGvWLMaPH8+4ceNITExk7ty5hIaGMm/evNM+xmKxEBMT49mio6NPOcZut1c4pmnTpnX5Mmom5jz4/Z/dt5c+BEd2m1tPNSQeuwy28YA6QouIiPczNQCVlJSwdu1akpOTPfusVivJycmsXLnytI/Ly8ujTZs2xMfHM2zYMDZt2nTKMcuXL6dFixZ07tyZCRMmkJWVddrzFRcX43A4Kmz1rt8kaDMASvLgwz+Cy1n/NdSAWoBERMSXmBqADh8+jNPpPKUFJzo6mvT09Eof07lzZ+bNm8fHH3/Mf/7zH1wuF/3792f//v2eY4YMGcKbb75JSkoKzzzzDCtWrODyyy/H6aw8VMycOZPw8HDPFh8fX3svsqqsNveosKAmsG8VfDe7/muogfIAtCMzj+Iy3wpvIiLif0y/BHau+vXrx+jRo+nVqxe/+93vWLRoEc2bN+fVV1/1HDNy5EiuuuoqunfvzvDhw/nss8/44YcfWL58eaXnnDp1Kjk5OZ5t37599fRqTtK0DVzxnPv2109B2i/m1FENLSNCCA8JpMxlsC3dtyd2FBGRhs/UABQVFYXNZiMjI6PC/oyMDGJiYqp0jsDAQM4//3x27Nhx2mPatWtHVFTUaY+x2+2EhYVV2EzTcyR0vcq9cvyiO6C00LxazoHFYuG8lroMJiIivsHUABQUFETv3r1JSUnx7HO5XKSkpNCvX78qncPpdLJhwwZiY2NPe8z+/fvJyso64zFew2JxrxXWOBoObYEvHzO7oirTkhgiIuIrTL8ENnnyZF577TXeeOMNUlNTmTBhAvn5+YwbNw6A0aNHM3XqVM/xjz/+OF988QW7du3ip59+4uabb+bXX3/l9ttvB9wdpB944AFWrVrFnj17SElJYdiwYXTo0IHBgweb8hrPWaNm7lXjAVa/Aju/NreeKlJHaBER8RUBZhcwYsQIDh06xPTp00lPT6dXr14sXbrU0zF67969WK3Hc9rRo0cZP3486enpNG3alN69e/P999+TmJgIgM1mY/369bzxxhtkZ2cTFxfHZZddxhNPPIHdbjflNVZLx0HQ5zb48XX46E6483sI8eKh/BwPQKlpuThdBjarxeSKREREKmcxDB+ceriOORwOwsPDycnJMbc/UEk+vPpbyNoB510H171uXi1V4HQZnDfjvxSWOvly8u/o0KKx2SWJiIgfOZfPb9MvgckZBDWCq/8BFhtsXAgbFppd0RnZrBa6xDYBdBlMRES8mwKQt2vVG373oPv24smQs//Mx5us/DLYZnWEFhERL6YA5Asuuh9a9oaiHHd/IJfL7IpOSyPBRETEFygA+QJboPtSWGAo7F4Ba149+2NMcuJIMHUvExERb6UA5CuiOsBlT7pvL5sBmanm1nManaKbYLNaOFpQSlpOkdnliIiIVEoByJf0uRU6DAJnMSwaD2UlZld0iuBAGx2Pjf7aeEAdoUVExDspAPkSiwWGvQQhkZC+AZbPNLuiSiV6LoOpH5CIiHgnBSBf0yQGhv7dffu72fDrSlPLqYw6QouIiLdTAPJFiVdBr1FguODD/4Mi7woa53mGwusSmIiIeCcFIF815GkIbw3Zv8LSqeBFI67KL4EdzCniaL739VMSERFRAPJVwWFw9VzAAj//B1Ie95oQ1CQ4kDbNQgFdBhMREe+kAOTLEgbAkGMdob+dBf992GtCkFaGFxERb6YA5Ot+MwGu+Kv79qqXYfH9XjFTtDpCi4iIN1MAagj6joerXgIs8OPr8Old4HKaWlKiWoBERMSLKQA1FBfcAtf8AyxWWPcf9+gwZ5lp5ZRfAtt1OJ+CEvPqEBERqYwCUEPS4wa47l9gDYAN78PCcabNFt2iSTDNm9gxDEhNyzWlBhERkdNRAGpoug2HEf8BWxCkfgLvjYZSc9bk6ubl8wGVOl1sSXew6Kf9PLUklX/8bycul3d0IhcRkboVYHYBUgc6Xw43vgPvjoJtn8O7N8KItyAotF7L6BYXxvKth7yiI/TR/BJS0xxsTnOQmpZLapqDHZl5lDgrdhjfmZnPzGu6Y7VaTKpUpPaVlLnDfkJUI8KCA80uR8QrKAA1VB2SYdT78PYI2PkVvH0D3Pgu2BvXWwlmjARzugz2ZOWTmuY4trnDzulWpm9iD6BrbBjxkaF8uG4/C37ch4HB09f08IsQlFtUShN9IDZYhmGQkprJX5aksvtwPoE2C79p14xBidEkd40mLiLE7BJFTGMxDC+ZOMaLOBwOwsPDycnJISwszOxyambvKvjPdVCSC/FJ7lAUHF4vT/1rVj6/e245QTYrmx4fTKCtdq+45hWXseVY0NmclsvmNAfb0nMpLK18BFzryFC6xjaha2wYibFhdI0No1XTECwWd9D59JeD3LvgZ5wug+t6t+KZa3tga8Ah6PVvd/PEZ5u5LDGa567vSXiIglBDsiXdwZOfpfLtjsMABAVYKSmr2OJ5XsswBnWN4bJu0XSJaeL5XRDxVefy+a0AVIkGFYAA9q+F/1wNRTkQdwHc/AGERtb507pcBj0f+4Lc4jKW3H2RZ2h8deQVl7HxQA7r92fzy/4cNh7I4desgkqPDQ600jkmjMQTwk7nmCZVaun4bP1B7nnXHYKuvaAVz17XMEPQur1HuX7uSsqO9XmKjwzh5Zt6071V/YRjqTuH84qZtWwb767Zi8uAIJuVWwe2ZeIl7TmUW8yyzRks25zB2r1HK8yb2qppCIMSoxmUGE3fhEgCavkPFpH6oABUQw0uAAGk/QJvDofCIxDTHW75CBpF1fnTjnh1Jat3H+G563pwfZ/4Kj2mqNTJ5jQH6/dls/5ADuv357DzUF6lk1xHh9k9rTnlW9uoRjUKLYvXp3H3u+twugyuOb8lz13fs0GFIEdRKVe+8A37jhRyUcco9mTls+9IIUE2K9OGJnJzUmu1BPig4jInb3y/hxdTdpBb7J564oruMTw0pCutm53a/+9QbjFfbXGHoW+2H6b4hNah8JBAft+lBYMSo/ltp+Y0tqu3hPgGBaAaapABCCBjM7w5DPIzoXkXGP0JNImu06d8/NPNzPtuN2P7J/DoVd1Oub/U6WJbRi7r9x9r3dmXw7aMXE/LxIniwoPp3iqcHq0i6NEqnG5x4UQ2CqqTupdsSOOud9wh6OrzW/LXBhKCDMPg7nd/5tNfDtKqaQiL774IDJiy8BeWbc4A4KqecTx1TXd96PkIwzD4YnMGTy1J9bSKdosLY/ofEklq16xK5ygoKeOb7YdZtjmDlNQMjhaUeu4LslkZ0KEZgxJjSO7aghZhwXXyOkRqgwJQDTXYAARweDu8cRXkHoRmHdwhKLxlnT3dB2v3c//7v9A3IZJ37/gNuw7n8cu+HDYcyOGX/dlsPuio8JdnuWaNguhxQtjp0SqC5k3sdVZnZT4/FoLKXAbDesXx/PU9ff6ywPs/7uOBheuxWS2893/96N2mKeD+EH392908/fkWylwG7Zo34pVRvekc08TkiuVMNh3M4YnPNrNq1xEAmjex8+Dgzlx7Qatqd+J3ugzW/nqULzalsyw145RLzb3iIxiUGM1lidF0aNFYrYXiVRSAaqhBByCAI7vdIShnL0S0gTGfQtM2dfJUW9IdDJn9DQFWC8GBNvKKT50VuklwAD1ahdO9ZQQ9W4XTIz6CuPBgr/iPdenGdCa9/RNlLoOresYx6wbfDUE7D+Xxhxe+pbDUyQODOzPxkg6nHPPjniNMensd6Y4iggOtPDm8O9f1bmVCtXImh3KLef6Lre5Ri4a7g/MdF7VjwsXtaVSLLXeGYbA9M49lmzP4YnMGv+zLrnD/eS3DeOrq7vRoFVFrzylSEwpANdTgAxBA9j54Yygc3Q1hrWDMJ9Csfa0/TanTRe8nluEocgef4EAr58W5W3R6xofTvWU4Cc0aefWQ8/9uSmfiW+4QNLRnHH/zwRBUXObk6jnfsznNQf/2zfj3bUmnvaSXlVfMfe/9wv+2HQJgRJ94HhvWjeBAW32WLJUoKnUy77vdvPz1Ts8fE3/oEctDl3ehVdO6n+crw1Hk6US9cmcWJU4XVgvcflE77kvuREhQw/o3YhgGBSVOjuSXuLeCEjBgQIcoggJ86/8Af6EAVEN+EYAAHGnw5lVweBs0jnG3BDXvVOtPs/FADlvTc+nWMowOzRv7XHgA+GJTOhPf/olSp8GVPWL5+4hePvU6yvtiRTYK4vN7LiL6LP04XC6DOV/v4G9fbsNlQJeYJrw86gLaNa+/eaTkOMMw+HxjOk8tSWX/0UIAerYKZ9ofEumTUPcjOiuTlVfM459t5uOfDwLQplkoT1/Tg37tq9bvyAylThdHjwWZI/klHM0v5Uh+MUfySzlaUEJWfon7/hMCz8lTBwD0a9eMeWMvbHCBryFQAKohvwlAAHmZ7tFhmZugUXMY/TFEn9pZWeDLzRlMeGutOwR1j2X2yF61PrdRXfhqSwa3zv8RgHlj+/D7LlXv+P79jsPc/e46DueV0NgewDPX9uDKHrF1VapUYsN+dz+fNXvc/XxiwoL50+WdGdazpVe0nKakZvDIhxtJd7gnG72xb2umXtHF9Bmn92YV8No3u9h4MMcTaHKLqrcwc1CAlWaNgmgaGsSerHwKSpxc1DGK10b3Ucuol1EAqiG/CkAABUfg38PdQ+VDmrqHyMf1Mrko75SSmsGE//xEidPF5efF8MKN53t1CMpwFHH537/hSH4J4wYkMGPouYfbTEcRk95Zx5rd7g/gMf3a8PCVXbEH6D/+upTpKOK5/25l4U/7MQz35eP/+217/u937QgN8q4Reo6iUp75fAtvrd4LuKen+Mvw7iQn1u0o08rsOpTHnK938tHPB3BWMprUYoGmoUE0DQ2kWSM7TRsFEtkoiMhjAaf89onfhwbZPH0Sf9hzhNGvr6Gw1MklnZsz95be+l3wIgpANeR3AQigMBv+cy0c+BHs4XDTAmjTz+yqvNJXWzL447/dIWhItxhevMk7Q5DTZXDL66v5fmcWibFhfDixf7X/oy5zupi1bBsvL98JuC+/vHTTBcRH1u/6cv4gp7CU+d/t4dX/7aSgxD2r+fBecTw4pIvXL12xalcWD32wnj3HRo4N7RnHjKGJRDWu+xGc2zNyeenrHXz6y0HKc8/vOjXnhj7xtAiz0zQ0iGaNgggLCazxlBYrd2Yxbv4aikpdDEqM5uVRF3jl/wH+SAGohvwyAAEU58JbN8De78EaCFc+D73HmF2VV/p6Syb/9++1lDhdDO4WzYs3XuB1nSLnfL2D5/67lZBAG5/dPZD2tdB/56stGdy34BdyCksJDwlk1g09ubRr/f+V3xAdyS9h3re7eeP7PZ6JDM9vHcH0PyRyfuumJldXdUWlTv725TZe+98uXAY0DQ1kxtBuDOsVVycjO7ekO3jxqx0s2ZDmmSw1uWsLJv2+I73iI2r9+cp9s/0Qt73xIyVlLq7oHsMLI8/3qX6BDZUCUA35bQACKMmHjybA5o/d3184HobMBJvWiTrZ11uPhaAy91+Bc27ynhD007GlLpwug2ev68ENVZyFuyr2Hy1g0tvr+PnYkOj/+107Hriss/7zr6ZMRxH/+N8u3lq917OOXafoxtz1+478oUesV0wHUR3r92fz4ML1bEnPBeCSzs35y9Xda60Va+OBHF78ajv/3ZTh2Te4WzR3/b4j57WsnyVdvt6ayf+96f5D6KqecfxtRK8GMWGqL1MAqiG/DkAAhgHf/BW+etL9fcJFcP38elk6w9cs35rJHcdCUHJXd1O42SHIUVTKFX//hv1HCxnaM44XRvaq9Q/RkjIXMz9P5V/f7QGgb0IkL9x4PjHhmiW4qvYfLeDVFbtY8OM+z0ij7i3DmfT7DgzqGu0VHZxrqtTp4tUVO3khZQclTheNgmw8dHkXRiW1qfbr+3lfNi+mbCdlSybg7tNzRfdY7vp9B7rE1P//18s2ZzDhP2spcxlcc0FL/npdzwbx3pUzDIP/bT9McIC1yjOLm0kBqIb8PgCV27IEFt3hXkk+vDWMfAtie5hdlddZse0Q49/88VgIasGcUReY1inSMAzuemcdn61PIz7SvdRFXY7GWbIhjQcXrievuIxmjYL4+8jzGdhRQflMdh3K45XlO/lw3QHPki992jRl0u878LtOzX22xedMdmTm8qcPNrD216OAOzDPvLb7OV2WXfvrEV5I2cGKY/NTWS3uPkaTLulAx2hzZyz/fEMak44tnTPywnieurp7gwhBP+/L5snPNvPjsfftkSu6Mv637Uyu6swUgGpIAegEmVvg3RvhyC4IDIXhL0O3q82uyuv871gIKi5z8fsuLXjlZnNC0Hs/7OPBD9YTYLXw/h/71UvfkT2H85nw1k+kpjmwWGB4r5aEhwRitViwWcFqtWCzWLBZLcf2nXibSva5j7da3fcHWK00b2InLjyE6HC7z4642ZLuYM7XO1m8/ngn3YEdopj0+w4ktY1skMHnRC6Xwb9X/cozS7dQUOIkKMDKvckdGX9RuzN2IF61K4sXv9rOdzuyALBZLQzv1ZKJl7T3qnmpPvnlIPe+uw6XATf/pjVPDDvPZ9/T/UcLeO6/Wz1zPAXaLJQ63f9obxvYlkeu6Oq1AU8BqIYUgE5SeBQW3gY7U9zfXzQFLnkErOrzcaJvtx/mtjd+oLjMxSWdm/PKzb3rdY6QHZl5DH3RvdTFg0M6c+fFpy51UVeKSp089ukm3lmzr86fK6pxEDHhwcSEhRAXEUxMeDCx4cHEhocQGx5MdFiwV83N8su+bF76eodnsVlwd9KdeEkHn+rcXFv2Hy3g4Q83emYa7xYXxjPX9qjQb8cwDL7fmcXfU7Z7pl8IsFq4rncr7ry4Q6Wr23uDRT+51z40DBg3IIHpf0j0qRCUW1TKK8t38s9vd1NS5sJigWvOb8WUwZ345OeDzPx8C+Buefvr9T288o8RBaAaUgCqhMsJXz4K37/g/r7T5XDNPyBYP58TfbfDHYKKSl1c3Lk5c+spBBWVOrn65e9JTXMwoEMz/n1rkil/oX29JZN1+7JxuQychuH+euJtw8DpouL9hvsY17GvThee2y7DoLjMxaHcYg5mF1a6cG5lmjUKqhCMTg5JMeF1H5LW7D7CS1/v8HzQl/dVmXhxBxLj/Pv3xjAMFv10gMc/20xOYSk2q4X/+2077r6047EWnx2ey2VBNivX92nFhIvb18tyHzVV3goLcMdv2zH18i5eH4LKnC4W/LiPvy3bxuG8EgB+0y6SP1+ZWCGYfrTuAFPe/4Uyl0G/ds14dXRv0ye8PJkCUA0pAJ3BLwvgk7vAWQxRnWHk2xBVfy0NvuD7HYe59VgIig0P5s5LOnBDn1Z1+tfSo59sYv73e4hsFMTSey6ixVmWuvBFhmGQXVBKWk4R6Y5CDmYXkZ5TxMGcQtJzjt8uKq16SIqLcAeiuAh3a5L7e/ftFk2Cz3lEj2EYfLvjMC9+tcPTclF+yWbCxe3p0MJ7Ltl4g0O5xTz6ySYWb0gDoLE9wLPGWVCAlZv6tub/fteO2HDvnv/oZG+t/pVHPtwIwKRLOjBlcGeTKzq95VszeWpJKtsy8gBoG9WIqZd3YVBidKXB7Zvth/jjv9eSX+KkS0wT3ri171mX1qlPCkA1pAB0Fgd+gndHQe5B96SJ182DjslmV+VVVu3K4p5315HhKAYg7lgQur4OgtCXmzO4/U33Uhf/Gnshl3RpUavn9yWGYZBT6A5JaTmF7q/ZRZ7QlJZd9ZAUYLUQHRZMXETwsVB0LCSFhxAbEUzLiBDCQwKxWCwYhsGXqZm89NV2ftmfAxxvufjj79prwsizWLoxnWkfb+RQbjHBgVZuTmrDHb9t59NBfv53u3n0080ATB7Uibsv7WhyRRVtTc/lL0tSPS2UEaGB3HNpR0YltTnrSNaNB3IY+68fOJxXTMuIEN649UI6tDC3I3o5nwtAc+bM4bnnniM9PZ2ePXvy4osv0rdv30qPnT9/PuPGjauwz263U1RU5PneMAxmzJjBa6+9RnZ2NgMGDOCVV16hY8eq/QNUAKqC3Ax47xbYtxosVkh+FPrf7W7nF8B9WWrBD/t4efmOU4LQDX3ia2W4fIajiCGz/8fRglJuHdCW6UMTa3zOhq48JB3ILjwWjgo5cOzrwexjLUuOokqXUThZSKCN2IhgDAN2H84H3EtW3NTX/QGuaQGqLqewlOVbMxnQIapeZo6uD//8ZhdPLk4FqPd+eadzKLeYWcu2seCHvbgMdwfnMf0SuOv3HQkPrfrlrH1HChg9bw27D+cTERrI62P60LuNOQvznsinAtCCBQsYPXo0c+fOJSkpidmzZ/P++++zdetWWrQ49S/Z+fPnc88997B161bPPovFQnT08dlon3nmGWbOnMkbb7xB27ZtmTZtGhs2bGDz5s0EB5/9PyQFoCoqK4bF98O6f7u/7349XPUiBPpWc3VdqywItYwI4c5L2nN97+oHIafL4OZ/rmblriy6xYWx6M7qL3UhFTldhrvf0bFQlJZd5A5MOe6AlJZT6OkrUa6xPYDR/dpw68C2DeYDXGru5eU7eHap+/Pqz1d25faLzBlGXlTq5PVvd/Py1zvIP7bEyuXnxfCnIV1IiGpUrXMeyS/h1vk/8PO+bOwBVl688Xwu6xZTm2WfM58KQElJSVx44YW89NJLALhcLuLj47nrrrt46KGHTjl+/vz53HvvvWRnZ1d6PsMwiIuL4/7772fKlCkA5OTkEB0dzfz58xk5cuRZa1IAOgeGAT/8Ez7/ExhOiO3lni8ovJXZlXmdolIn767Zy8vLd5KZezwITbykA9f1bnXOQah8qYvQIBuf3TXQq4YE+4OiUqen35GjsJR+7aLO6S9o8R9//3I7f/tyGwCPXdWNMf0T6u25XS6DT9cf5NmlWzmQXQhAj1bh/PnKRPq2rXmLTUFJGXe9vY6ULZlYLfDE8PMYldSmxuetrnP5/DZ1HHNJSQlr164lOfl4/xGr1UpycjIrV6487ePy8vJo06YN8fHxDBs2jE2bNnnu2717N+np6RXOGR4eTlJS0mnPWVxcjMPhqLBJFVks0Hc8jP4YQptB2s/wj4th7yqzK/M6wYE2xg5oy/8evIQZQxNp0cTOgexCHv5wA5f8dTlvr97rmRH4bNb+epRZy47/h6rwU/+CA20kRDWif/sohpwXq/Ajp3X3pR2YdIn78teMTzbx1upf6+V5f9hzhKtf/o573v2ZA9mFxIUHM3tELz66c0CthB+A0KAAXr2lNyP6xOMy4JEPNzLri614Qe+aszI1AB0+fBin01nh8hVAdHQ06enplT6mc+fOzJs3j48//pj//Oc/uFwu+vfvz/79+wE8jzuXc86cOZPw8HDPFh9fe+sm+Y22F8H4ryG6O+Qfgvl/gLXzza7KKwUH2hh3QhBqflIQemfNmYNQTmEpdx+bdXZYrziu663WNhFvZrFYuP+yTvzfsVmUH/lwI+/9UHdzZv2alc+db63l+rkr+WV/Do2CbDwwuDNfTbmY4ee3rPUpMgJsVp6+truno/cLX+3goQ82UOas2h90Zgkwu4Bz1a9fP/r16+f5vn///nTt2pVXX32VJ554olrnnDp1KpMnT/Z873A4FIKqo2kbuO2/8NGdsPkj+PQeSN8AQ57WYqqVKA9CN/Ztzdur9/LKip0cyC5k6qINzPl6B5Mu6cC1vVtVmCXXMAweXrSBA9mFxEeG8ORw351tVsSfWCwWHrq8CyVOF//6bg9/WrSewAALV59/7n/AGIbB0YJS9h8tYP/RQvYdcX8t/3734XzKXAZWC4y4sDX3DepIiyZ12yHfYrEweVAnYsKC+fNHG1jw4z4O5RXz0k3nExrknVHD1KqioqKw2WxkZGRU2J+RkUFMTNU6UgUGBnL++eezY8cOAM/jMjIyiI2NrXDOXr16VXoOu92O3a5Oi7UiqJF74dRvnncvpvrDPyEzFW54U4upnkZwoI1bB7blpqTWvLV6L68s38n+o4U8tGgDL50UhN77cR+LN6QRYLXwwsjzaeJlk5CJyOlZLBam/yGRUqeL/6zay/3v/UKA1crQnnEVjisfqVgeavYdOR5uyveVd2Q+nYs6RvHIlV3rfYHYm5Ja07yJnUlv/8RXWzK58bXVzBvTh2ZeODDAKzpB9+3blxdffBFwd4Ju3bo1kyZNqrQT9MmcTifdunXjiiuuYNasWZ5O0FOmTOH+++8H3C06LVq0UCfo+rb1c/hgvHsx1YAQiOoIUZ2ObcduN2uvUWMnKSxx8vYadxA6nOfuLN2qaQg3/6YNf/9yO4WlTv40pAsTLm5vcqUiUh0ul8HDH27g3R/2eWbBLihxVmjFKZ8Q8kyiw+y0ahpKq6YhxB/72qppKG2auW+b2Tq89tcj3PbGj2QXlNI2qhFvjOtbL0uY+NQosAULFjBmzBheffVV+vbty+zZs3nvvffYsmUL0dHRjB49mpYtWzJz5kwAHn/8cX7zm9/QoUMHsrOzee655/joo49Yu3YtiYnuOVCeeeYZnn766QrD4NevX69h8GY4tBUW3AyHt53mAAtEtD41GEV1crcY+fHlncISJ2+t/pW5K3Z5ghC4F9B889a+XrsYoYicnctl8MDC9Xzw0/7THhPV2E58ZEglIcc9Mac3rXlXmR2ZeYyZt4YD2YVENbYzf9yFFZbWqAvn8vlt+oW5ESNGcOjQIaZPn056ejq9evVi6dKlnk7Me/fuxXrCoptHjx5l/PjxpKen07RpU3r37s3333/vCT8ADz74IPn5+dxxxx1kZ2czcOBAli5dWqXwI7WseWe4cxUc2e0OQVnb3V8Pb3eHo6JsyP7Vve1YVvGxwREnBaNj4ahpgl/0KQoJsnH7Re0YldTGE4SCA63MuqGnwo+Ij7NaLTx7XQ/at2jEpoMOT+tNedBpGRFCSJB3B5yz6dCiMYvu7M/Yf/1AapqDEa+uZO4tvbmoY3OzSwO8oAXIG6kFqJ4YBhRkHQtEx0JR+e2jvwKn+adpDXCHoOAI9+WzwNCTvla2r5L7ghpV/N5m9+oV7ssXDj2xU7SIiLdzFJXyx3+v5fudWQRYLTx3fY9qdf6u0nP50iUwb6QA5AVKC+HIrlOD0eHtUFpQh09scQcsq+2krwFgqWSfNcAdmip8bzt+rMXiPqfnUp6lkn2ctM9S+b7TnuPEr6e5/7TnO6mOOnHy66zqfVTtPmr5cXXyXNWllj5pGJwugy9TM9iW6V50tX/7Zlxw0ZVYOlxaq8/jU5fARCoVGALR3dzbiVwu9yKsR3ZBSb47DJUUuANT6clfq7jPWXzCExjgKnVvIiJSK2zAYGBweer4Fb4rKWVALQegc6EAJL7FanUvs1GbS224nO4gVFbsXs7DVXbC5jrpe2clxziPbSfvKwMM96W+8st55bcr/cqZjznr40/8yknnPcOxde2URmajhvef4dgqP646j6nl5zorH2ic1wUEqYZNB3P4Yc9ROrWqfNHz+qIAJGK1gb2xexMRkTrVDbBn5tKhRRNT61BvShEREalXZocfUAASERERP6QAJCIiIn5HAUhERET8jgKQiIiI+B0FIBEREfE7CkAiIiLidxSARERExO8oAImIiIjfUQASERERv6MAJCIiIn5HAUhERET8jgKQiIiI+B0FIBEREfE7AWYX4I0MwwDA4XCYXImIiIhUVfnndvnn+JkoAFUiNzcXgPj4eJMrERERkXOVm5tLeHj4GY+xGFWJSX7G5XJx8OBBDMOgdevW7Nu3j7CwMLPLkhM4HA7i4+P13ngZvS/eS++Nd9L7UrsMwyA3N5e4uDis1jP38lELUCWsViutWrXyNKWFhYXpH6aX0nvjnfS+eC+9N95J70vtOVvLTzl1ghYRERG/owAkIiIifkcB6AzsdjszZszAbrebXYqcRO+Nd9L74r303ngnvS/mUSdoERER8TtqARIRERG/owAkIiIifkcBSERERPyOApCIiIj4HQWgM5gzZw4JCQkEBweTlJTEmjVrzC7Jrz366KNYLJYKW5cuXcwuyy/973//Y+jQocTFxWGxWPjoo48q3G8YBtOnTyc2NpaQkBCSk5PZvn27OcX6kbO9L2PHjj3ld2jIkCHmFOtHZs6cyYUXXkiTJk1o0aIFw4cPZ+vWrRWOKSoqYuLEiTRr1ozGjRtz7bXXkpGRYVLF/kEB6DQWLFjA5MmTmTFjBj/99BM9e/Zk8ODBZGZmml2aX+vWrRtpaWme7dtvvzW7JL+Un59Pz549mTNnTqX3P/vss7zwwgvMnTuX1atX06hRIwYPHkxRUVE9V+pfzva+AAwZMqTC79A777xTjxX6pxUrVjBx4kRWrVrFsmXLKC0t5bLLLiM/P99zzH333cenn37K+++/z4oVKzh48CDXXHONiVX7AUMq1bdvX2PixIme751OpxEXF2fMnDnTxKr824wZM4yePXuaXYacBDA+/PBDz/cul8uIiYkxnnvuOc++7Oxsw263G++8844JFfqnk98XwzCMMWPGGMOGDTOlHjkuMzPTAIwVK1YYhuH+/QgMDDTef/99zzGpqakGYKxcudKsMhs8tQBVoqSkhLVr15KcnOzZZ7VaSU5OZuXKlSZWJtu3bycuLo527doxatQo9u7da3ZJcpLdu3eTnp5e4fcnPDycpKQk/f54geXLl9OiRQs6d+7MhAkTyMrKMrskv5OTkwNAZGQkAGvXrqW0tLTC70yXLl1o3bq1fmfqkAJQJQ4fPozT6SQ6OrrC/ujoaNLT002qSpKSkpg/fz5Lly7llVdeYffu3Vx00UXk5uaaXZqcoPx3RL8/3mfIkCG8+eabpKSk8Mwzz7BixQouv/xynE6n2aX5DZfLxb333suAAQM477zzAPfvTFBQEBERERWO1e9M3dJq8OIzLr/8cs/tHj16kJSURJs2bXjvvfe47bbbTKxMxDeMHDnSc7t79+706NGD9u3bs3z5ci699FITK/MfEydOZOPGjeq/6AXUAlSJqKgobDbbKT3wMzIyiImJMakqOVlERASdOnVix44dZpciJyj/HdHvj/dr164dUVFR+h2qJ5MmTeKzzz7j66+/plWrVp79MTExlJSUkJ2dXeF4/c7ULQWgSgQFBdG7d29SUlI8+1wuFykpKfTr18/EyuREeXl57Ny5k9jYWLNLkRO0bduWmJiYCr8/DoeD1atX6/fHy+zfv5+srCz9DtUxwzCYNGkSH374IV999RVt27atcH/v3r0JDAys8DuzdetW9u7dq9+ZOqRLYKcxefJkxowZQ58+fejbty+zZ88mPz+fcePGmV2a35oyZQpDhw6lTZs2HDx4kBkzZmCz2bjxxhvNLs3v5OXlVWg12L17Nz///DORkZG0bt2ae++9lyeffJKOHTvStm1bpk2bRlxcHMOHDzevaD9wpvclMjKSxx57jGuvvZaYmBh27tzJgw8+SIcOHRg8eLCJVTd8EydO5O233+bjjz+mSZMmnn494eHhhISEEB4ezm233cbkyZOJjIwkLCyMu+66i379+vGb3/zG5OobMLOHoXmzF1980WjdurURFBRk9O3b11i1apXZJfm1ESNGGLGxsUZQUJDRsmVLY8SIEcaOHTvMLssvff311wZwyjZmzBjDMNxD4adNm2ZER0cbdrvduPTSS42tW7eaW7QfONP7UlBQYFx22WVG8+bNjcDAQKNNmzbG+PHjjfT0dLPLbvAqe08A41//+pfnmMLCQuPOO+80mjZtaoSGhhpXX321kZaWZl7RfsBiGIZR/7FLRERExDzqAyQiIiJ+RwFIRERE/I4CkIiIiPgdBSARERHxOwpAIiIi4ncUgERERMTvKACJiIiI31EAEhGpguXLl2OxWE5Zr0lEfJMCkIiIiPgdBSARERHxOwpAIuITXC4XM2fOpG3btoSEhNCzZ08WLlwIHL88tXjxYnr06EFwcDC/+c1v2LhxY4VzfPDBB3Tr1g273U5CQgLPP/98hfuLi4v505/+RHx8PHa7nQ4dOvD6669XOGbt2rX06dOH0NBQ+vfvz9atW+v2hYtInVAAEhGfMHPmTN58803mzp3Lpk2buO+++7j55ptZsWKF55gHHniA559/nh9++IHmzZszdOhQSktLAXdwueGGGxg5ciQbNmzg0UcfZdq0acyfP9/z+NGjR/POO+/wwgsvkJqayquvvkrjxo0r1PHII4/w/PPP8+OPPxIQEMCtt95aL69fRGqXFkMVEa9XXFxMZGQkX375Jf369fPsv/322ykoKOCOO+7gkksu4d1332XEiBEAHDlyhFatWjF//nxuuOEGRo0axaFDh/jiiy88j3/wwQdZvHgxmzZtYtu2bXTu3Jlly5aRnJx8Sg3Lly/nkksu4csvv+TSSy8FYMmSJVx55ZUUFhYSHBxcxz8FEalNagESEa+3Y8cOCgoKGDRoEI0bN/Zsb775Jjt37vQcd2I4ioyMpHPnzqSmpgKQmprKgAEDKpx3wIABbN++HafTyc8//4zNZuN3v/vdGWvp0aOH53ZsbCwAmZmZNX6NIlK/AswuQETkbPLy8gBYvHgxLVu2rHCf3W6vEIKqKyQkpErHBQYGem5bLBbA3T9JRHyLWoBExOslJiZit9vZu3cvHTp0qLDFx8d7jlu1apXn9tGjR9m2bRtdu3YFoGvXrnz33XcVzvvdd9/RqVMnbDYb3bt3x+VyVehTJCINl1qARMTrNWnShClTpnDffffhcrkYOHAgOTk5fPfdd4SFhdGmTRsAHn/8cZo1a0Z0dDSPPPIIUVFRDB8+HID777+fCy+8kCeeeIIRI0awcuVKXnrpJV5++WUAEhISGDNmDLfeeisvvPACPXv25NdffyUzM5MbbrjBrJcuInVEAUhEfMITTzxB8+bNmTlzJrt27SIiIoILLriAhx9+2HMJ6umnn+aee+5h+/bt9OrVi08//ZSgoCAALrjgAt577z2mT5/OE088QWxsLI8//jhjx471PMcrr7zCww8/zJ133klWVhatW7fm4YcfNuPlikgd0ygwEfF55SO0jh49SkREhNnliIgPUB8gERER8TsKQCIiIuJ3dAlMRERE/I5agERERMTvKACJiIiI31EAEhEREb+jACQiIiJ+RwFIRERE/I4CkIiIiPgdBSARERHxOwpAIiIi4ncUgERERMTv/D9nymjTOQ1SYwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# draw Losses\n",
    "if is_graphic:\n",
    "    plt.plot(list(range(1, len(arr_train_loss) + 1)), arr_train_loss, label=\"train\")\n",
    "    plt.plot(list(range(1, len(arr_val_loss) + 1)), arr_val_loss, label=\"val\")\n",
    "\n",
    "    plt.title(\"Loss\")\n",
    "    plt.xlabel(\"epoch\")\n",
    "    plt.ylabel(\"loss\")\n",
    "    plt.legend(loc=\"upper left\")\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "8bda887e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# draw mae\n",
    "if is_graphic:\n",
    "    plt.plot(list(range(1, len(arr_train_mae) + 1)), arr_train_mae, label=\"train\")\n",
    "    plt.plot(list(range(1, len(arr_val_mae) + 1)), arr_val_mae, label=\"val\")\n",
    "\n",
    "    plt.title(\"MAE\")\n",
    "    plt.xlabel(\"epoch\")\n",
    "    plt.ylabel(\"mae\")\n",
    "    plt.legend(loc=\"upper right\")\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1cc21154",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# draw Accuracies\n",
    "if is_graphic:\n",
    "    plt.plot(list(range(1, len(arr_train_acc) + 1)), arr_train_acc, label=\"train\")\n",
    "    plt.plot(list(range(1, len(arr_val_acc) + 1)), arr_val_acc, label=\"val\")\n",
    "\n",
    "    plt.title(\"Accuracy\")\n",
    "    plt.xlabel(\"epoch\")\n",
    "    plt.ylabel(\"accuracy\")\n",
    "    plt.legend(loc=\"upper left\")\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0133ff45",
   "metadata": {},
   "source": [
    "# Eval model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "e095b271",
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils_french.split_all_docs import split_all_docs\n",
    "from rouge_score import rouge_scorer\n",
    "from utils_french.threshold_probs import threshold_probs_by_prop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "273c6eb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = RNN_RNN(device=device, vocab_size=vocab_size, word_embed=embmgr.getEmbeddings(), dim_word_embed=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4fc41994",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_epoch = np.argmin(np.array(arr_val_loss)) + 1\n",
    "model.load(checkpoints_folder + \"/\" + model_name + \"-\" + str(best_epoch) + \".pt\")\n",
    "model.to(device)\n",
    "model.eval()\n",
    "best_epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "16b5ecbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_set = df_test\n",
    "test_set[\"doc_splited\"] = split_all_docs(test_set[\"flat_contents\"], False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "bb8ab07a",
   "metadata": {},
   "outputs": [],
   "source": [
    "scorer = rouge_scorer.RougeScorer(['rouge1', 'rouge2', 'rougeL'], use_stemmer=True)\n",
    "criterion = nn.BCELoss()\n",
    "mae_fn = nn.L1Loss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "67659b8a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Eval model: 100%|██████████| 100/100 [00:01<00:00, 70.55batch/s, accuracy=0.754, mae=0.333, rouge1=0.262, rouge2=0.08, rougeL=0.163]  \n"
     ]
    }
   ],
   "source": [
    "arr_rouge1 = []\n",
    "arr_rouge2 = []\n",
    "arr_rougeL = []\n",
    "arr_loss = []\n",
    "mae = []\n",
    "accuracy = []\n",
    "\n",
    "idx = 0\n",
    "with tqdm(test_iter, unit=\"batch\", total=len(test_iter)) as tepoch:\n",
    "    for batch in tepoch:\n",
    "        tepoch.set_description(\"Eval model\")\n",
    "        features = []\n",
    "        doc_lens = []\n",
    "\n",
    "        for j in range(1):\n",
    "            doc_lens.append(len(batch[j][\"docs\"]))\n",
    "            features = features + batch[j][\"docs\"]\n",
    "                \n",
    "        features = torch.tensor(features, dtype=torch.long).to(device)\n",
    "\n",
    "        targets = [torch.tensor(batch[j][\"labels\"], dtype=torch.float) for j in range(1)]\n",
    "        targets = torch.cat(targets)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        y_pred = model(features, doc_lens)\n",
    "    \n",
    "        loss = criterion(y_pred, targets)\n",
    "\n",
    "        arr_loss.append(loss.item())\n",
    "\n",
    "        doc = test_set[\"doc_splited\"].iloc[idx]\n",
    "        summaries = test_set[\"summaries\"].iloc[idx]\n",
    "\n",
    "        probs = np.array(y_pred.tolist()) # compute_probs(y_pred)\n",
    "        probs_tensor = y_pred #torch.tensor(probs, dtype=torch.float).to(device)\n",
    "\n",
    "        #probs = threshold_probs_by_nb(probs=probs, doc_lens=[probs.shape[0]], average_number_of_sentences_per_document=average_number_of_sentences_per_document)\n",
    "        #probs = threshold_probs_by_prop(probs=probs, doc_lens=[probs.shape[0]], average_proportion_of_sentences_per_document=average_proportion_of_sentences_per_document)\n",
    "        indices = torch.argsort(y_pred, descending=True)\n",
    "\n",
    "        if indices.shape[0] <= len(doc):\n",
    "            y_pred_thresh = []\n",
    "\n",
    "            txt = \"\"\n",
    "\n",
    "            for i in range(min(len(doc), y_pred.shape[0])):\n",
    "                txt = txt + \". \" + doc[indices[i]]\n",
    "                y_pred_thresh.append(indices[i])\n",
    "                if len(txt) >= len(summaries):\n",
    "                    break\n",
    "\n",
    "            y_pred_thresh.sort()\n",
    "\n",
    "            txt = \"\"\n",
    "\n",
    "            for i in y_pred_thresh:#range(min(len(doc), y_pred.shape[0])):\n",
    "                txt = txt + \". \" + doc[i]\n",
    "\n",
    "            n = min(len(txt), len(summaries))\n",
    "\n",
    "            while n < len(txt) and txt[n].isalnum():\n",
    "                n += 1\n",
    "\n",
    "            txt = txt[:n]\n",
    "\n",
    "            #assert len(txt) - len(summaries) <= 20\n",
    "\n",
    "            scores = scorer.score(summaries, txt)\n",
    "            arr_rouge1.append(scores[\"rouge1\"].recall)\n",
    "            arr_rouge2.append(scores[\"rouge2\"].recall)\n",
    "            arr_rougeL.append(scores[\"rougeL\"].recall)\n",
    "\n",
    "            mae.append(mae_fn(probs_tensor, targets).item())\n",
    "\n",
    "            accuracy.append(accuracy_nb_sent_per_doc_fn(probs=probs, targets=targets.cpu().detach().numpy(), doc_lens=[len(probs)], average_number_of_sentences_per_document=average_number_of_sentences_per_document))\n",
    "            #accuracy.append(accuracy_prop_sent_per_doc_fn(probs=probs, targets=targets.cpu().detach().numpy(), doc_lens=[len(probs)], average_proportion_of_sentences_per_document=average_proportion_of_sentences_per_document))\n",
    "\n",
    "        tepoch.set_postfix(accuracy = np.mean(accuracy), mae = np.mean(mae), rouge1 = np.mean(arr_rouge1), rouge2 = np.mean(arr_rouge2), rougeL = np.mean(arr_rougeL))\n",
    "        idx += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "5c1fd15f",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_metrics = {}\n",
    "test_metrics[\"accuracy\"] = np.mean(accuracy)\n",
    "test_metrics[\"mae\"]      = np.mean(mae)\n",
    "test_metrics[\"rouge1\"]   = np.mean(arr_rouge1)\n",
    "test_metrics[\"rouge2\"]   = np.mean(arr_rouge2)\n",
    "test_metrics[\"rougeL\"]   = np.mean(arr_rougeL)\n",
    "\n",
    "# Save to file in JSON format\n",
    "\n",
    "with open(checkpoints_folder + \"/test_metrics.json\", 'w') as fp:\n",
    "    json.dump(test_metrics, fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e4f4796c",
   "metadata": {},
   "outputs": [
    {
     "ename": "SystemExit",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "An exception has occurred, use %tb to see the full traceback.\n",
      "\u001b[0;31mSystemExit\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/raoufdine/anaconda3/lib/python3.9/site-packages/IPython/core/interactiveshell.py:3465: UserWarning: To exit: use 'exit', 'quit', or Ctrl-D.\n",
      "  warn(\"To exit: use 'exit', 'quit', or Ctrl-D.\", stacklevel=1)\n"
     ]
    }
   ],
   "source": [
    "# Does not excute this cell if you want to execute the following cells.\n",
    "import sys\n",
    "sys.exit()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.9.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "4f810f0bde7d7b5dc90f9e6832822547e13e234c53731a97fc91254323942f40"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}