{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deep learning NER\n",
    "\n",
    "In this notebook, we will discover two deep learning techniques for Named Entity Recognition (or NER).\n",
    "\n",
    "- LSTM (Long Short Term Memory)\n",
    "- Transformers\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The generated **BIO** file by the script will be loaded as well as a corpus of lambda sentences that don't necessarily contain either an _arrival_ or _departure_ location.\n",
    "\n",
    "Since the goal is to extract the locations from a user's query, the times in which random sentences might be inputted should be accounted for. Therefore, to stay rational with the frequency of that happening, the ratio will be $1:10$ for the lambda sentences which will be extracted from `data/french_text/1k_unlabeled_samples.txt`. In addition, more complex sentences where the **departure** and **arrival** locations are not in the same sentence. Their ratio will be the same as the unlabeled sentences.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sentence Loading and Preprocessing\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "os.environ[\"TF_USE_LEGACY_KERAS\"] = \"0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt_tab to /Users/az-r-\n",
      "[nltk_data]     ow/nltk_data...\n",
      "[nltk_data]   Package punkt_tab is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "from app.travel_resolver.libs.nlp import data_processing as dp\n",
    "\n",
    "sentences, labels, vocab, unique_labels = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/10k_train_small_samples.bio\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "lambda_sentences, lambda_labels, _, __ = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/1k_train_unlabeled_samples.bio\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "large_sentences, large_labels, _, __ = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/1k_train_large_samples.bio\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "sentences = sentences + lambda_sentences + large_sentences\n",
    "labels = labels + lambda_labels + large_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_sentences_and_labels(\n",
    "    sentences,\n",
    "    labels,\n",
    "    rm_stopwords: bool = False,\n",
    "    stemming: bool = True,\n",
    "    return_tokens: bool = False,\n",
    "):\n",
    "    \"\"\"\n",
    "    Process the sentences and labels using the process_sentence function from the data_processing module.\n",
    "\n",
    "    Args:\n",
    "    sentences (list): List of sentences to process.\n",
    "    labels (list): List of labels to process.\n",
    "    rm_stopwords (bool): Whether to remove stopwords from the sentences.\n",
    "    stemming (bool): Whether to apply stemming to the sentences.\n",
    "    return_tokens (bool): Whether to return the tokens of the sentences.\n",
    "\n",
    "    Returns:\n",
    "    processed_sentences (list): List of processed sentences.\n",
    "    processed_labels (list): List of processed labels.\n",
    "    \"\"\"\n",
    "    processed_sentences = []\n",
    "    processed_labels = []\n",
    "\n",
    "    for sentence, label in zip(sentences, labels):\n",
    "        sentence, label = dp.process_sentence(\n",
    "            sentence,\n",
    "            labels_to_adapt=label,\n",
    "            rm_stopwords=rm_stopwords,\n",
    "            stemming=stemming,\n",
    "            return_tokens=return_tokens,\n",
    "        )\n",
    "        processed_sentences.append(sentence)\n",
    "        processed_labels.append(label)\n",
    "\n",
    "    return processed_sentences, processed_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_sentences, processed_labels = process_sentences_and_labels(\n",
    "    sentences, labels, return_tokens=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.hist([len(sentence) for sentence in processed_sentences], bins=50)\n",
    "plt.title(\"Histogram of sentence lengths\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "  This variable will control the maximum length of the sentence \n",
    "  as well as the embedding size\n",
    "\"\"\"\n",
    "\n",
    "MAX_LEN = 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stanza Pipeline for POS\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning:\n",
      "\n",
      "IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "\n",
      "2025-01-05 14:29:26 INFO: Checking for updates to resources.json in case models have been updated.  Note: this behavior can be turned off with download_method=None or download_method=DownloadMethod.REUSE_RESOURCES\n",
      "Downloading https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.9.0.json: 392kB [00:00, 103MB/s]                     \n",
      "2025-01-05 14:29:26 INFO: Downloaded file to /Users/az-r-ow/stanza_resources/resources.json\n",
      "2025-01-05 14:29:26 WARNING: Language fr package default expects mwt, which has been added\n",
      "2025-01-05 14:29:27 INFO: Loading these models for language: fr (French):\n",
      "===============================\n",
      "| Processor | Package         |\n",
      "-------------------------------\n",
      "| tokenize  | combined        |\n",
      "| mwt       | combined        |\n",
      "| pos       | combined_charlm |\n",
      "===============================\n",
      "\n",
      "2025-01-05 14:29:27 WARNING: GPU requested, but is not available!\n",
      "2025-01-05 14:29:27 INFO: Using device: cpu\n",
      "2025-01-05 14:29:27 INFO: Loading: tokenize\n",
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/stanza/models/tokenization/trainer.py:82: FutureWarning:\n",
      "\n",
      "You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "\n",
      "2025-01-05 14:29:27 INFO: Loading: mwt\n",
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/stanza/models/mwt/trainer.py:201: FutureWarning:\n",
      "\n",
      "You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "\n",
      "2025-01-05 14:29:27 INFO: Loading: pos\n",
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/stanza/models/pos/trainer.py:139: FutureWarning:\n",
      "\n",
      "You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "\n",
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/stanza/models/common/pretrain.py:56: FutureWarning:\n",
      "\n",
      "You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "\n",
      "/Users/az-r-ow/Developer/TravelOrderResolver/venv/lib/python3.12/site-packages/stanza/models/common/char_model.py:271: FutureWarning:\n",
      "\n",
      "You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "\n",
      "2025-01-05 14:29:28 INFO: Done loading processors!\n"
     ]
    }
   ],
   "source": [
    "import stanza\n",
    "\n",
    "nlp = stanza.Pipeline(\n",
    "    \"fr\", processors=\"tokenize,pos\", use_gpu=True, pos_batch_size=3000\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "docs = [stanza.Document([], text=sentence) for sentence in sentences]\n",
    "\n",
    "pos_tags_docs = nlp(docs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_sentences_pos_tags(sentences: list[str]):\n",
    "    \"\"\"\n",
    "    Get the POS tags of the words in the sentences.\n",
    "\n",
    "    Args:\n",
    "    sentences (list): List of sentences to get the POS tags from.\n",
    "\n",
    "    Returns:\n",
    "    pos_tags (list): List of POS tags for each sentence.\n",
    "    \"\"\"\n",
    "    docs = [stanza.Document([], text=sentence) for sentence in sentences]\n",
    "    pos_tags_docs = nlp(docs)\n",
    "\n",
    "    sentences_pos_tags = []\n",
    "    for doc in pos_tags_docs:\n",
    "        sentences_pos_tags.append([word.upos for word in doc.sentences[0].words])\n",
    "\n",
    "    return sentences_pos_tags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "pos_tags_sentences = []\n",
    "\n",
    "for doc in pos_tags_docs:\n",
    "    pos_tags_sentences.append([word.upos for word in doc.sentences[0].words])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_pos_tags = []\n",
    "\n",
    "for doc in pos_tags_docs:\n",
    "    pos_tags = [word.upos for sent in doc.sentences for word in sent.words]\n",
    "    all_pos_tags.extend(pos_tags)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ADJ',\n",
       " 'ADP',\n",
       " 'ADV',\n",
       " 'AUX',\n",
       " 'CCONJ',\n",
       " 'DET',\n",
       " 'INTJ',\n",
       " 'NOUN',\n",
       " 'NUM',\n",
       " 'PRON',\n",
       " 'PROPN',\n",
       " 'PUNCT',\n",
       " 'SCONJ',\n",
       " 'SYM',\n",
       " 'VERB',\n",
       " 'X']"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_pos_tags = sorted(set(all_pos_tags))\n",
    "\n",
    "unique_pos_tags"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generating inputs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "\n",
    "def encode_and_pad_sentence(sentence: str, vocab: list[str], max_length: int = MAX_LEN):\n",
    "    \"\"\"\n",
    "    Encode a sentence into a list of integers\n",
    "\n",
    "    Parameters:\n",
    "    sentence (str): The sentence to encode\n",
    "    vocab (list): The vocabulary\n",
    "\n",
    "    Returns:\n",
    "    list: The list of integers\n",
    "    \"\"\"\n",
    "    encoded_sentence = [\n",
    "        vocab.index(word) if word in vocab else vocab.index(\"<UNK>\")\n",
    "        for word in sentence\n",
    "    ]\n",
    "\n",
    "    return tf.keras.utils.pad_sequences(\n",
    "        [encoded_sentence], maxlen=max_length, padding=\"post\", value=0\n",
    "    )[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def encode_and_pad_sentence_pos(\n",
    "    sentence_pos: str, pos_tags: list[str] = unique_pos_tags, max_length: int = MAX_LEN\n",
    "):\n",
    "    \"\"\"\n",
    "    Encode a sentence into a list of integers\n",
    "\n",
    "    Parameters:\n",
    "    sentence (str): The sentence to encode\n",
    "    pos_tags (list): The vocabulary\n",
    "\n",
    "    Returns:\n",
    "    list: The list of integers\n",
    "    \"\"\"\n",
    "    encoded_sentence = [pos_tags.index(pos) for pos in sentence_pos]\n",
    "\n",
    "    return tf.keras.utils.pad_sequences(\n",
    "        [encoded_sentence], maxlen=max_length, padding=\"post\", value=0\n",
    "    )[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "get_vocab_from_corpus = lambda corpus: [\"\", \"<UNK>\"] + list(\n",
    "    set([word for sentence in corpus for word in sentence])\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab = get_vocab_from_corpus(processed_sentences)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded_sentences = [\n",
    "    encode_and_pad_sentence(sentence, vocab) for sentence in processed_sentences\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "pos_tags_sentences = [\n",
    "    encode_and_pad_sentence_pos(sentence_pos) for sentence_pos in pos_tags_sentences\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "padded_labels = tf.keras.preprocessing.sequence.pad_sequences(\n",
    "    processed_labels, maxlen=MAX_LEN, padding=\"post\", value=-1\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2025-01-05 14:33:50.868923: I tensorflow/core/framework/local_rendezvous.cc:405] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence\n"
     ]
    }
   ],
   "source": [
    "dataset = tf.data.Dataset.from_tensor_slices((encoded_sentences, padded_labels))\n",
    "\n",
    "dataset = dataset.shuffle(len(encoded_sentences), seed=42)\n",
    "\n",
    "# Split the dataset into a training and testing dataset\n",
    "train_dataset, test_dataset = tf.keras.utils.split_dataset(dataset, left_size=0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2025-01-05 14:31:08.581526: I tensorflow/core/framework/local_rendezvous.cc:405] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence\n"
     ]
    }
   ],
   "source": [
    "pos_dataset = tf.data.Dataset.from_tensor_slices(\n",
    "    ((encoded_sentences, pos_tags_sentences), padded_labels)\n",
    ")\n",
    "\n",
    "\n",
    "pos_dataset = pos_dataset.shuffle(len(encoded_sentences), seed=42)\n",
    "\n",
    "pos_train_dataset, pos_test_dataset = tf.keras.utils.split_dataset(\n",
    "    pos_dataset, left_size=0.8\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Masked loss and metrics\n",
    "\n",
    "Before training the model, we need to create your own function to compute the accuracy. Tensorflow has built-in accuracy metrics but we cannot pass values to be ignored. This will impact the calculations, since we must remove the padded values.\n",
    "\n",
    "Usually, the metric that inputs true labels and predicted labels and outputs how many times the predicted and true labels match is called accuracy. In some cases, however, there is one more step before getting the predicted labels. This may happen if, instead of passing the predicted labels, a vector of probabilities is passed. In such case, there is a need to perform an `argmax` for each prediction to find the appropriate predicted label. Such situations happen very often, therefore Tensorflow has a set of functions, with prefix `Sparse`, that performs this operation in the backend. Unfortunately, it does not provide values to ignore in the accuracy case. This is what you will work on now.\n",
    "\n",
    "Note that the model's prediction has 3 axes:\n",
    "\n",
    "- the number of examples (batch size)\n",
    "- the number of words in each example (padded to be as long as the longest sentence in the batch)\n",
    "- the number of possible targets (the 17 named entity tags).\n",
    "\n",
    "Another important function is the loss function. In this case, we will use the Cross Entropy loss, but we need a multiclass implementation of it, also we may look for its Sparse version. Tensorflow has a SparseCategoricalCrossentropy loss function, which it is already imported by the name SparseCategoricalCrossEntropy.\n",
    "\n",
    "SparseCategoricalCrossentropy: The Sparse Categorical Crossentropy Loss Function.\n",
    "\n",
    "The arguments you will need:\n",
    "\n",
    "1. `from_logits`: This indicates if the values are raw values or normalized values (probabilities). Since the last layer of the model finishes with a LogSoftMax call, the results are not normalized - they do not lie between 0 and 1.\n",
    "2. `ignore_class`: This indicates which class should be ignored when computing the crossentropy. Remember that the class related to padding value is set to be 0.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomSparseCategoricalCrossentropy(tf.keras.losses.Loss):\n",
    "    def __init__(self, from_logits=False, ignore_class=-1):\n",
    "        super().__init__()\n",
    "        self.from_logits = from_logits\n",
    "        self.ignore_class = ignore_class\n",
    "\n",
    "    def call(self, y_true, y_pred):\n",
    "        # Ensure inputs are tensors\n",
    "        y_true = tf.convert_to_tensor(y_true)\n",
    "        y_pred = tf.convert_to_tensor(y_pred)\n",
    "\n",
    "        # Generate a mask that is False where y_true equals ignore_class and True elsewhere\n",
    "        mask = tf.not_equal(y_true, self.ignore_class)\n",
    "\n",
    "        # Use this mask to filter out ignored values from y_true and y_pred\n",
    "        y_true_filtered = tf.boolean_mask(y_true, mask)\n",
    "        y_pred_filtered = tf.boolean_mask(y_pred, mask)\n",
    "\n",
    "        # Compute the sparse categorical crossentropy on filtered targets and predictions\n",
    "        loss = tf.keras.losses.sparse_categorical_crossentropy(\n",
    "            y_true_filtered, y_pred_filtered, from_logits=self.from_logits\n",
    "        )\n",
    "\n",
    "        # Return the mean loss value\n",
    "        return tf.reduce_mean(loss)\n",
    "\n",
    "\n",
    "def masked_loss(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    Calculate the masked sparse categorical cross-entropy loss.\n",
    "\n",
    "    Parameters:\n",
    "    y_true (tensor): True labels.\n",
    "    y_pred (tensor): Predicted logits.\n",
    "\n",
    "    Returns:\n",
    "    loss (tensor): Calculated loss.\n",
    "    \"\"\"\n",
    "\n",
    "    # Calculate the loss for each item in the batch. Remember to pass the right arguments, as discussed above!\n",
    "    loss_fn = CustomSparseCategoricalCrossentropy(from_logits=True, ignore_class=-1)\n",
    "    # Use the previous defined function to compute the loss\n",
    "    loss = loss_fn(y_true, y_pred)\n",
    "\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def masked_accuracy(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    Calculate masked accuracy for predicted labels.\n",
    "\n",
    "    Parameters:\n",
    "    y_true (tensor): True labels.\n",
    "    y_pred (tensor): Predicted logits.\n",
    "\n",
    "    Returns:\n",
    "    accuracy (tensor): Masked accuracy.\n",
    "    \"\"\"\n",
    "\n",
    "    # Calculate the loss for each item in the batch.\n",
    "    # We must always cast the tensors to the same type in order to use them in training. Since we will make divisions, it is safe to use tf.float32 data type.\n",
    "    y_true = tf.cast(y_true, tf.float32)\n",
    "    # Create the mask, i.e., the values that will be ignored\n",
    "    mask = tf.not_equal(y_true, -1.0)\n",
    "\n",
    "    mask = tf.cast(mask, tf.float32)\n",
    "\n",
    "    # Perform argmax to get the predicted values\n",
    "    y_pred_class = tf.math.argmax(y_pred, axis=-1)\n",
    "    y_pred_class = tf.cast(y_pred_class, tf.float32)\n",
    "    # Compare the true values with the predicted ones\n",
    "    matches_true_pred = tf.equal(y_true, y_pred_class)\n",
    "    matches_true_pred = tf.cast(matches_true_pred, tf.float32)\n",
    "    # Multiply the acc tensor with the masks\n",
    "    matches_true_pred *= mask\n",
    "\n",
    "    # Compute masked accuracy (quotient between the total matches and the total valid values, i.e., the amount of non-masked values)\n",
    "    masked_acc = tf.reduce_sum(matches_true_pred) / tf.reduce_sum(mask)\n",
    "\n",
    "    return masked_acc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition, we will need a function that computes the accuracy of the model when it comes to the tagged entities. This will come in handy when comparing the performance of the models on sentences with different length. Because otherwise, using the `masked_accuracy` the score will be biases (because of the `O` token).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def entity_accuracy(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    Calculate the accuracy based on the entities. Which mean that correct `O` tags will not be taken into account.\n",
    "\n",
    "    Parameters:\n",
    "    y_true (tensor): True labels.\n",
    "    y_pred (tensor): Predicted logits.\n",
    "\n",
    "    Returns:\n",
    "    accuracy (tensor): Tag accuracy.\n",
    "    \"\"\"\n",
    "\n",
    "    y_true = tf.cast(y_true, tf.float32)\n",
    "    # We ignore the padding and the O tag\n",
    "    mask = y_true > 0\n",
    "    mask = tf.cast(mask, tf.float32)\n",
    "\n",
    "    y_pred_class = tf.math.argmax(y_pred, axis=-1)\n",
    "    y_pred_class = tf.cast(y_pred_class, tf.float32)\n",
    "\n",
    "    matches_true_pred = tf.equal(y_true, y_pred_class)\n",
    "    matches_true_pred = tf.cast(matches_true_pred, tf.float32)\n",
    "\n",
    "    matches_true_pred *= mask\n",
    "\n",
    "    masked_acc = tf.reduce_sum(matches_true_pred) / tf.reduce_sum(mask)\n",
    "\n",
    "    return masked_acc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Eval Dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_sentences, eval_labels, _, __ = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/800_eval_small_samples.bio\"\n",
    ")\n",
    "\n",
    "eval_unlabeled, eval_unlabeled_labels, _, __ = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/eval_unlabeled_samples.bio\"\n",
    ")\n",
    "\n",
    "eval_large, eval_large_labels, _, __ = dp.from_bio_file_to_examples(\n",
    "    \"./data/bio/fr.bio/100_eval_large_samples.bio\"\n",
    ")\n",
    "\n",
    "eval_short_sentences, eval_short_labels = process_sentences_and_labels(\n",
    "    eval_sentences, eval_labels, return_tokens=True\n",
    ")\n",
    "eval_unlabeled_sentences, eval_unlabeled_labels = process_sentences_and_labels(\n",
    "    eval_unlabeled, eval_unlabeled_labels, return_tokens=True\n",
    ")\n",
    "eval_large_sentences, eval_large_labels = process_sentences_and_labels(\n",
    "    eval_large, eval_large_labels, return_tokens=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_short_sentences_pos = get_sentences_pos_tags(eval_sentences)\n",
    "eval_unlabeled_sentences_pos = get_sentences_pos_tags(eval_unlabeled)\n",
    "eval_large_sentences_pos = get_sentences_pos_tags(eval_large)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded_eval_short_sentences_pos = [\n",
    "    encode_and_pad_sentence_pos(sentence_pos)\n",
    "    for sentence_pos in eval_short_sentences_pos\n",
    "]\n",
    "\n",
    "encoded_eval_unlabeled_sentences_pos = [\n",
    "    encode_and_pad_sentence_pos(sentence_pos)\n",
    "    for sentence_pos in eval_unlabeled_sentences_pos\n",
    "]\n",
    "\n",
    "encoded_eval_large_sentences_pos = [\n",
    "    encode_and_pad_sentence_pos(sentence_pos)\n",
    "    for sentence_pos in eval_large_sentences_pos\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded_eval_short_sentences = [\n",
    "    encode_and_pad_sentence(sentence, vocab) for sentence in eval_short_sentences\n",
    "]\n",
    "\n",
    "encoded_eval_short_sentences = tf.convert_to_tensor(encoded_eval_short_sentences)\n",
    "\n",
    "padded_eval_short_labels = tf.keras.preprocessing.sequence.pad_sequences(\n",
    "    eval_short_labels, maxlen=MAX_LEN, padding=\"post\", value=-1\n",
    ")\n",
    "\n",
    "encoded_eval_unlabeled_sentences = [\n",
    "    encode_and_pad_sentence(sentence, vocab) for sentence in eval_unlabeled_sentences\n",
    "]\n",
    "\n",
    "encoded_eval_unlabeled_sentences = tf.convert_to_tensor(\n",
    "    encoded_eval_unlabeled_sentences\n",
    ")\n",
    "\n",
    "padded_eval_unlabeled_labels = tf.keras.preprocessing.sequence.pad_sequences(\n",
    "    eval_unlabeled_labels, maxlen=MAX_LEN, padding=\"post\", value=-1\n",
    ")\n",
    "\n",
    "encoded_eval_large_sentences = [\n",
    "    encode_and_pad_sentence(sentence, vocab) for sentence in eval_large_sentences\n",
    "]\n",
    "\n",
    "encoded_eval_large_sentences = tf.convert_to_tensor(encoded_eval_large_sentences)\n",
    "\n",
    "padded_eval_large_labels = tf.keras.preprocessing.sequence.pad_sequences(\n",
    "    eval_large_labels, maxlen=MAX_LEN, padding=\"post\", value=-1\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Custom Confusion Matrix\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "def tf_round(x, decimals=0):\n",
    "    multiplier = tf.constant(10**decimals, dtype=x.dtype)\n",
    "    return tf.round(x * multiplier) / multiplier\n",
    "\n",
    "\n",
    "def confusion_matrix(\n",
    "    y_true,\n",
    "    y_pred,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix\",\n",
    "    return_ax: bool = False,\n",
    "    ax: plt.Axes = None,\n",
    "):\n",
    "    \"\"\"\n",
    "    Calculate the confusion matrix for the model's predictions.\n",
    "\n",
    "    Parameters:\n",
    "    y_true (tensor): True labels.\n",
    "    y_pred (tensor): Predicted logits.\n",
    "    unique_labels (list): List of unique labels.\n",
    "    title (str): Title of the plot.\n",
    "    return_ax (bool): Whether to return the axis or not if not the plot will be shown.\n",
    "    \"\"\"\n",
    "\n",
    "    # Perform argmax to get the predicted values\n",
    "    y_pred_class = tf.math.argmax(y_pred, axis=-1)\n",
    "\n",
    "    print(y_pred_class.shape)\n",
    "    # Cast the tensors to float32\n",
    "    y_true = tf.cast(y_true, tf.float32)\n",
    "    y_pred_class = tf.cast(y_pred_class, tf.float32)\n",
    "    # Create the mask, i.e., the values that will be ignored\n",
    "    mask = tf.not_equal(y_true, -1.0)\n",
    "    mask2 = tf.not_equal(y_true, -100)\n",
    "\n",
    "    mask = tf.cast(mask, tf.float32)\n",
    "    mask2 = tf.cast(mask2, tf.float32)\n",
    "\n",
    "    # Multiply the true values by the mask\n",
    "    y_true *= mask * mask2\n",
    "    # Multiply the predicted values by the mask\n",
    "    y_pred_class *= mask * mask2\n",
    "\n",
    "    # Flattening to match the confusion matrix function signature\n",
    "    y_true_flat = tf.reshape(y_true, [-1])\n",
    "    y_pred_class_flat = tf.reshape(y_pred_class, [-1])\n",
    "    mask_flat = tf.reshape(mask, [-1])\n",
    "\n",
    "    # Compute the confusion matrix\n",
    "    cm = tf.math.confusion_matrix(\n",
    "        y_true_flat,\n",
    "        y_pred_class_flat,\n",
    "        num_classes=len(unique_labels.items()),\n",
    "        weights=mask_flat,\n",
    "    )\n",
    "\n",
    "    cm_normalized = tf.math.divide_no_nan(cm, tf.reduce_sum(cm, axis=1, keepdims=True))\n",
    "\n",
    "    cm_normalized = tf_round(cm_normalized, 2)\n",
    "\n",
    "    hm = sns.heatmap(cm_normalized, annot=True, cmap=\"Blues\", ax=ax)\n",
    "\n",
    "    ax = ax if ax else hm\n",
    "\n",
    "    ax.set_xticklabels(labels=unique_labels, rotation=45)\n",
    "    ax.set_yticklabels(labels=unique_labels, rotation=45)\n",
    "\n",
    "    ax.set_xlabel(\"Predicted labels\")\n",
    "    ax.set_ylabel(\"True labels\")\n",
    "    ax.set_title(title)\n",
    "\n",
    "    if return_ax:\n",
    "        return ax\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bootstrapping\n",
    "\n",
    "The bootstrapping function will resample the evaluation data with $B = 30$.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tqdm import tqdm\n",
    "from sklearn.metrics import f1_score\n",
    "import transformers\n",
    "\n",
    "bootstrap_eval_sentences = tf.concat(\n",
    "    [\n",
    "        encoded_eval_short_sentences,\n",
    "        encoded_eval_unlabeled_sentences,\n",
    "        encoded_eval_large_sentences,\n",
    "    ],\n",
    "    axis=0,\n",
    ")\n",
    "\n",
    "bootstrap_eval_sentences_pos = tf.concat(\n",
    "    [\n",
    "        encoded_eval_short_sentences_pos,\n",
    "        encoded_eval_unlabeled_sentences_pos,\n",
    "        encoded_eval_large_sentences_pos,\n",
    "    ],\n",
    "    axis=0,\n",
    ")\n",
    "\n",
    "bootstrap_eval_labels = tf.concat(\n",
    "    [padded_eval_short_labels, padded_eval_unlabeled_labels, padded_eval_large_labels],\n",
    "    axis=0,\n",
    ")\n",
    "\n",
    "\n",
    "def bootstrap_evaluation(\n",
    "    model,\n",
    "    sentences=bootstrap_eval_sentences,\n",
    "    pos_sentences=None,\n",
    "    labels=bootstrap_eval_labels,\n",
    "    num_bootstrap_samples=30,\n",
    "    from_logits=False,\n",
    "    has_mask=False,\n",
    "):\n",
    "    \"\"\"\n",
    "    Perform bootstrapping on the evaluation dataset and calculate both accuracy and entity accuracy.\n",
    "\n",
    "    Parameters:\n",
    "    model (tf.keras.Model): The trained model.\n",
    "    sentences (list): List of sentences to evaluate.\n",
    "    labels (list): List of true labels corresponding to the sentences.\n",
    "    vocab (list): The vocabulary.\n",
    "    unique_labels (dict): Dictionary of unique labels.\n",
    "    max_len (int): The maximum length of the sentences.\n",
    "    num_bootstrap_samples (int): Number of bootstrap samples to generate.\n",
    "    from_logits (bool): If the predictions are logits\n",
    "    has_mask (bool): Set to true if the inputs contains masked values (-100)\n",
    "\n",
    "    Returns:\n",
    "    dict: Dictionary containing the accuracy and entity accuracy for each bootstrap sample.\n",
    "    \"\"\"\n",
    "    accuracies = []\n",
    "    entity_accuracies = []\n",
    "    f1_scores = []\n",
    "    num_sentences = None\n",
    "\n",
    "    if isinstance(sentences, transformers.tokenization_utils_base.BatchEncoding):\n",
    "        num_sentences = sentences[\"input_ids\"].shape[0]\n",
    "    else:\n",
    "        num_sentences = tf.shape(sentences)[0]\n",
    "\n",
    "    for _ in tqdm(range(num_bootstrap_samples)):\n",
    "        indices = tf.random.uniform(\n",
    "            [num_sentences], maxval=num_sentences, dtype=tf.int32\n",
    "        )\n",
    "\n",
    "        if isinstance(sentences, transformers.tokenization_utils_base.BatchEncoding):\n",
    "            sampled_input_ids = tf.gather(sentences[\"input_ids\"], indices)\n",
    "            sampled_attention_mask = tf.gather(sentences[\"attention_mask\"], indices)\n",
    "            sampled_sentences = dict()\n",
    "            sampled_sentences[\"input_ids\"] = sampled_input_ids\n",
    "            sampled_sentences[\"attention_mask\"] = sampled_attention_mask\n",
    "        else:\n",
    "            sampled_sentences = tf.gather(sentences, indices)\n",
    "\n",
    "        sampled_labels = tf.gather(labels, indices)\n",
    "\n",
    "        if pos_sentences is not None:\n",
    "            sampled_pos_sentences = tf.gather(pos_sentences, indices)\n",
    "            sampled_sentences = [sampled_sentences, sampled_pos_sentences]\n",
    "\n",
    "        predictions = model.predict(sampled_sentences, verbose=0)\n",
    "\n",
    "        if from_logits:\n",
    "            predictions = predictions.logits\n",
    "\n",
    "        acc = masked_accuracy(sampled_labels, predictions).numpy()\n",
    "        entity_acc = entity_accuracy(sampled_labels, predictions).numpy()\n",
    "\n",
    "        positive_labels = np.unique(sampled_labels.numpy().flatten())\n",
    "        positive_labels = positive_labels[positive_labels > 0]\n",
    "\n",
    "        predictions = tf.math.argmax(predictions, axis=-1)\n",
    "\n",
    "        if has_mask:\n",
    "            mask = tf.not_equal(sampled_labels, -100)\n",
    "            mask_value = tf.cast(-100, dtype=predictions.dtype)\n",
    "            predictions = tf.where(mask, predictions, mask_value)\n",
    "\n",
    "        f_1 = f1_score(\n",
    "            sampled_labels.numpy().flatten(),\n",
    "            predictions.numpy().flatten(),\n",
    "            average=\"micro\",\n",
    "            labels=positive_labels,\n",
    "        )\n",
    "\n",
    "        accuracies.append(acc)\n",
    "        entity_accuracies.append(entity_acc)\n",
    "        f1_scores.append(f_1)\n",
    "\n",
    "    return {\n",
    "        \"accuracies\": accuracies,\n",
    "        \"entity_accuracies\": entity_accuracies,\n",
    "        \"f1_scores\": f1_scores,\n",
    "    }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LSTM with POS\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext tensorboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_input = tf.keras.layers.Input(shape=(MAX_LEN,), name=\"word_input\")\n",
    "pos_input = tf.keras.layers.Input(shape=(MAX_LEN,), name=\"pos_input\")\n",
    "\n",
    "mask = tf.keras.layers.Masking(mask_value=0)(word_input)\n",
    "\n",
    "emb_size = 32\n",
    "\n",
    "word_embedding = tf.keras.layers.Embedding(len(vocab), emb_size, name=\"word_embedding\")(\n",
    "    word_input\n",
    ")\n",
    "\n",
    "pos_embedding = tf.keras.layers.Embedding(\n",
    "    len(unique_pos_tags),\n",
    "    emb_size,\n",
    "    name=\"pos_embedding\",\n",
    ")(pos_input)\n",
    "\n",
    "concatenated = tf.keras.layers.Concatenate()([word_embedding, pos_embedding])\n",
    "\n",
    "masked_cat = tf.keras.layers.Masking(mask_value=0)(concatenated)\n",
    "\n",
    "lstm_layer_with_pos = tf.keras.layers.LSTM(\n",
    "    emb_size, return_sequences=True, name=\"lstm_layer\"\n",
    ")(masked_cat)\n",
    "\n",
    "dropout = tf.keras.layers.Dropout(0.2)(lstm_layer_with_pos)\n",
    "\n",
    "output = tf.keras.layers.Dense(len(unique_labels), activation=tf.nn.log_softmax)(\n",
    "    dropout\n",
    ")\n",
    "\n",
    "lstm_with_pos = tf.keras.Model(inputs=[word_input, pos_input], outputs=output)\n",
    "\n",
    "lstm_with_pos.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(0.01),\n",
    "    loss=masked_loss,\n",
    "    metrics=[entity_accuracy],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m297/297\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 22ms/step - entity_accuracy: 0.7112 - loss: 0.1691 - val_entity_accuracy: 0.9696 - val_loss: 0.0141\n",
      "Epoch 2/10\n",
      "\u001b[1m297/297\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 21ms/step - entity_accuracy: 0.9678 - loss: 0.0168 - val_entity_accuracy: 0.9814 - val_loss: 0.0090\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x2dc3c53d0>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lstm_with_pos.fit(\n",
    "    pos_train_dataset.batch(32),\n",
    "    validation_data=pos_test_dataset.batch(32),\n",
    "    epochs=10,\n",
    "    callbacks=[\n",
    "        tf.keras.callbacks.EarlyStopping(\n",
    "            monitor=\"val_loss\", min_delta=0.01, restore_best_weights=True\n",
    "        ),\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded_eval_short_sentences_pos_array = np.array(encoded_eval_short_sentences_pos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m24/24\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step\n",
      "(758, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lstm_with_pos_eval_short_predictions = lstm_with_pos.predict(\n",
    "    [encoded_eval_short_sentences, encoded_eval_short_sentences_pos_array]\n",
    ")\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_short_labels,\n",
    "    lstm_with_pos_eval_short_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix on short sentences\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 8ms/step \n",
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "encoded_eval_unlabeled_sentences_pos_array = np.array(\n",
    "    encoded_eval_unlabeled_sentences_pos\n",
    ")\n",
    "\n",
    "lstm_with_pos_eval_unlabeled_predictions = lstm_with_pos.predict(\n",
    "    [encoded_eval_unlabeled_sentences, encoded_eval_unlabeled_sentences_pos_array]\n",
    ")\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_unlabeled_labels,\n",
    "    lstm_with_pos_eval_unlabeled_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix on unlabeled sentences\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m3/3\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 9ms/step \n",
      "(96, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "encoded_eval_large_sentences_pos_array = np.array(encoded_eval_large_sentences_pos)\n",
    "\n",
    "lstm_with_pos_eval_large_predictions = lstm_with_pos.predict(\n",
    "    [encoded_eval_large_sentences, encoded_eval_large_sentences_pos_array]\n",
    ")\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_large_labels,\n",
    "    lstm_with_pos_eval_large_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix on large sentences\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [00:05<00:00,  5.64it/s]\n"
     ]
    }
   ],
   "source": [
    "lstm_with_pos_results = bootstrap_evaluation(\n",
    "    lstm_with_pos, pos_sentences=bootstrap_eval_sentences_pos\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.8164736164736165,\n",
       " 0.7985812009120852,\n",
       " 0.784516765285996,\n",
       " 0.8030188679245283,\n",
       " 0.8226950354609929,\n",
       " 0.8065573770491803,\n",
       " 0.8091872791519434,\n",
       " 0.8013398608606029,\n",
       " 0.7972406996797241,\n",
       " 0.801345059493016,\n",
       " 0.8103008669046405,\n",
       " 0.820815175596001,\n",
       " 0.802,\n",
       " 0.8129829984544049,\n",
       " 0.8045859872611465,\n",
       " 0.8225765628954695,\n",
       " 0.8179299924452279,\n",
       " 0.8139656070870245,\n",
       " 0.8120261437908497,\n",
       " 0.8071500503524672,\n",
       " 0.8032786885245902,\n",
       " 0.8046324269889225,\n",
       " 0.8132375577219086,\n",
       " 0.8171772978402813,\n",
       " 0.8084672277480235,\n",
       " 0.8151873271310033,\n",
       " 0.8107396683337721,\n",
       " 0.8050568900126422,\n",
       " 0.8066098631551769,\n",
       " 0.8025316455696202]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lstm_with_pos_results[\"f1_scores\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "with open(\"lstm_with_pos_results.pickle\", \"wb\") as f:\n",
    "    pickle.dump(lstm_with_pos_results, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LSTM without POS\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm = tf.keras.models.Sequential(\n",
    "    layers=[\n",
    "        tf.keras.layers.Embedding(len(vocab) + 1, MAX_LEN, mask_zero=True),\n",
    "        tf.keras.layers.LSTM(MAX_LEN, return_sequences=True),\n",
    "        tf.keras.layers.Dropout(0.2),\n",
    "        tf.keras.layers.Dense(len(unique_labels), activation=tf.nn.log_softmax),\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(0.01),\n",
    "    loss=masked_loss,\n",
    "    metrics=[entity_accuracy],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m149/149\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 67ms/step - entity_accuracy: 0.7096 - loss: 0.1811 - val_entity_accuracy: 0.9831 - val_loss: 0.0105\n",
      "Epoch 2/10\n",
      "\u001b[1m149/149\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 66ms/step - entity_accuracy: 0.9802 - loss: 0.0090 - val_entity_accuracy: 0.9864 - val_loss: 0.0084\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x2dc58f200>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lstm.fit(\n",
    "    train_dataset.batch(64),\n",
    "    validation_data=test_dataset.batch(64),\n",
    "    epochs=10,\n",
    "    shuffle=True,\n",
    "    callbacks=[\n",
    "        tf.keras.callbacks.EarlyStopping(\n",
    "            monitor=\"val_loss\", min_delta=0.01, restore_best_weights=True\n",
    "        )\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2368/2368\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m13s\u001b[0m 5ms/step\n"
     ]
    }
   ],
   "source": [
    "test_predictions = lstm.predict(test_dataset.batch(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m24/24\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step\n",
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step\n",
      "\u001b[1m3/3\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step\n"
     ]
    }
   ],
   "source": [
    "eval_short_predictions = lstm.predict(encoded_eval_short_sentences)\n",
    "eval_unlabeled_predictions = lstm.predict(encoded_eval_unlabeled_sentences)\n",
    "eval_large_predictions = lstm.predict(encoded_eval_large_sentences)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Eval short sentences accuracy:  0.9421664\n",
      "Eval unlabeled sentences accuracy:  0.9981966\n",
      "Eval large sentences accuracy:  0.97053784\n"
     ]
    }
   ],
   "source": [
    "print(\n",
    "    \"Eval short sentences accuracy: \",\n",
    "    masked_accuracy(padded_eval_short_labels, eval_short_predictions).numpy(),\n",
    ")\n",
    "print(\n",
    "    \"Eval unlabeled sentences accuracy: \",\n",
    "    masked_accuracy(padded_eval_unlabeled_labels, eval_unlabeled_predictions).numpy(),\n",
    ")\n",
    "print(\n",
    "    \"Eval large sentences accuracy: \",\n",
    "    masked_accuracy(padded_eval_large_labels, eval_large_predictions).numpy(),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The accuracy of the `large_sentences` corpus is much higher than the `short_sentences` but that is influenced by the **higher** number of `O` tags and doesn't reflect the actual accuracy of **entities recognition**.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Eval short sentences tag accuracy:  0.738223\n",
      "Eval large sentences tag accuracy:  0.6543779\n"
     ]
    }
   ],
   "source": [
    "print(\n",
    "    \"Eval short sentences tag accuracy: \",\n",
    "    entity_accuracy(padded_eval_short_labels, eval_short_predictions).numpy(),\n",
    ")\n",
    "print(\n",
    "    \"Eval large sentences tag accuracy: \",\n",
    "    entity_accuracy(padded_eval_large_labels, eval_large_predictions).numpy(),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(758, 100), dtype=int64, numpy=\n",
       "array([[0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       ...,\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0]])>"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.math.argmax(eval_short_predictions, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0,  0,  0, ..., -1, -1, -1], dtype=int32)"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "padded_eval_short_labels.flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7795969773299748"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "\n",
    "f1_score(\n",
    "    padded_eval_short_labels.flatten(),\n",
    "    tf.math.argmax(eval_short_predictions, axis=-1).numpy().flatten(),\n",
    "    average=\"micro\",\n",
    "    labels=[1, 2],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(758, 100)\n",
      "(100, 100)\n",
      "(96, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x3000 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(3, 1, figsize=(10, 30))\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_short_labels,\n",
    "    eval_short_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for LSTM on evaluation (short sentences) dataset\",\n",
    "    return_ax=True,\n",
    "    ax=axs[0],\n",
    ")\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_unlabeled_labels,\n",
    "    eval_unlabeled_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for LSTM on evaluation (unlabeled sentences) dataset\",\n",
    "    return_ax=True,\n",
    "    ax=axs[1],\n",
    ")\n",
    "\n",
    "confusion_matrix(\n",
    "    padded_eval_large_labels,\n",
    "    eval_large_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for LSTM on evaluation (large sentences) dataset\",\n",
    "    return_ax=True,\n",
    "    ax=axs[2],\n",
    ")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(\n",
    "    sentence: str,\n",
    "    model: tf.keras.Model,\n",
    "    vocab: list[str] = vocab,\n",
    "    max_len: int = MAX_LEN,\n",
    "):\n",
    "    \"\"\"\n",
    "    Predict the labels for a given sentence.\n",
    "\n",
    "    Parameters:\n",
    "    sentence (str): The sentence to predict.\n",
    "    model (tf.keras.Model): The trained model.\n",
    "    vocab (list): The vocabulary.\n",
    "    max_len (int): The maximum length of the sentence.\n",
    "\n",
    "    Returns:\n",
    "    list: The predicted labels.\n",
    "    \"\"\"\n",
    "    processed_sentence = dp.process_sentence(\n",
    "        sentence, stemming=True, return_tokens=True\n",
    "    )\n",
    "\n",
    "    unstemmed_sentence = dp.process_sentence(\n",
    "        sentence, stemming=False, return_tokens=True\n",
    "    )\n",
    "\n",
    "    encoded_sentence = encode_and_pad_sentence(processed_sentence, vocab, max_len)\n",
    "\n",
    "    prediction = model.predict(tf.convert_to_tensor([encoded_sentence]))\n",
    "\n",
    "    for word, label in zip(\n",
    "        unstemmed_sentence, tf.math.argmax(prediction, axis=-1)[0].numpy()\n",
    "    ):\n",
    "        print(f\"{word}: {list(unique_labels.keys())[label]}\")\n",
    "\n",
    "    return prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [00:10<00:00,  2.73it/s]\n"
     ]
    }
   ],
   "source": [
    "lstm_results = bootstrap_evaluation(lstm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "with open(\"lstm_results.pickle\", \"wb\") as f:\n",
    "    pickle.dump(lstm_results, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 24ms/step\n",
      "J'ai: O\n",
      "passé: O\n",
      "de: O\n",
      "belles: O\n",
      "vacances: O\n",
      "à: O\n",
      "Paris: LOC-ARR\n",
      ".: O\n",
      "C'est: O\n",
      "l'heure: O\n",
      "de: O\n",
      "rentrer: O\n",
      "à: O\n",
      "Montpellier: O\n",
      ".: O\n"
     ]
    }
   ],
   "source": [
    "test_sentence = (\n",
    "    \"J'ai passé de belles vacances à Paris. C'est l'heure de rentrer à Montpellier.\"\n",
    ")\n",
    "\n",
    "test_pred = predict(test_sentence, lstm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparison\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "blue"
         },
         "name": "LSTM",
         "type": "box",
         "y": [
          0.7618257261410788,
          0.7644217107132708,
          0.7775585001409642,
          0.774390243902439,
          0.7608447488584474,
          0.7607361963190185,
          0.7933425797503467,
          0.7749303621169916,
          0.7679465776293823,
          0.7556066705002875,
          0.7724331926863572,
          0.7743201569946734,
          0.8006952491309386,
          0.7796135536264351,
          0.7630090497737556,
          0.7779024116657319,
          0.7739227756015669,
          0.7729957805907173,
          0.7790178571428571,
          0.7672242699177771,
          0.7670149675232985,
          0.780650542118432,
          0.7711670480549199,
          0.7820829898323716,
          0.7791563275434243,
          0.7692307692307693,
          0.7648757943385326,
          0.7610810810810811,
          0.7864849517319704,
          0.7941501103752759
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "red"
         },
         "name": "LSTM with POS",
         "type": "box",
         "y": [
          0.8164736164736165,
          0.7985812009120852,
          0.784516765285996,
          0.8030188679245283,
          0.8226950354609929,
          0.8065573770491803,
          0.8091872791519434,
          0.8013398608606029,
          0.7972406996797241,
          0.801345059493016,
          0.8103008669046405,
          0.820815175596001,
          0.802,
          0.8129829984544049,
          0.8045859872611465,
          0.8225765628954695,
          0.8179299924452279,
          0.8139656070870245,
          0.8120261437908497,
          0.8071500503524672,
          0.8032786885245902,
          0.8046324269889225,
          0.8132375577219086,
          0.8171772978402813,
          0.8084672277480235,
          0.8151873271310033,
          0.8107396683337721,
          0.8050568900126422,
          0.8066098631551769,
          0.8025316455696202
         ]
        }
       ],
       "layout": {
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "title": {
         "text": "F1-scores for LSTM vs LSTM with POS"
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import plotly.express as px\n",
    "import plotly.graph_objects as go\n",
    "import pandas as pd\n",
    "\n",
    "df_lstm_results = pd.DataFrame(lstm_results)\n",
    "df_lstm_with_pos_results = pd.DataFrame(lstm_with_pos_results)\n",
    "\n",
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=df_lstm_results[\"f1_scores\"],\n",
    "        name=\"LSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"blue\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=df_lstm_with_pos_results[\"f1_scores\"],\n",
    "        name=\"LSTM with POS\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"red\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.update_layout(title=\"F1-scores for LSTM vs LSTM with POS\")\n",
    "\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion of using LSTM\n",
    "\n",
    "The limitation of a simple LSTM is that the inputs are treated **sequentially**. Which means that it does not benefit from the context of the words that comes after, only what comes before. Therefore, we will try the **BiLSTM model** hoping that treating the sentence from both sides will leverage more context.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## BiLSTM\n",
    "\n",
    "As mentioned in the previous section, we will be testing a **Bidirectional LSTM** model in order to improve performance.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](./assets/bi-lstm.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### BiLSTM without POS\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "bilstm = tf.keras.models.Sequential(\n",
    "    layers=[\n",
    "        tf.keras.layers.Embedding(len(vocab) + 1, MAX_LEN, mask_zero=True),\n",
    "        tf.keras.layers.Bidirectional(\n",
    "            tf.keras.layers.LSTM(MAX_LEN, return_sequences=True)\n",
    "        ),\n",
    "        tf.keras.layers.Dropout(0.3),\n",
    "        tf.keras.layers.Dense(len(unique_labels), activation=tf.nn.log_softmax),\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "bilstm.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(0.01),\n",
    "    loss=masked_loss,\n",
    "    metrics=[masked_accuracy],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_1\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_1\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ embedding_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Embedding</span>)         │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ bidirectional (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Bidirectional</span>)   │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ ?                      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ embedding_1 (\u001b[38;5;33mEmbedding\u001b[0m)         │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ bidirectional (\u001b[38;5;33mBidirectional\u001b[0m)   │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_3 (\u001b[38;5;33mDropout\u001b[0m)             │ ?                      │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                 │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "bilstm.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m149/149\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 94ms/step - loss: 0.1662 - masked_accuracy: 0.9349 - val_loss: 0.0034 - val_masked_accuracy: 0.9992\n",
      "Epoch 2/10\n",
      "\u001b[1m149/149\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 128ms/step - loss: 0.0022 - masked_accuracy: 0.9994 - val_loss: 6.5582e-04 - val_masked_accuracy: 0.9998\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x2dd4068a0>"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bilstm.fit(\n",
    "    train_dataset.batch(64),\n",
    "    validation_data=test_dataset.batch(64),\n",
    "    epochs=10,\n",
    "    shuffle=True,\n",
    "    callbacks=[\n",
    "        tf.keras.callbacks.EarlyStopping(\n",
    "            monitor=\"val_loss\", min_delta=0.01, restore_best_weights=True\n",
    "        )\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m24/24\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 19ms/step\n",
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step\n",
      "\u001b[1m3/3\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 20ms/step\n"
     ]
    }
   ],
   "source": [
    "blstm_eval_short_predictions = bilstm.predict(encoded_eval_short_sentences)\n",
    "blstm_eval_unlabeled_predictions = bilstm.predict(encoded_eval_unlabeled_sentences)\n",
    "blstm_eval_large_predictions = bilstm.predict(encoded_eval_large_sentences)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [00:20<00:00,  1.49it/s]\n"
     ]
    }
   ],
   "source": [
    "bilstm_results = bootstrap_evaluation(bilstm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "with open(\"bilstm_results.pickle\", \"wb\") as f:\n",
    "    pickle.dump(bilstm_results, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(758, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlEAAAHuCAYAAAC/CFnjAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8ekN5oAAAACXBIWXMAAA9hAAAPYQGoP6dpAACBhUlEQVR4nO3dd1hUR9sG8Ht3aWuhFzsgNrBhRUSsRDRGReNrV+zGiqKJomJXYu9dY+8FNXZji7HHAsZYEOwKgiBW2u58f/ixcV1wARcW8f69115vHOacmbMczj77zJw5EiGEABERERFlilTfHSAiIiL6GjGIIiIiIsoCBlFEREREWcAgioiIiCgLGEQRERERZQGDKCIiIqIsYBBFRERElAUMooiIiIiygEEUERERURZ8dUFUWFgYGjduDDMzM0gkEuzevVun+79//z4kEgnWrFmj0/3mBQ4ODujWrZvO9hcVFYU2bdrAysoKEokEc+fO1dm+v1S3bt3g4OCg725QJp08eRISiQQnT57US/vjx4+HRCLRS9sA8OjRI5iYmODMmTOZ2k4ikWDgwIHZ1CvKbWrVqoVffvnli/bBz8oPshREhYeHo2/fvihZsiRMTExgamoKDw8PzJs3D+/fv9d1H9X4+vri+vXrmDJlCtavX4/q1atna3t50b///ovx48fj/v37eu3H0KFDcfjwYQQEBGD9+vVo0qRJtrYnkUjUXvnz54eLiwsmT56Md+/eZWh7bR80SqUS69atg5ubGywtLVGwYEGUKVMGXbt2xfnz5wF8CEY/7Utar9SLU+q/e/XqlWabo0ePVtWJiYnJ3JtCmfbu3TuMHz9eb4Ha50ycOBFubm7w8PDQd1fSdfbsWYwfPx4vX77Ud1fS9fTpU4wfPx7Xrl3Td1eyxYgRI7Bo0SJERkbqpf3cdg4sXrw468GgyKR9+/YJuVwuzM3NxeDBg8Xy5cvFwoULRfv27YWhoaHo3bt3ZneZYe/evRMAxOjRo7OtDaVSKd6/fy9SUlKyrQ192759uwAgTpw4kantEhISRFJSks76YWdnJzp16qSz/WkDQHz33Xdi/fr1Yv369WLJkiWiY8eOAoBo06aNWt2kpCSRkJCgsf2AAQM+28aAAQMEANGyZUsxb948sWjRIjF48GBRunRpMW7cOCGEEMHBwao+rF+/XnTo0EEAEHPmzFErDw8PV7VrYmIizM3NRWJiokabjo6OwsTERAAQ0dHRX/AOff1OnDiRpXM7M6KjowUA1e/zY8nJyeL9+/fZ1vbnPH/+XBgaGopNmzZletuMnNu6MmPGDAFA3Lt3L0fay4pLly4JAGL16tX67kq2UCgUolChQiIwMDDL+7h3716W36Pcdg6UL19e1KtXL0vbGmQm4Lp37x7at28Pe3t7HD9+HIULF1b9bMCAAbh79y7279+ftWguA6KjowEA5ubm2daGRCKBiYlJtu3/ayOEQEJCAuRyOYyNjXW67+fPn+v0d5mQkAAjIyNIpeknWMuUKYPOnTur/v3TTz8hKSkJu3btQkJCgup3b2homOn2o6KisHjxYvTu3RvLly9X+9ncuXNV56+Pj4/azyIjI7F582b4+PikO4TYpEkT7N27FwcPHkTLli1V5WfPnsW9e/fw448/YufOnZnuM+mWgYEBDAwydVnVmQ0bNsDAwADNmzfXS/vavH37Fvnz59d3NwiAVCpFmzZtsG7dOkyYMEGvQ9BfvcxEXD/99JMAIM6cOZOh+snJyWLixImiZMmSwsjISNjb24uAgACNb/j29vaiWbNm4vTp06JGjRrC2NhYODo6irVr16rqjBs3TgBQe9nb2wshhPD19VX998dSt/nYkSNHhIeHhzAzMxP58+cXZcqUEQEBAaqfpxddHzt2TNSpU0fky5dPmJmZiRYtWoh///03zfbCwsKEr6+vMDMzE6ampqJbt27i7du3Wt+vevXqifLly4uQkBBRt25dIZfLhZOTk9i+fbsQQoiTJ0+KmjVrChMTE1GmTBlx9OhRte3v378v+vXrJ8qUKSNMTEyEpaWlaNOmjVq0v3r1ao33ER99c0/9XRw6dEhUq1ZNGBsbizlz5qh+5uvrK4T4kLGrX7++sLa2FlFRUar9JyYmigoVKoiSJUuKN2/epHmc6fUhVXh4uGjTpo2wsLAQcrlcuLm5iX379qntIzXjsHnzZjF69GhRpEgRIZFIRFxcXLrvL9L5tj1w4EAhk8lEcnKyqiytcyq97VOdO3dOABBr1qxJt05atH0rS223fv36om3btmo/69+/v6hYsaLq3MtIJurKlSuiSZMmomDBgiJ//vyiYcOG4ty5c2p1Un9Hf/31lxg6dKiwtrYW+fLlEz4+PuL58+cZOq6bN2+KH3/8UVhYWAhjY2NRrVo1sWfPHtXPU7/tp/V+HTp0SAAQv//+uxAiY+e2EGlnoj4+bz9Wr149tW+fiYmJIjAwUFStWlWYmpqKfPnyiTp16ojjx4+r6qReHz59pWal0rrm6PI6+Dl169YV9evX1yi/c+eOaN26tbCzsxPGxsaiaNGiol27duLly5eqOqnnWHBwsChfvrwwMjISLi4u4uDBgxr7y8z5c/LkSdGvXz9hY2MjzM3N07yOf+7cz2j/hRBi/fr1omrVqsLExERYWFiIdu3aiYcPH6rVSb3G3rhxQ9SvX1/I5XJRpEgRMW3aNFWd1HPo09fHnwnnz58X3t7ewtTUVMjlclG3bl3x119/qbWV2c+D9evXixo1aqhGejw9PcXhw4fV6hw4cED1OVSgQAHx/fffi3/++UetzrNnz0S3bt1E0aJFhZGRkShUqJBo0aKFxnu8Z88eAUBcuXIl3fc+VVxcnPD19RWmpqbCzMxMdO3aVVy9elXjfQkJCRG+vr7C0dFRGBsbCzs7O9G9e3cRExOj8b6kdw789ttvokGDBsLGxkYYGRkJZ2dnsXjxYo0+Xbp0STRu3FhYWVkJExMT4eDgILp3765WR6FQiDlz5ggXFxdhbGwsbG1tRZ8+fURsbKyqjr29vUZfMpOVylQQVbRoUVGyZMkM1/f19VUNlSxatEh07dpVABA+Pj5q9ezt7UXZsmWFnZ2dGDVqlFi4cKGoWrWqkEgkqhMkJCREzJkzRwAQHTp0EOvXrxfBwcGqdjISRP3zzz/CyMhIVK9eXcybN08sXbpUDB8+XNStW1dVJ60g6ujRo8LAwECUKVNGTJ8+XUyYMEFYW1sLCwsLtRMztb0qVaqI1q1bi8WLF4tevXoJAOKXX37R+n7Vq1dPFClSRBQvXlz8/PPPYsGCBcLFxUXIZDKxZcsWUahQITF+/Hgxd+5cUbRoUWFmZiZevXql2n779u2icuXKYuzYsWL58uVi1KhRwsLCQtjb26v+aMPDw8XgwYMFADFq1CjV0FFkZKTqd1GqVClhYWEhRo4cKZYuXaoWYH38YRQRESEKFCggWrVqpSobOXKkkEgk4tSpU+keZ3h4uFi/fr3G8JoQQkRGRgo7OztRsGBBMXr0aDF79mxRuXJlIZVKxa5du1T7SL3Qubi4CFdXVzF79mwRFBT02WAVgOjZs6eIjo4W0dHR4v79+2Ljxo2iYMGCokuXLmp1sxJEPX36VAAQzZo1y1DQnCqjQdTy5cuFXC4Xr1+/FkJ8+HC2sbERQUFBGQ6i/vnnH5E/f35RuHBhMWnSJPHrr7+qLnjnz59X1Uv9EKxSpYpo2LChWLBggRg2bJiQyWQagVx67ZiZmQkXFxcxbdo0sXDhQlG3bl0hkUjUfo8lS5YU33//vcb23bt3FxYWFqrh44yc20J8WRAVHR0tChcuLPz9/cWSJUvE9OnTRdmyZYWhoaG4evWqEEKIN2/eiCVLlggAolWrVqpzNyQkRAiRdhCly+tgepKSkoRcLhf+/v5q5YmJicLR0VEUKVJETJ48WaxcuVJMmDBB1KhRQ9y/f19VD4CoXLmy6ryYO3euKFmypMiXL5/aB2Bmzx8XFxdRr149sWDBAvHrr7+KkJCQNIev0/vCldH+T548WUgkEtGuXTuxePFi1TXawcFB7YvVx9dYPz8/sXjxYtGwYUMBQBw4cEAI8eEaNHHiRAFA9OnTR2N4/dixY8LIyEi4u7uLWbNmiTlz5ohKlSoJIyMjceHCBVVbmfk8GD9+vAAgateuLWbMmCHmzZsnOnbsKEaMGKGqs27dOiGRSESTJk3EggULxLRp04SDg4MwNzdXu3bUrl1bmJmZiTFjxoiVK1eKqVOnigYNGmhckx8/fiwAiAULFqT53qdSKpWibt26QiqViv79+4sFCxaIhg0bikqVKml8Vs6cOVN4enqKiRMniuXLlws/Pz8hl8tFzZo1hVKpFEIIredAjRo1RLdu3cScOXPEggULROPGjQUAsXDhQlU7UVFRwsLCQpQpU0bMmDFDrFixQowePVo4Ozur9b1Xr17CwMBA9O7dWyxdulSMGDFC5M+fX9SoUUN1bQkODhbFihUT5cqVU/XlyJEjn31PPpbhICo+Pl411yMjrl27JgCIXr16qZUPHz5cAFD7dpcaCf7555+qsufPnwtjY2MxbNgwVVlqgDNjxgy1fWY0iEoNwj73QZNWEOXq6ipsbW3FixcvVGUhISFCKpWKrl27arTXo0cPtX22atVKWFlZpdtmqnr16gkAanMabt26JQAIqVSqdpE6fPiwRj/fvXunsc/U7Mi6detUZZ+bE5X6uzh06FCaP/v0w2jZsmUCgNiwYYM4f/68kMlkYsiQIVqPVYi0g5IhQ4YIAOL06dOqstevXwtHR0fh4OAgFAqFEOK/D8uSJUumedzptZfWy8fHRyMrkJUgSgih+oC0sLAQrVq1EjNnzhQ3b9787DYZDaJiY2OFkZGRKuDcv3+/kEgk4v79+xkOonx8fISRkZHqA0GID8FfwYIF1b5MpH4Ienl5qS5+QggxdOhQIZPJNLIAn2rUqJGoWLGi2vuqVCpF7dq1RenSpVVlAQEBwtDQUO2bYWJiojA3N1f7O8rouf0lQVRKSorGnLO4uDhhZ2en1pfPzYn69JqTHdfBtNy9ezfND8TUbEFqNjs9AISRkZG4e/euqiwkJERjn5k9f+rUqaMxvzQz82Ey0v/79+8LmUwmpkyZolZ+/fp1YWBgoFaeeo39+JxJTEwUhQoVEj/++KOqLL05UUqlUpQuXVp4e3ur/V28e/dOODo6iu+++05VltHPg7CwMCGVSkWrVq1U17eP2xPiwzXQ3NxcY85xZGSkMDMzU5XHxcWl+RmZHiMjI9GvX7/P1tm9e7cAIKZPn64qS0lJEZ6enhn6DNq8ebPGef25cyCtfXh7e6slcIKDgwUAcenSpXT7ffr0aQFAbNy4Ua08Ncv9cfmXzInK8N15r169AgAULFgwQ/UPHDgAAPD391crHzZsGABozJ1ycXGBp6en6t82NjYoW7YsIiIiMtpFrVLn3+zZswdKpTJD2zx79gzXrl1Dt27dYGlpqSqvVKkSvvvuO9Vxfuynn35S+7enpydevHiheg8/p0CBAmjfvr3q32XLloW5uTmcnZ3h5uamKk/974/fH7lcrvrv5ORkvHjxAqVKlYK5uTmuXLmSgaP9wNHREd7e3hmq26dPH3h7e2PQoEHo0qULnJycMHXq1Ay39akDBw6gZs2aqFOnjqqsQIEC6NOnD+7fv49///1Xrb6vr6/acWvTsmVLHD16FEePHsWePXsQEBCAQ4cOoWPHjhBCZLnfqVavXo2FCxfC0dERwcHBGD58OJydndGoUSM8efLki/ZtYWGBJk2aYPPmzQCATZs2oXbt2rC3t8/Q9gqFAkeOHIGPjw9KliypKi9cuDA6duyIv/76S+Mc7dOnj9p8CU9PTygUCjx48CDddmJjY3H8+HG0bdsWr1+/RkxMDGJiYvDixQt4e3sjLCxM9V60a9cOycnJ2LVrl2r7I0eO4OXLl2jXrp2qTFfn9ufIZDIYGRkB+HCXZWxsLFJSUlC9evUst5FT18EXL14A+HCOfMzMzAwAcPjwYa13oHp5ecHJyUn170qVKsHU1FTVdlbOn969e0Mmk3223c/JSP937doFpVKJtm3bqs61mJgYFCpUCKVLl8aJEyfU6hcoUEBtXqSRkRFq1qyZoc+aa9euISwsDB07dsSLFy9Ubb19+xaNGjXCn3/+qfHZou3zYPfu3VAqlRg7dqzGfM7Uv72jR4/i5cuX6NChg9oxymQyuLm5qY5RLpfDyMgIJ0+eRFxcnNbjsbCw0HpH74EDB2BgYIB+/fqpymQyGQYNGqRR9+O/04SEBMTExKBWrVoAkOG/oY/3ER8fj5iYGNSrVw8RERGIj48H8N9n+b59+5CcnJzmfrZv3w4zMzN89913au9ZtWrVUKBAAY3zIqsyHESZmpoCAF6/fp2h+g8ePIBUKkWpUqXUygsVKgRzc3ONi3CJEiU09mFhYZGhEyGj2rVrBw8PD/Tq1Qt2dnZo3749tm3b9tmAKrWfZcuW1fiZs7Oz6g/oY58eS+qFLSPHUqxYMY1JfmZmZihevLhG2af7fP/+PcaOHYvixYvD2NgY1tbWsLGxwcuXL1UnX0Y4OjpmuC4ArFq1Cu/evUNYWBjWrFmTqaDmUw8ePEj3vU79+Zf0tVixYvDy8oKXlxdatGiBqVOnYvLkydi1axf27duX5X6nkkqlGDBgAC5fvoyYmBjs2bMHTZs2xfHjx9WC46zq2LEjjh49iocPH2L37t3o2LFjhreNjo7Gu3fv0n1/lUolHj16pFaelXP57t27EEIgMDAQNjY2aq9x48YB+HBTAQBUrlwZ5cqVw9atW1Xbb926FdbW1mjYsKGqTFfntjZr165FpUqVYGJiAisrK9jY2GD//v1ZbiOnr4OffhFwdHSEv78/Vq5cCWtra3h7e2PRokVpHo+2trNy/mT27/NTGel/WFgYhBAoXbq0xvl28+ZN1bmWKq1rbEbf47CwMAAfvrx92tbKlSuRmJio8d5q+xsKDw+HVCqFi4uL1nYbNmyo0e6RI0dUx2hsbIxp06bh4MGDsLOzQ926dTF9+vR0lzIQQmidVP7gwQMULlwYBQoUUCtP6zyIjY2Fn58f7OzsIJfLYWNjozoHMvo3dObMGXh5eSF//vwwNzeHjY0NRo0apbaPevXq4ccff8SECRNgbW2Nli1bYvXq1UhMTFTtJywsDPHx8bC1tdV4z968eaNxXmRVhm8jMTU1RZEiRfDPP/9kqoGMzvpP79tKRrID6bWhUCjU/i2Xy/Hnn3/ixIkT2L9/Pw4dOoStW7eiYcOGOHLkyBd9Y/rYlxxLettmZJ+DBg3C6tWrMWTIELi7u6sWJG3fvn2GM28AMh0EnTx5UnXyXr9+He7u7pna/kt8ScCWqlGjRgCAP//8U6d3NllZWaFFixZo0aIF6tevj1OnTuHBgwcZzhylpUWLFjA2Noavry8SExPRtm1bnfU3LVk5l1PPteHDh6eb0fw4qGjXrh2mTJmCmJgYFCxYEHv37kWHDh3U7nL7knP7c9eHj49vw4YN6NatG3x8fPDzzz/D1tYWMpkMQUFBCA8P/2wb2mT3ddDKygpA2sHtrFmz0K1bN+zZswdHjhzB4MGDERQUhPPnz6NYsWJf3Pbn6OLvU1v/lUolJBIJDh48mOYxfPrh/yXHmXquzZgxA66urmnW0WV7n7a7fv16FCpUSOPnH/+tDBkyBM2bN8fu3btx+PBhBAYGIigoCMePH0eVKlXUtnv58iWsra0z3A9t2rZti7Nnz+Lnn3+Gq6srChQoAKVSiSZNmmToMyg8PByNGjVCuXLlMHv2bBQvXhxGRkY4cOAA5syZo9qHRCLBjh07cP78efz+++84fPgwevTogVmzZuH8+fOqdm1tbbFx48Y027KxsdHJMWfqXtwffvgBy5cvx7lz57R+UNrb20OpVCIsLEyVRQA+3Ab+8uXLL/og+ZSFhUWai3alNeQglUrRqFEjNGrUCLNnz8bUqVMxevRonDhxAl5eXmkeBwDcvn1b42e3bt2CtbV1rrltd8eOHfD19cWsWbNUZQkJCRrvjS5vZ3327BkGDRqExo0bw8jISPXBmdXfr729fbrvderPdS0lJQUA8ObNG53vO1X16tVx6tQpPHv27IuOQS6Xw8fHBxs2bEDTpk0zdQG0sbFBvnz50n1/pVKpRsYzK1KHegwNDdP8m/pUu3btMGHCBOzcuRN2dnZ49eqVRtYuo+d2Wj53ffh4WGrHjh0oWbIkdu3apfY3kpo9S5WZv5+cug6WKFECcrkc9+7dS/PnFStWRMWKFTFmzBicPXsWHh4eWLp0KSZPnpzhNnR1/mTl+vO5/js5OUEIAUdHR5QpUybT+85MH1OHO01NTTN0bmeEk5MTlEol/v3333QDs9R2bW1tM9Suk5MThg0bhmHDhiEsLAyurq6YNWsWNmzYoKrz5MkTJCUlqZ2XabG3t8exY8fw5s0btQDx0/MgLi4Ox44dw4QJEzB27FhVeWoW7WPpvb+///47EhMTsXfvXrUMXnpDb7Vq1UKtWrUwZcoUbNq0CZ06dcKWLVvQq1cvODk54Y8//oCHh4fWYP5LPhMztWL5L7/8gvz586NXr16IiorS+Hl4eDjmzZsHAPj+++8BQONRHrNnzwYANGvWLCv9TZOTkxPi4+MRGhqqKnv27BmCg4PV6sXGxmpsm3rSfpwG/FjhwoXh6uqKtWvXql2I//nnHxw5ckR1nLmBTCbT+HazYMECjYxcatCni9Vie/fuDaVSiVWrVmH58uUwMDBAz549s/zt9fvvv8fFixdx7tw5Vdnbt2+xfPlyODg4fDblnVW///47gA9DS18iMjJSY84WACQlJeHYsWNpDutkxfDhwzFu3DgEBgZmajuZTIbGjRtjz549aqvVR0VFYdOmTahTp45q2P5L2Nraon79+li2bBmePXum8fPU9bJSOTs7o2LFiti6dSu2bt2KwoULo27duhp9z8i5nRYnJyecP38eSUlJqrJ9+/ZpDD2lZgw+bufChQtq5yIA5MuXD0DG/n5y6jpoaGiI6tWr4++//1Yrf/XqlepLQqqKFStCKpWme81Lj67On8xcfzLS/9atW0Mmk2HChAka54gQQjVfLDPS62O1atXg5OSEmTNnpvml69NzOyN8fHwglUoxceJEjWxN6vF4e3vD1NQUU6dOTXMOUGq77969Q0JCgtrPnJycULBgQY3f9+XLlwEAtWvX/mz/vv/+e6SkpGDJkiWqMoVCgQULFqjVS+vvB9A894H039+09hEfH4/Vq1er1YuLi9No59PP8rZt20KhUGDSpEka7aekpKi1nT9//ix/HmYqE+Xk5IRNmzahXbt2cHZ2RteuXVGhQgUkJSXh7Nmz2L59u+rZapUrV4avry+WL1+Oly9fol69erh48SLWrl0LHx8fNGjQIEsdTkv79u0xYsQItGrVCoMHD8a7d++wZMkSlClTRm0y28SJE/Hnn3+iWbNmsLe3x/Pnz7F48WIUK1ZMbSLzp2bMmIGmTZvC3d0dPXv2xPv377FgwQKYmZlh/PjxOjuOL/XDDz9g/fr1MDMzg4uLC86dO4c//vhDlepP5erqCplMhmnTpiE+Ph7GxsZo2LAhbG1tM9Xe6tWrsX//fqxZs0Y1LLBgwQJ07twZS5YsQf/+/TN9DCNHjsTmzZvRtGlTDB48GJaWlli7di3u3buHnTt3fnYhzYy4c+eO6tvYu3fvcP78eaxduxalSpVCly5dtG7/999/p/ntvX79+jAxMUHNmjXRsGFDNGrUCIUKFcLz58+xefNmhISEYMiQITpJnVeuXDnLAd/kyZNx9OhR1KlTB/3794eBgQGWLVuGxMRETJ8+/Yv7lmrRokWoU6cOKlasiN69e6NkyZKIiorCuXPn8PjxY4SEhKjVb9euHcaOHQsTExP07NlT4/ec0XM7Lb169cKOHTvQpEkTtG3bFuHh4diwYYPaJOrUNnbt2oVWrVqhWbNmuHfvHpYuXQoXFxe1D0y5XA4XFxds3boVZcqUgaWlJSpUqIAKFSpotJ2T18GWLVti9OjRePXqlSqYOX78OAYOHIj//e9/KFOmDFJSUrB+/XrIZDL8+OOPmW5DF+dPtWrVAHx4XFH79u1haGiI5s2bp5nRz0j/nZycMHnyZAQEBOD+/fvw8fFBwYIFce/ePQQHB6NPnz4YPnx4po7TyckJ5ubmWLp0KQoWLIj8+fPDzc0Njo6OWLlyJZo2bYry5cuje/fuKFq0KJ48eYITJ07A1NRU9aUso0qVKoXRo0dj0qRJ8PT0ROvWrWFsbIxLly6hSJEiCAoKgqmpKZYsWYIuXbqgatWqaN++PWxsbPDw4UPs378fHh4eWLhwIe7cuYNGjRqhbdu2cHFxgYGBAYKDgxEVFaWR3T169ChKlCihMcT3qebNm8PDwwMjR47E/fv34eLigl27dmnMcTI1NVXNwUpOTkbRokVx5MiRNLOj6Z0DqSMazZs3R9++ffHmzRusWLECtra2al/I1q5di8WLF6NVq1ZwcnLC69evsWLFCpiamqq+uNSrVw99+/ZFUFAQrl27hsaNG8PQ0BBhYWHYvn075s2bhzZt2qj6s2TJEkyePBmlSpWCra2t2pzMz8rKLX137twRvXv3Fg4ODsLIyEgULFhQeHh4iAULFqjd0pycnCwmTJggHB0dhaGhoShevPhnF5n71Ke3IKe3xIEQHxbRrFChgjAyMhJly5YVGzZs0Ljd+NixY6Jly5aiSJEiwsjISBQpUkR06NBB3LlzR6ONT29t/eOPP4SHh4eQy+XC1NRUNG/ePN3FNj+9zTz1dl9tt/SmLgT3qfTeH3xyy31cXJzo3r27sLa2FgUKFBDe3t7i1q1bad7ivWLFClGyZEkhk8nUbglPr63Un6Xu59GjR8LMzEw0b95co16rVq1E/vz5RURExGeP99P+p0pdbNPc3FyYmJiImjVrprvYprZbtz9t7+OXTCYTxYoVE3369FFbMFSI9Jc4SO81adIk8erVKzFv3jzh7e0tihUrJgwNDUXBggWFu7u7WLFihdot0R/L6BIHn5PZxTa9vb1FgQIFRL58+USDBg3E2bNn1eqknrOf3kKcmceqhIeHi65du4pChQoJQ0NDUbRoUfHDDz+IHTt2aNQNCwtTvZefLlooRMbP7fT6N2vWLFG0aFFhbGwsPDw8xN9//61xfVEqlWLq1KnC3t5eGBsbiypVqoh9+/aleS6cPXtWVKtWTRgZGaktd5DeYpu6vA6mJyoqShgYGKiWwBDiw1puPXr0EE5OTqpFShs0aCD++OMPtW3TO8fSunZ8yfmTatKkSaJo0aJCKpV+9tzPaP+FEGLnzp2iTp06In/+/CJ//vyiXLlyYsCAAeL27duqOuldY9P6He/Zs0e4uLgIAwMDjc+Eq1evitatWwsrKythbGws7O3tRdu2bcWxY8dUdTL7efDbb7+JKlWqCGNjY2FhYSHq1aunsaDyiRMnhLe3tzAzMxMmJibCyclJdOvWTfz9999CCCFiYmLEgAEDRLly5UT+/PmFmZmZcHNzE9u2bVPbj0KhEIULFxZjxozRfNPT8OLFC9GlSxfVYptdunRJc7HNx48fi1atWglzc3NhZmYm/ve//6nWz/t0SZD0zoG9e/eKSpUqqRbQnDZtmvjtt9/U6ly5ckV06NBBlChRQrWI5g8//KB6Hz62fPlyUa1aNSGXy0XBggVFxYoVxS+//CKePn2qqhMZGSmaNWsmChYsKIDMLbYpEUIH93UTEZHe9ezZE3fu3MHp06f13RXKxVLv7A0PD1d7fBtlHoMoIqI84uHDhyhTpgyOHTsGDw8PfXeHcil3d3d4enrqdAj/W8UgioiIiCgLvmyWLhEREdE3ikEUERERURYwiCIiIiLKAgZRRERERFnAIIqIiIgoCzK1YjllH3mVgfruAuUicZcW6rsLlMukKHgjNf2ngLHunoGaHl19Lr2/mnevZwyiiIiISJOEg1Xa8B0iIiIiygJmooiIiEiTJPuHDL92DKKIiIhIE4fztOI7RERERJQFzEQRERGRJg7nacUgioiIiDRxOE8rvkNEREREWcBMFBEREWnicJ5WDKKIiIhIE4fztGIQRURERJqYidKKYSYRERFRFjATRURERJo4nKcVgygiIiLSxOE8rRhmEhEREWUBM1FERESkicN5WjGIIiIiIk0cztOKYSYRERFRFjATRURERJo4nKcVgygiIiLSxCBKK75DRERERFnATBQRERFpknJiuTYMooiIiEgTh/O0YhBFREREmrjEgVYMM4mIiIiygJkoIiIi0sThPK0YRBEREZEmDudpxTCTiIiIKAuYiSIiIiJNHM7TikEUERERaeJwnlYMM4mIiIiygJkoIiIi0sThPK0YRBEREZEmDudpxTCTiIiIKAuYiSIiIiJNHM7TikEUERERaeJwnlYMooiIiEgTM1Fa8R0iIiIiygJmooiIiEgTM1FaMYgiIiIiTZwTpRXDTCIiIqIsYCaKiIiINHE4TysGUURERKSJw3laMcwkIiIiygJmooiIiEgTh/O0YhBFREREmjicpxXDTCIiIqIsYCaKiIiINEiYidKKQRQRERFpYBClHYMoIiIi0sQYSivOiSIiIiLKAmaiiIiISAOH87RjEEVEREQaGERpx+E8IiIioixgJiqLhBCM0omIKM/iZ5x2zERl0vv375GYmIhHjx4hISFB393J9fq2rYtb+ycg7vwc/LluOKqXt0+3roGBFAF9muDG3nGIOz8HF7aOxHe1ndXqSKUSjO3fDDf3jUfsudm4sXccRvZukt2HQXqwZdNGNP2uIWpUqYhO7f+H66Gh+u4S6di2LRvxQ5OGcK9eCV07tsU/1z//Oz565BBat2gK9+qV0LZ1c/x1+pTqZ8nJyZg/Zybatm4Oj5pV4N3IE2NHjUD086jsPow8SyKR6OSVlzGIyoSbN2+ic+fOqF69OpycnODu7o6RI0fqu1u5VpvGVTFtWCtMWXYQ7h2nIfTOE+xdPAA2FgXSrD++f3P0+rEO/KdvR5UfJ2Pljr+wdVZvVC5bTFVnWLfv0LuNJ4b+uh2urSdjzPw98Pf1Qv8O9XLqsCgHHDp4ADOnB6Fv/wHYsj0YZcuWQ7++PfHixQt9d4105MihA5g941f0+WkANm7dhTJly2LgT70Qm87vOOTaFYweMQw+rdpg07Zg1G/ohWF+A3E37A4AICEhAbdu/otefftj49admDl7Ae7fv4ehg/vn5GHRN4ZBVAZdv34d7u7uKFy4MIYMGYJt27bB3t4ec+fORfPmzZGcnKzvLuY6gzs3xOpdZ7F+73nciojEoClb8D4hCb4+7mnW7/hDTUxfdQSH//oX95+8wIrtf+HwmX/h16Whqk6tyiWx71QoDv11Aw+fxSL4j2s4dv7WZzNc9PVZv3Y1WrdpC59WP8KpVCmMGTcBJiYm2L1rp767RjqyYd0atPrxf2jh8yNKOpXCqMAJMJGbYM/utH/Hmzeuh7tHHXTt3hOOJZ3Qf6Afyjm7YNuWjQCAggULYvHy39DYuykcHEuiYmVXjBgViJv/3sCzZ09z8tDyDomOXnkYg6gMiI6Ohq+vL/r164eFCxeiZ8+eaNWqFVasWIFZs2bhxIkT6NKli767masYGshQxbk4jl+4rSoTQuD4hduoWckxzW2MDA2QkKQejL5PSELtKk6qf58PiUCDmmVRqoQtAKBimaJwdy2JI2f+zYajIH1ITkrCzX9voJZ7bVWZVCpFrVq1ERpyVY89I11JTk7CrZs3ULOW+u+4pps7rodcS3Ob0JBrcHOrrVbmXtsDoenUB4A3b15DIpGgYEFTXXT7m8PhPO04sTwDHj9+jOTkZHTq1AkKhQIymQxKpRI2Njbo3LkzXr58ialTp2L37t3w8fHRd3dzBWuLAjAwkOF57Gu18ucvXqGsg12a2/xx7iYGd26Iv67cRcSjGDSoWRYtG7pCJvvvj3Dm6qMwLWCCkOAxUCgEZDIJxi3ahy0H/87W46GcE/cyDgqFAlZWVmrlVlZWuHcvQk+9Il16GZfe79ga9+/dS3ObFzExsPykvqWVNV7ExKRZPzExEfPnzIR302YoUCDtKQREX4pBVAaEhITg7t27qFChAoAPGRWp9EMSz8zMDB07dsSMGTNw9+7dDO0vMTERiYmJamVCqYBEKtNtx78yw2fswOLADgjZFQghBCIex2Dd3vPwbVlLVadN46po37QGuo1ai3/Dn6FS2aKYMbwNnkXHY+PvF/TYeyLKLZKTkzFy+BAIAQSMGa/v7ny18noWSRc4nJcBpUqVAgDs3PlhrP7TE8vR0RElS5bEkydPMrS/oKAgmJmZqb1Soi7rttN6FhP3BikpCthaFlQrt7UyReSLV+lu09Z/Baxq+6Ps92NRudUkvH2XiHtP/ptoOnWID2auPorthy/jxt2n2Lz/EhZsPI6fu3+XrcdDOcfC3AIymUxjEvmLFy9gbW2tp16RLplbpPc7jkn3d2xlba0x6Tz2RQysPqmfnJyMkT8PxbNnT7F4+Spmob4Ah/O0YxCVAQ4ODjA1NcW6devw4MEDVblSqQQAxMXFQS6Xo1q1ahnaX0BAAOLj49VeBnYZ2/ZrkZyiwNWbj9DArayqTCKRoEHNMrgYmna6PlViUgqeRsfDwEAKn0au2Hfyv9ue5SZGUAqlWn2F8r/MIH39DI2M4OxSHhfOn1OVKZVKXLhwDpUqV9Fjz0hXDA2NUM65PC5dUP8dX7pwHhUru6a5TaXKrrj4UX0AuHD+LCp9VD81gHr04AGWLF8Nc3OL7Oj+N4NBlHb85MmAYsWKYcmSJTh06BACAwNx48YNAFB9cM+ePRtPnz6Fp6dnhvZnbGwMU1NTtVdeHMqbv+E4ureqjU7N3VDW0Q7zR7VDPrkx1u05DwBYOakLJg5qoapfo4I9WjasDIeiVvCo4oS9CwdAKpVg9po/VHUO/HkdI3p6o0md8ihR2BItGlTC4M4NsPd4SI4fH2WfLr7dsWvHNuzdHYyI8HBMnjge79+/h0+r1vruGulI567dELxzO37fE4x7EeEImvzhd9zC58PveOyoEVgwb5aqfodOXXD27F9Yv/Y33LsXgWWLF+DfGzfQtn0nAB8CqBHD/HDzxj+Y/OsMKJQKxMREIyYmGsnJSXo5Rsq6RYsWwcHBASYmJnBzc8PFixc/W3/u3LkoW7Ys5HI5ihcvjqFDh+bIWo6cE5VBLVu2xPz58zFw4EBcvHgRHh4eKFy4MO7du4eDBw/i2LFjsLfnbfYf23HkCqwtCmBsv2awsyqI0NtP0HLAItVk8+KFLKFUClV9Y2NDjBvwAxyLWuPNu0QcPnMDPQPXIf7Ne1Ud/2nbMa7/D5g3qh1sLArgWXQ8Vu04g6nLD+b48VH2adL0e8TFxmLxwvmIiYlG2XLOWLxspcbQDX29Gjf5HnFxsVi6eAFexESjTFlnLFiyAlZWH37HkZFPIZH+l8Wo7FoVU36diSUL5mLR/DkoUcIBs+YtRKnSZQAA0c+jcOrkcQBAh//5qLW1bNVaVK/hljMHlpfoKYm0detW+Pv7Y+nSpXBzc8PcuXPh7e2N27dvw9bWVqP+pk2bMHLkSPz222+oXbs27ty5g27dukEikWD27NnZ2leJEEJor0apLly4gOnTp+P27dswNzdH5cqVMWjQIJQrV+6L9iuvMlBHPaS8IO7SQn13gXKZFAUv1fSfAsbZH+FYd9uik/3ErGmfqfpubm6oUaMGFi78cB1UKpUoXrw4Bg0alOYC1wMHDsTNmzdx7NgxVdmwYcNw4cIF/PXXX1/WeS2YicokNzc3bNu2DVKpFBKJBEqlkvNxiIiI0pHWHenGxsYwNjbWqJuUlITLly8jICBAVSaVSuHl5YVz585p1AeA2rVrY8OGDbh48SJq1qyJiIgIHDhwIEfWb+SnfxakBlAAbwElIqK8SVcTy9O6Iz0oKCjNNmNiYqBQKGBnp76eoJ2dHSIjI9PcpmPHjpg4cSLq1KkDQ0NDODk5oX79+hg1apTO35NPMYjKgo8DJwZRRESUF+kqiErrjvSPM01f6uTJk5g6dSoWL16MK1euYNeuXdi/fz8mTZqkszbSw+E8IiIiyjbpDd2lxdraGjKZDFFRUWrlUVFRKFSoUJrbBAYGokuXLujVqxcAoGLFinj79i369OmD0aNHZ+uUG2aiiIiISJMeHkBsZGSEatWqqU0SVyqVOHbsGNzd0354/bt37zQCJZnsw7JB2X3vHDNRREREpEFf01X8/f3h6+uL6tWro2bNmpg7dy7evn2L7t27AwC6du2KokWLquZVNW/eHLNnz0aVKlXg5uaGu3fvIjAwEM2bN1cFU9mFQRQRERHlGu3atUN0dDTGjh2LyMhIuLq64tChQ6rJ5g8fPlTLPI0ZMwYSiQRjxozBkydPYGNjg+bNm2PKlCnZ3leuE5VLcJ0o+hjXiaJPcZ0o+lhOrBNVqPcOnewnckUbnewnN2ImioiIiDTw7nPtGEQRERGRBgZR2vHuPCIiIqIsYCaKiIiINDERpRWDKCIiItLA4TztOJxHRERElAXMRBEREZEGZqK0YxBFREREGhhEacfhPCIiIqIsYCaKiIiINDERpRWDKCIiItLA4TztOJxHRERElAXMRBEREZEGZqK0YxBFREREGhhEaccgioiIiDQwiNKOc6KIiIiIsoCZKCIiItLERJRWDKKIiIhIA4fztONwHhEREVEWMBNFREREGpiJ0o5BFBEREWlgDKUdh/OIiIiIsoCZKCIiItLA4TztGEQRERGRBsZQ2nE4j4iIiCgLmIkiIiIiDRzO045BFBEREWlgDKUdgygiIiLSIJUyitKGc6KIiIiIsoCZKCIiItLA4TztGEQRERGRBk4s147DeURERERZwEwUERERaWAiSjsGUURERKSBw3nacTiPiIiIKAuYiSIiIiINzERpxyCKiIiINDCG0o7DeURERERZwEwUERERaeBwnnYMooiIiEgDYyjtGEQRERGRBmaitOOcKCIiIqIsYCaKiIiINDARpR2DKCIiItLA4TztOJxHRERElAXMRBEREZEGJqK0YxBFREREGjicpx2H84iIiIiygJmoXOLBn3P03QXKRSya83wgdc92DdZ3FyhXkWV7C0xEaccgioiIiDRwOE87DucRERERZQEzUURERKSBiSjtGEQRERGRBg7naccgioiIiDQwhtKOc6KIiIiIsoCZKCIiItLA4TztGEQRERGRBgZR2nE4j4iIiCgLmIkiIiIiDUxEaccgioiIiDRwOE87DucRERERZQEzUURERKSBiSjtGEQRERGRBg7nacfhPCIiIqIsYBBFREREGiQS3byyYtGiRXBwcICJiQnc3Nxw8eLFz9Z/+fIlBgwYgMKFC8PY2BhlypTBgQMHstZ4JnA4j4iIiDRI9TSct3XrVvj7+2Pp0qVwc3PD3Llz4e3tjdu3b8PW1lajflJSEr777jvY2tpix44dKFq0KB48eABzc/Ns7yuDKCIiItKgrylRs2fPRu/evdG9e3cAwNKlS7F//3789ttvGDlypEb93377DbGxsTh79iwMDQ0BAA4ODjnSVw7nERERUbZJTEzEq1ev1F6JiYlp1k1KSsLly5fh5eWlKpNKpfDy8sK5c+fS3Gbv3r1wd3fHgAEDYGdnhwoVKmDq1KlQKBTZcjwfYxBFREREGiQSiU5eQUFBMDMzU3sFBQWl2WZMTAwUCgXs7OzUyu3s7BAZGZnmNhEREdixYwcUCgUOHDiAwMBAzJo1C5MnT9b5e/IpDucRERGRBqmOhvMCAgLg7++vVmZsbKybnQNQKpWwtbXF8uXLIZPJUK1aNTx58gQzZszAuHHjdNZOWhhEERERUbYxNjbOcNBkbW0NmUyGqKgotfKoqCgUKlQozW0KFy4MQ0NDyGQyVZmzszMiIyORlJQEIyOjrHdeCw7nERERkQZdDedlhpGREapVq4Zjx46pypRKJY4dOwZ3d/c0t/Hw8MDdu3ehVCpVZXfu3EHhwoWzNYACGEQRERFRGvS1TpS/vz9WrFiBtWvX4ubNm+jXrx/evn2ruluva9euCAgIUNXv168fYmNj4efnhzt37mD//v2YOnUqBgwYoKu3Il0cziMiIqJco127doiOjsbYsWMRGRkJV1dXHDp0SDXZ/OHDh5BK/8sBFS9eHIcPH8bQoUNRqVIlFC1aFH5+fhgxYkS291UihBDZtfOXL1/myGJXecHz18n67gLlIvZtF+q7C5TLPNs1WN9doFzEXC7TXukL/bDskk72s69vDZ3sJzfS2XDetGnTsHXrVtW/27ZtCysrKxQtWhQhISG6aoaIiIhygFSim1deprMgaunSpShevDgA4OjRozh69CgOHjyIpk2b4ueff9ZVM0RERES5gs7mREVGRqqCqH379qFt27Zo3LgxHBwc4ObmpqtmiIiIKAdk9s66b5HOMlEWFhZ49OgRAODQoUOqJduFEDmy9DoRERHpjr7uzvua6CwT1bp1a3Ts2BGlS5fGixcv0LRpUwDA1atXUapUKV01Q0RERDlAmtcjIB3QWRA1Z84cODg44NGjR5g+fToKFCgAAHj27Bn69++vq2aIiIiIcgWdBVGGhoYYPny4RvnQoUN11QQRERHlECaitPuiIGrv3r0ZrtuiRYsvaYqIiIhyECeWa/dFQZSPj0+G6kkkEk4uJyIiojzli4Kojx/2R0RERHkHE1HaZcuz8xISEmBiYpIduyYiIqIcwLvztNPZOlEKhQKTJk1C0aJFUaBAAURERAAAAgMDsWrVKl01kytl4+MHiYiIKJfSWRA1ZcoUrFmzBtOnT4eRkZGqvEKFCli5cqWumtG7iIgIBAYGol+/fpg/fz6AD3O+GEgREVFeItHRKy/TWRC1bt06LF++HJ06dYJM9t/TpStXroxbt27pqhm9CgkJgYeHB86cOYMTJ05gyJAh6Nu3LwDexUBERHmLRCLRySsv01kQ9eTJkzRXJlcqlUhOTtZVM3pz/fp11K5dGz179sTRo0dx4sQJzJkzBytWrMD27dv13T0iIiKdkkp088rLdBZEubi44PTp0xrlO3bsQJUqVXTVjF7ExcWhTZs2KF26NCZPngyZTIbChQujSZMmsLOzyxNBIhEREWWOzu7OGzt2LHx9ffHkyRMolUrs2rULt2/fxrp167Bv3z5dNaMXCoUCHTt2xIIFCxAUFISAgABVeXx8PCwsLPTcQyIiIt3K60NxuqCzIKply5b4/fffMXHiROTPnx9jx45F1apV8fvvv+O7777TVTM56s6dO4iPj0eNGjUwYMAAmJiYICgoCGZmZujRowcaN26M3r17qx62TERElFcwhtJOp+tEeXp64ujRo7rcpd6EhISgSpUqmDdvHmrUqAFra2v07NkTQgiMGjUKgwcPxtChQzFjxgwIISCEgFSqs9FRIiIiyuV0vtjm33//jZs3bwL4ME+qWrVqum4i24WEhKB27doYNWoUBg0apCq3trZG165dYWhoiKlTp6qG8SQSCVdvJyKiPIXDedrpLIh6/PgxOnTogDNnzsDc3BwA8PLlS9SuXRtbtmxBsWLFdNVUtgoNDYWHhweGDBmCyZMnq8o3b96MDh06oEiRIujSpQtSUlIwdepUGBgY4JdfflFb1oGIiOhrl9fvrNMFnY0/9erVC8nJybh58yZiY2MRGxuLmzdvQqlUolevXrpqJls9ffoUrq6u6NixI6ZMmaIqnzZtGjp16oS///4bAGBra4uePXtizJgxGDlyJObOnaunHhMREZG+6CwTderUKZw9exZly5ZVlZUtWxYLFiyAp6enrprJVsbGxqhcuTIuXbqEsLAwlC5dGr/++itmzpyJw4cPo3r16hBCQCKRwNraGj169IChoSEnlhMRUZ7D4TztdBZEFS9ePM31khQKBYoUKaKrZrKNEAJWVlb4448/0LRpU7Rt2xaNGjXC2rVrsXnzZtUdhqkn1Y0bN1C2bFn4+fnxRCMiojyHn2za6Ww4b8aMGRg0aJBqyAv4MMncz88PM2fO1FUz2c7KygqHDh2CqakpZs+ejenTp2ss0TBixAi0b98er1+/ZgBFRET0jfqiTJSFhYVaEPH27Vu4ubnBwODDblNSUmBgYIAePXrAx8fnizqaHZ49e4bz58/j8OHDEEKgadOmqFGjBooWLYrg4GD88MMPmDVrFmrXrq0aphw3bhwWLFiAkydPcpFNIiLKs6RMEmj1RUHU1zyh+saNG+jWrRvMzMwQGRkJuVyOFStWoGXLlvD394enpyf2798Pb29vtGrVCn/88QdWrFiBadOm4cyZM1/l0g1EREQZxRhKuy8Konx9fXXVjxwVEhKCevXqoU+fPujbty+cnJwAAKtXr8awYcPw/v17TJkyBdWqVcPBgwfRvHlzFCtWDHK5nAEUERF9EzhdRbtsWWI7ISEBr169UnvlFjdu3ECdOnXg5+eH6dOnw8HBAQCgVCrRvXt3LFy4ECdOnMDWrVsBfJgjtWfPHrRr144BFBEREanoLIh6+/YtBg4cCFtbW+TPnx8WFhZqr9wgISEBHTp0gKWlJbp27QoAkMlkqmULAKBjx44YMmQIlixZgidPngAAbGxssHHjRri6uuqr61+tXds243/NG6NR7aro49sB//5z/bP1T/xxGJ1+bI5GtavCt10rnPvrT7Wfe1avkOZr07rfsvMwSEf6/lAZt9b0QNyeQfhzTntUL2P32foDfaogZIUvYncPQti6Xpjepx6MDf9b2LaA3BAz+tbD7TU9Ebt7EE7MaodqWvZJucf2LZvg09QLnjVd0aNzO9y4HvrZ+seOHEJbn2bwrOmKjm1a4szpU+nW/XXyeLi5umDzhnW67vY3QyLRzSsv01kQ9csvv+D48eNYsmQJjI2NsXLlSkyYMAFFihTBunW54yQ2MTHBtGnTYGhoiMmTJyMkJATAfylLhUIBAPDy8oJUKkV0dLRqWz4XL/OOHTmIhXOmo1vvfli5YTtKlSmLYYP6Ii72RZr1r4dcxYTRv6BZy1ZYtXE7POs3xKjhgxFxN0xVZ/ehk2qvkWMnQSKRoH7Dr/Mh19+SNnXLYFqfupiy8TzcB21E6L0Y7J3cGjZm8jTrt6tfFpO618HUjefh2mctfpp7BG3qlsHEbh6qOkv8vkPDKvboMfMQqvdbhz+uPMD+qT+iiFX+nDosyqKjhw9i3qxp6Nm3P9Zu3oFSZcrBr38fxKZzfQi9dhWBAT+juU9rrNuyE3UbNMIvQwch/KPrQ6qTx//AP6EhsLGxze7DyNOkEolOXnmZziKD33//HYsXL8aPP/4IAwMDeHp6YsyYMZg6dSo2btyoq2a+SOodeIsWLcLx48cxd+5cXL/+ITMikUhUwdTDhw9RvHhxlChRQp/d/ept3bgOzX3aoFmLVnAs6YThAWNhYmKC/XuD06y/Y8sG1HT3QMeuPeDg6IRe/QahTDkX7Nq2SVXHytpa7fXXqROoUr0mihQrnlOHRVk0uFVVrD74D9Yf/Re3HsZi0II/8D4xBb6NK6RZv5ZzEZz79ym2nryNh89f4diVh9h28jaqly0EADAxksGnTmmMXnUaZ/55gohn8Ziy8TzCn75E72aVc/LQKAs2r1+Dlq3/h+Y+rVHSqRRGjhkHExMT/L57V5r1t25aj1q166BLt55wLOmEnwYMRllnF2zfov758jwqCjN/nYKJU6er7hQnyi46C6JiY2NRsmRJAICpqSliY2MBAHXq1MGff/75uU1zjEQigRAC3t7eWL58OY4fP45Zs2YhNPRDClkqlSI5ORkXLlyAm5sb8ufnt9msSk5Oxp1b/6KaWy1VmVQqRfWatXAjNCTNbf4JDUH1mu5qZTXda+Of62nXj30Rg3N//YkfWrbWXccpWxgaSFGltB2OX3uoKhMCOH7tIWo6F05zm/M3n6JKKVvVkJ9DITN413DAoUv3AAAGMikMZFIkJKeobZeQlILa5XP/Ar/fsuTkJNy6+S9qfnJ9qOHmjuuh19Lc5nroNdRwU78+1HL3wPWPridKpRLjx4xEZ98eKFmqdLb0/VvC4TztdBamlyxZEvfu3UOJEiVQrlw5bNu2DTVr1sTvv/+ueiBxbpCSkgJDQ0N4e3tj2bJl6Nu3LwDA398flSpVwpQpU7Bv3z4cP34cxsbGeu7t1yv+ZRwUCgUsLa3Uyi0srfDg/r00t4l9EaNR39LSGrEvYtKsf3DfXuTLnw91G3jpptOUbaxN5TCQSfE87p1a+fO4dyhbLO05k1tP3oaVqRzHZraDRAIYGsiwfH8IZmy9BAB48z4Z5/99ioAObrj9MBZRL9+hbb2ycCtXGOHPXmb3IdEXeBn38sP1wcpardzSygoP7kekuc2LmBhYWn1yfbCyxouY/64P61avhEwmQ7uOnXXf6W8Q787TTmdBVPfu3VVLB4wcORLNmzfHwoULkZycjNmzZ+uqmQz7eLJ4KoVCAUNDQzx+/Bjx8fFo0qQJli9fjj59+kAulyMlJQUbNmzAmTNnUK5cuWzrW2JiIhITE9XLkqQM2jLpwN5gfNfkB75veZRnxWL4uV1N+C06jku3n8GpiDlm9q2PZx3c8OvmCwCAHjMPYdnQxojY2AcpCiWu3X2Obaduo0opzoX51tz89wa2blqPdZt38sOfcozOgqihQ4eq/tvLywu3bt3C5cuXUapUKVSqVElXzWRIREQEdu7ciZs3b2LmzJmwtLSEUqmETCbDgwcP4OLiAn9/f0ycOBHe3t5YsWIFWrduDaVSibNnz6JKlSrZ2r+goCBMmDBBrWz4yDH4edTYbG03J5mZW0Amk2lMEo2LfQGrT759prK0staoHxsbo/FtFQBCrl7Gwwf3MCFohu46Tdkm5tV7pCiUsLXIp1Zua5EPkZ9kp1KN61obm4/fxJrD/wAAbtx/gXzGhlg02AvTtlyAEMC9Z/Fo/Mt25DM2gGk+Y0TGvcX6kd/jXmR8th8TZZ25hfmH68MnWebYFy9gaZ329cHK2hqxLz65PryIgdX/17925TLiYmPRsmkj1c8VCgXmz56OrRvXYffBP3R8FHkfb6fSLtveI3t7e7Ru3TrHA6jr16/D29sb9+7dg5mZGfLl+3DRlkqleP78OWrVqoWuXbtiwoQJqjlSjRs3xpEjRxASEpLtARQABAQEID4+Xu01eNiIbG83JxkaGqJMORdcvnhBVaZUKnH50gWUr5T2pN8KlSrj8qXzamV/XziHChU16+/bswtlnV1Qqkz2ZQxJd5JTlLgaFoUGrv/dACCRAA1ci+PizWdpbiM3NoBSCLUypVL8/7bqmYZ3iSmIjHsL8wLG8Kpmj33n0x4SotzB0NAI5ZxdcOnif3/vSqUSly6eR8VKrmluU7GSK/6+qH59uHj+HCr+//Xk+x9aYOP23Vi/dZfqZWNji86+PTBvyYpsO5a8LPWGqy995WVflImaP39+husOHjz4S5rKkLCwMDRq1Ag9evTAlClTIJN9WE8mdWjvn3/+QUBAAAYNGqTxi61du3a29y+VsbGxxhBUwuvkHGs/p7Tr1BVTx49GOZfycC5fAds3bcD79+/xfXMfAMDksQGwtrXFTwM/ZDHbtO+MQX26Y8uGNXCvUxfHDh/ErX9v4OdR49X2+/bNG5z84wgGDBmew0dEX2J+8BWsGOaNy2HP8fftSAz0qYJ8xoZYd/QGAGDlMG88ffEGY9ecAQAcuBCBwa2rIiT8OS7eioRTEXOM7VobBy5EqIIpr6r2kEiAO4/j4FTEHFN7euLO4zisO3JDb8dJGdOhSzdMDAyAs0sFuFSoiC0b1yHh/Xv80LIVAGD8mJGwsbXFgMH+AIB2Hbvgp16+2LhuNTw86+HooQO4+e8/CBj7IatvZm4Os0/m3xoYGMDSyhr2Do45emz07fiiIGrOnDkZqieRSLI9iFIoFJg3bx68vLwwbtw4VQCV2j4ANGzYEA0bNtToG2WPRo2b4mVcHFYtXYjYFzEoVaYcZi5Yqhqei4p8BslH629VrFwF46ZMw4rFC7B80TwUK26PqTPna9xlc+zIQQgh4NXk+xw9HvoyO/68A2szOcZ2doedZT6EhkejZWAwnr/8MJxX3LagWubp180fhuzGdfVAEasCiIl/h/0XIjB+7VlVHbP8xpjY3QNFrQsg9nUi9vwVhnFrzyBFoczx46PM+c67KV7GxWL5kgV4ERODMmXLYe7iZarh/qhnzyCV/Hd9qORaBZOmTsfSRfOxZMFcFC9hj+lzFsCJd+FlGyk/HrWSCPFJvvwrVr16dTRu3BhTp07V+JlCoYBMJoNSqYRUKlX9f27xPA9moijr7Nsu1HcXKJd5tiv7s/n09TCXy7RX+kL+e2/pZD+zW+TdaRe5J4r4AkIIvHnzBo8fP4bV/98C+2lsmBpA+fn54eXLl7kqgCIiIsptOCdKuzwRSQghYGJiAldXVwQHByMiIkL1i/s4mIqIiMDly5fx7FnaE1mJiIiIMipPBFFSqRQGBgbw8fHB2bNnsWrVKjx9+hSA+pyn9evXQy6Xo1ChQvrqKhER0VdBKtHNKy/7Kh8sFB4ejk2bNuHChQswNjZG8eLFMWbMGPz0008IDw9HUFAQ4uPj0blzZ9SqVQuhoaFYvXo11q5di1OnTsHCIu0VkomIiOiDPD4SpxNfXRAVGhqKxo0bo0aNGjA3N0dkZCSOHz+Offv2Yd68eZg2bRrkcjnmzZuHFStWwNLSEubm5jA2Nsbx48dRsWJFfR8CERER5QE6DaJOnz6NZcuWITw8HDt27EDRokWxfv16ODo6ok6dOl+8/wcPHqBZs2bo0aMHJk2aBJlMhpSUFISGhqJPnz7o27cvdu7ciYkTJ6JRo0aIiopCREQE3Nzc4OzszGE8IiKiDJIyFaWVzoKonTt3okuXLujUqROuXr2qejZcfHw8pk6digMHDnxxG3v37kX58uURGBgImUwGIQQMDAxQtWpVbN26FS1atMDQoUNx9uxZ1KtX74vbIyIi+lbliUnT2Uxn79HkyZOxdOlSrFixAoaGhqpyDw8PXLlyRSdtXL16FQqFAnK5XOMBw46OjvDz88PVq1d11h4RERFRenSWibp9+zbq1q2rUW5mZoaXL1/qpA2JRKJ2193HgZRUKoWXlxcSExPx4pOHVBIREVHmcDRPO51logoVKoS7d+9qlP/1118oWbLkF+07da0nT09PREZGYvHixQA+BFIpKSkAPjy8MiYmBi4uLnB05HOSiIiIvoRUItHJKy/TWRDVu3dv+Pn54cKFC6qM0caNGzF8+HD069cv0/tLSEjQKGvWrBmKFSuGWbNmYdOmTQA+PGAS+JCJCg4ORv78+WFpafllB0NERESkhc6G80aOHAmlUolGjRrh3bt3qFu3LoyNjTF8+HAMGjQoU/t68uQJhg4din79+qFBgwaqjJONjQ12796NunXrYuTIkTh9+jQGDhyI8PBw/Pnnn1i2bBlOnz7NIIqIiOgL5fEkkk7oLIiSSCQYPXo0fv75Z9y9exdv3ryBi4sLChQokOl9JSYm4vHjx5g1axaMjIzg4eEBAwMDJCcnw9HREWfOnMGIESOwe/durFixAo6OjihRogTOnDmDSpUq6eqQiIiIvll5fbVxXZCIT5/Um0uEhYVh8ODBEEIgMDAQHh4eAICkpCQYGRnh/fv3SExMxOHDh9G8eXMkJyfDzMxMz73Ouuevk/XdBcpF7Nsu1HcXKJd5tmuwvrtAuYi5XJbtbUw8qjnPOSvGfldKJ/vJjXSWiUoddkvP8ePHM7W/0qVLY/78+Rg8eDAmTZqkCqSMjIygUCgglUoxdepUREREoHXr1siXL9+XHgIRERFRhulsYrmrqysqV66serm4uCApKQlXrlzJ8qNWUgMpiUSCSZMm4cyZMwAAhUKB4cOHY/bs2Rg9erTaulRERET05SQS3bzyMp1loubMmZNm+fjx4/HmzZss7/fTjNTIkSNx8OBBrFq1CpcuXUKVKlWyvG8iIiJKG+dEaZftq7p37twZv/322xftIzWQMjQ0ROvWrTF37lz89ddfDKCIiIhIb7I9iDp37hxMTEy+eD+lS5fGzJkz4enpiStXrqBq1ao66B0RERGlRaKj/+VlOhvOa926tdq/hRB49uwZ/v77bwQGBuqkjbJly2LHjh2cA0VERJTNOJynnc6CqE+XF5BKpShbtiwmTpyIxo0b66oZBlBERESUK+gkiFIoFOjevTsqVqwICwsLXeySiIiI9IiZKO10MidKJpOhcePGePnypS52R0RERHomkUh08srLdDaxvEKFCoiIiNDV7oiIiIhyNZ0FUZMnT8bw4cOxb98+PHv2DK9evVJ7ERER0ddDKtHNKy/74iBq4sSJePv2Lb7//nuEhISgRYsWKFasGCwsLGBhYQFzc3POkyIiIvrK6HPF8kWLFsHBwQEmJiZwc3PDxYsXM7Tdli1bIJFI4OPjk7WGM+mLJ5ZPmDABP/30E06cOKGL/hAREVEuINXTfKatW7fC398fS5cuhZubG+bOnQtvb2/cvn0btra26W53//59DB8+HJ6enjnW1y8OooQQAIB69ep9cWeIiIjo2zZ79mz07t0b3bt3BwAsXboU+/fvx2+//YaRI0emuY1CoUCnTp0wYcIEnD59OsdudNPJnKi8PvueiIjoW6OrOVGJiYka86QTExPTbDMpKQmXL1+Gl5fXf/2QSuHl5YVz586l29eJEyfC1tYWPXv21Pn78Dk6CaLKlCkDS0vLz76IiIjo66GrOVFBQUEwMzNTewUFBaXZZkxMDBQKBezs7NTK7ezsEBkZmeY2f/31F1atWoUVK1bo/D3QRieLbU6YMEFjxXIiIiKigIAA+Pv7q5UZGxvrZN+vX79Gly5dsGLFClhbW+tkn5mhkyCqffv2n53sRURERF8XqY4eHmxsbJzhoMna2hoymQxRUVFq5VFRUShUqJBG/fDwcNy/fx/NmzdXlSmVSgCAgYEBbt++DScnpy/o/ed98XAe50MRERHlPfpY4sDIyAjVqlXDsWPHVGVKpRLHjh2Du7u7Rv1y5crh+vXruHbtmurVokULNGjQANeuXUPx4sW/9G34LJ3dnUdERET0pfz9/eHr64vq1aujZs2amDt3Lt6+fau6W69r164oWrQogoKCYGJiggoVKqhtb25uDgAa5dnhi4Oo1LQZERER5R36Wm28Xbt2iI6OxtixYxEZGQlXV1ccOnRINdn84cOHkEp19sCVLyIRTCXlCs9fJ+u7C5SL2LddqO8uUC7zbNdgfXeBchFzuSzb21h+/oFO9tOnlr1O9pMb5Y5QjoiIiOgro5O784iIiChv4X1j2jGIIiIiIg36enbe14RBFBEREWlgDKUd50QRERERZQEzUURERKSBWRbtGEQRERGRBj6RRDsGmkRERERZwEwUERERaWAeSjsGUURERKSBSxxox+E8IiIioixgJoqIiIg0MA+lHYMoIiIi0sDRPO04nEdERESUBcxEERERkQauE6UdgygiIiLSwKEq7RhEERERkQZmorRjoElERESUBcxEERERkQbmobRjEEVEREQaOJynHYOoXOLl22R9d4FykXPLeum7C5TLNJ5zWt9doFzk4qj6+u4CgUEUERERpYGTprVjEEVEREQaOJynHQNNIiIioixgJoqIiIg0MA+lHYMoIiIi0sDRPO04nEdERESUBcxEERERkQYpB/S0YhBFREREGjicpx2DKCIiItIgYSZKK86JIiIiIsoCZqKIiIhIA4fztGMQRURERBo4sVw7DucRERERZQEzUURERKSBw3naMYgiIiIiDQyitONwHhEREVEWMBNFREREGrhOlHYMooiIiEiDlDGUVhzOIyIiIsoCZqKIiIhIA4fztGMQRURERBp4d552DKKIiIhIAzNR2nFOFBEREVEWMBNFREREGnh3nnYMooiIiEgDh/O043AeERERURYwE0VEREQaeHeedgyiiIiISANjKO04nEdERESUBcxEERERkQYpx/O0YhBFREREGhhCacfhPCIiIqIsYCaKiIiINDEVpRWDKCIiItLAxTa1YxBFREREGjivXDvOiSIiIiLKAmaiiIiISAMTUdoxiCIiIiJNjKK04nAeERERURYwE5UJ0dHRuHfvHqRSKVxcXJAvXz59d4mIiChb8O487ZiJyqDr16+jfv366NatG2rWrIlBgwbh5s2b+u4WERFRtpBIdPPKyxhEZUBoaChq1aqFFi1aYMeOHVi8eDHWrFmDI0eO6LtrREREpCcMorS4efMmatSogX79+iEoKAguLi7o0KEDChUqhAMHDkChUKjVVyqVeuopERGR7kh09MrLGERpcfnyZSQnJ6NEiRJ49eoVAGDRokV49uwZ3r17h+HDh2P16tU4deoUAECS13OXRET0bdBjFLVo0SI4ODjAxMQEbm5uuHjxYrp1V6xYAU9PT1hYWMDCwgJeXl6fra9LnFiuRefOnRETE4MhQ4bA2NgYUVFRmDNnDlatWgVzc3PcuXMHq1evxv3792FtbY3OnTtj6NChDKaIiIiyYOvWrfD398fSpUvh5uaGuXPnwtvbG7dv34atra1G/ZMnT6JDhw6oXbs2TExMMG3aNDRu3Bg3btxA0aJFs7WvEiGEyNYW8ohZs2bh559/hkQiQXBwMFq0aKH6WXR0NB4/foy5c+dizJgxKF26dKb3fyfynS67S1+5d0kK7ZXom9Jnw2V9d4FykYuj6md7G1cfvNbJfqrYF8xUfTc3N9SoUQMLFy4E8GGaTPHixTFo0CCMHDlS6/YKhQIWFhZYuHAhunbtmqU+ZxQzUZ949OgR/vjjDyiVSjg7O6N27doAgGHDhsHU1BR9+/ZFeHg44uPjYWZmBgCwsrKCjY0N1q5dq8+uExER6YyuBlQSExORmJioVmZsbAxjY2ONuklJSbh8+TICAgJUZVKpFF5eXjh37lyG2nv37h2Sk5NhaWn5ZR3PAM6J+khoaCg8PT2xZMkSjBgxAr6+vggODlb9vHfv3pg5cyaGDRuGZcuWqeZISaUf3kYm9YiIKK/Q1ZSooKAgmJmZqb2CgoLSbDMmJgYKhQJ2dnZq5XZ2doiMjMxQv0eMGIEiRYrAy8srk0eceQyi/l9oaCjc3d3Rvn17nDhxAps2bcLbt2+xevVqvH//HsnJyQAAf39/zJw5E4GBgZgzZw5ev/4v3cl5UEREROoCAgIQHx+v9vo406RLv/76K7Zs2YLg4GCYmJhkSxsf43AePgzhNWrUCM2bN8evv/4KAGjcuDGKFCmCsLAwpKSkQC6Xq+r7+/vj9evXmDdvHgYOHKivbhMREWUfHeUF0hu6S4u1tTVkMhmioqLUyqOiolCoUKHPbjtz5kz8+uuv+OOPP1CpUqUs9zczmIkCEBYWBgcHByQkJKjGXIOCgnDlyhWYmpqic+fO6NWrF5YsWYJHjx5BoVBg3LhxCAsLg5WVlZ57T0REpHsSHf0vM4yMjFCtWjUcO3ZMVaZUKnHs2DG4u7unu9306dMxadIkHDp0CNWrV8/yMWcW7877f/v378eCBQsgl8tha2uL4OBgLFq0CDVr1sTff/+NGzduYOHChTAxMUHlypWxd+9eCCFU86G+FO/Oo4/x7jz6FO/Oo4/lxN15oY/e6GQ/lYoXyFT9rVu3wtfXF8uWLUPNmjUxd+5cbNu2Dbdu3YKdnR26du2KokWLquZVTZs2DWPHjsWmTZvg4eGh2k+BAgVQoEDm2s6sb344LyUlBQYGBmjWrBmUSiXmz5+PQ4cOYcqUKfjf//4HALC3t8ePP/6Ifv364cSJE6hatSokEgnnQBERUZ6lr4+4du3aITo6GmPHjkVkZCRcXV1x6NAh1WTzhw8fqiUwlixZgqSkJLRp00ZtP+PGjcP48eOzta/fZCYqOjoacrlcFaEmJyfD0NAQAHDo0CHMnj0bRkZGGD16tCp9+HGd7MBMFH2MmSj6FDNR9LGcyET981g3magKxbI3G6RP39ycqIcPH8Le3h5+fn7o378/3rx5o5ZRatKkCQYOHIiUlBRMnDgR58+fB4BsDaCIiIjo6/PNDec9ePAACQkJqFy5Ms6fPw93d3c0adIEzZo1Q/369QEALVq0gFKpxIoVK+Dv74958+ahRo0a+u34V2p/8Fbs2rIWcbEv4OhUBn39RqCMc4U06z64F46Nvy1G+J2beB75DL0GDkfL/3VSq7Np9VJsXrNMraxoCQcsXR8M+voc3rMNv29fj5exL2DvVBrdB/yMUuXSPj+OHQjGn0f349H9cACAY2lndOjRP936lPu1qVYEnd1KwKqAEcKi3mDmkTD8+yz9VbILGBugX31HNChrDVMTQ0TGJ2D2H3dxNjxWVcemgBEGNnRC7ZKWMDaU4nHce0zadxs3I3Wz+vY3hTNWtPrmgihPT08MGDAADx48wKZNm7By5UrcvXsXLVq0QOfOnVGrVi107doVPj4+MDExwbJlyzQW/aKMOX38MFYumoUB/qNRxqUC9m7fhLHD+2Ppht0wt9BcSTYxIQGFihRDnfrfYeXCWenut4SjEybPWqr6t1Qmy5b+U/Y6e/II1i2bg16DA1DauQIO7NqMqQGDMOe3nTBL4/y4EXIZtRt4o6xLJRgaGWPP1rWYMnIgZq3cBktrzedpUe7m5WyDIY1K4ddDd3Dj6Su0r1EM89tXwv+WXUTcu2SN+gZSCRZ2qITYd8kYuesGol8noZCZMd4kpKjqFDQxwIquVXH5QRz8tobi5btkFLeU41WC5v5Iu8zeWfct+qaCKCEEJBIJ3Nzc8Ntvv+H169fo1asXgA8PMNy3bx8OHTqEpUuXon379ujZsyfq16+fIwt25UW7t22A9w+t4fV9SwBA/2Gjcen8aRw9sBv/69RDo34Z5/Io41weALB2+fx09yuTyWBhZZ09naYcs3/nRjRq6oMGTT48h7KXXwCuXPgLJw7vhU/7bhr1BwdMVvv3T/5jcPGv47h+9SLqffdDTnSZdKhjzeLYfe0Z9oV+WIX614N34FHKCs0rF8a6cw816reoXBimckP0XHcVCuWHqbzP4hPU6nStVQLPXydg0v7bqrKnn9Qh0qVvak5U6tynzp07IzIyUrWwpq+vLx4/foytW7fixIkTsLKywubNmxEXF8cAKouSk5Nx985NVK7mpiqTSqVwreaG2zdCv2jfTx8/hG/r79Cr/Q+YOWkUnkc9+9LuUg5LSU5GxJ1bqFhV/fyoWLUmwv7N2PmRmJiAlJQUFCholl3dpGxiIJWgXOGCuHQ/TlUmAFy6F4eKRU3T3MaztBWuP3mFX7xL46BfbWzuXQPdapeA9KNkiWcZK9x89hpBrVxwyK821veohpauhbP5aPIuiUQ3r7zsm8pEAR+e7iyTyTB+/Hhs374djRo1wr///ot9+/ahWrVqAIDff/8dcXFxsLCw0HNvv16v4uOgVChg8cmwjLmFFR4/vJ/l/ZZxroAhIyeiaAl7xL2IweY1yzByUA8sXLMD+fLl/8JeU055Ff8SSqVCY9jOzMISTx/dz9A+Nq5cAEsra1SsWjMbekjZyTyfIQykEsS+TVIrj32bBHurfGluU9RCjupmJjj8TxSGbg1FMQs5RniXgYFUgpV/PfhQx1yO1lWLYtOFR1h99iFcChfEsO9KIUWhxP7rUWnul9KXx+MfnfimgqjUAAoAypcvj19++QVv3rzBtWvXUKxYMQAfVkaVSqXZGkCl9UTrpEQFjDK4LP63rHqtOqr/dnQqgzLOFdGz3ff468QRNG7WSo89o5y0e8sanD15BONmLoOREf9uvgVSAHFvkzD14G0oBXAr8g1sCxqjc63iqiBKKgFuPnuNJafuAQDuRL2Bk01+tK5ShEFUVjCK0ipPDueltfRVagD15MkT3LhxA+XLl8fIkSNhYWGBlJT/JibqagXyz0nridbLFszM9nZzkqmZBaQyGeLiYtXKX8a9gIWl7h6VU6BgQRQpVgLPnjzS2T4p+5mamUMqlSH+k/MjPi4W5hafPz9+374ee7asweighbAvWTo7u0nZ5OW7ZKQoBSzzG6mVW+Y3wotPslOpYt4m4WHseyg/urzfi3kH6wLGMPj/Mb2YN0m4F6O+5t79F+9gZ8ZpGZQ98lwQFRERgZkzZ6JHjx6Ijf1wgVYqlZDJZHjw4AHKlCmDLVu2AABq1qwJS0tLnDp1Kkf7mNYTrfsOGp6jfchuhoaGKFXGGaGXL6jKlEolQq5cRNnyunsw5Pt37xD59DEsLDnR/GtiYGiIkmXK4frVi6oypVKJf65eQmmX9M+PPVvXYueGlQiYugBOZV1yoquUDVKUAreevUYNB3NVmQRAdQcLXH/yKs1tQh7Fo5iFXC05UsJKjujXiUj5/8gq9HE87K3katuVsJQjkpPLs0Qfz8772uSpIOr69evw9vbGvXv3YGZmhnz5PoytS6VSPH/+XLV8wYQJEwAAVatWhYWFBRYuXKiWjcpuxsbGMDU1VXvlxaE8n7adcXh/MI4d2otH9yOwePZUJLx/D6+mH+7Wmz1ljNpdeMnJyYgIu42IsNtISU7Gi5jniAi7jaeP/7tTZ9Xi2bh+7W9EPXuKm/9cw9Qx/pBKpajn1STHj4++TLMfO+H4gd04dWQfHj+4h5Xzg5CY8B71vZsDABZOG4tNqxaq6u/Zsgbb1i5Fv+FjYVuoMF7GxuBlbAwS3nO1/6/RpouP0NK1CJpVtIODVT6MaFoGckMp9oV+uFFkfPNy6F/fUVV/55WnMJUbYFjjUihhKYeHkyW61bbHjstPPtrnY1QoYoputUugmIUc3i628HEtgu0f1aGM48Ry7fLMnKiwsDA0atQIPXr0wJQpU1Rzn1KXNfjnn38QEBCAQYMGQSKRqIb3Fi9eDCEEDAzyzFuRa3g29Eb8yzhs/G0J4mJfoGSpspgwY5FqOC/6eSQkHw2fxsZEw69Xe9W/g7esQ/CWdajgWg1B81YCAF5ER2HmxAC8ehUPM3MLuFR0xcwl62BmrrmuEOVutes3xquXcdi2dilexr2Ag1MZBExdoBrOe/E8ElLJf+fH0X07kZKcjNkTR6jtp02X3vhf17452nf6cn/cjIZFPiP0qesIq/xGuBP1Bn5bQxH79sOaTnamJmpDd89fJ8JvSyiGeJXCxl5FEP06EVsvPVZbDuHms9f4ZecN9K/viJ51HPD05XvM/uMuDt94ntOHR9+IPPHsPIVCAT8/P8TGxmLVqlWQy+XaN8pl+Ow8+hifnUef4rPz6GM58ew8XX0ulSmU9h2XeUGeGM6TyWQ4f/48HBwc0gygFIoPH0hKpVLt/4mIiCgdEh298rCvPogSQuDNmzd4/PgxrKysVGUfk8lkUCqV8PPzw8uXL3PkDjwiIiLK2776aEIIARMTE7i6uiI4OBgRERGqlck/DqYiIiJw+fJlPHvG1a2JiIi04d152n31QZRUKoWBgQF8fHxw9uxZrFq1Ck+fPgXw32NeAGD9+vWQy+UoVKiQvrpKRET01eDdedp9dbekhYeHY9OmTbhw4QKMjY1RvHhxjBkzBj/99BPCw8MRFBSE+Ph4dO7cGbVq1UJoaChWr16NtWvX4tSpU3yUCxEREenEVxVEhYaGonHjxqhRowbMzc0RGRmJ48ePY9++fZg3bx6mTZsGuVyOefPmYcWKFbC0tIS5uTmMjY1x/PhxVKxYUd+HQERE9FXI40kknfhqgqgHDx6gWbNm6NGjByZNmgSZTIaUlBSEhoaiT58+6Nu3L3bu3ImJEyeiUaNGiIqKQkREBNzc3ODs7MxhPCIiosxgFKXVV7NO1IIFC7B//34EBwdDLperFtEEPgzxtWjRAmZmZjh79qyee5o1XCeKPsZ1ouhTXCeKPpYT60RFROvmcTklbfLuswu/monlV69ehUKh0AigAMDR0RF+fn64evUqrly5osdeEhER0bfiqxnOk0gkanfdfRxISaVSeHl5ITExES9evNBnN4mIiPKEvH5nnS7k+kxU6mijp6cnIiMjsXjxYgAfAqnUhwYrlUrExMTAxcUFjo6O6e6LiIiIMoYLlmuXK4OohATNcdhmzZqhWLFimDVrFjZt2gQAqocGS6VSBAcHI3/+/LC05INoiYiIKPvluiDqyZMn6Nq1K06cOAHgv4yTjY0Ndu/ejaSkJIwcORL9+vXDjRs3sHfvXgwfPhwLFy7EsmXLGEQRERHpAlNRWuW6ICoxMRGPHz/GrFmzcObMGQAfMk7JyclwdHTEmTNn4OHhgd27d6Ny5coYNmwYrl69ijNnzsDV1VW/nSciIsoj+NgX7XLlEgdhYWEYPHgwhBAIDAyEh4cHACApKQlGRkZ4//49EhMTcfjwYTRv3hzJyckwMzPTc6+/DJc4oI9xiQP6FJc4oI/lxBIHD14k6mQ/9lbGOtlPbpTrMlEAULp0acyfPx8SiQSTJk1SZaSMjIygUCgglUoxdepUbN++HYaGhl99AEVERJTb8Nl52uXKIApIP5BSKBQYPnw4Zs+ejdGjR8PQ0FDPPSUiIsp7OCVKu1wbRAGagdTJkycRGBiIVatW4dKlS6hSpYq+u0hERETfqFy/2GZqIOXv74/WrVvj7du3OHfuHAMoIiKibJTXh+J0IVdnolKVLl0aM2fOhKenJ65cuYKqVavqu0tERER5HAf0tMn1mahUZcuWxY4dOzgHioiIKAcwE6XdV5GJSsUAioiIiHKLryYTRURERDmHiSjtGEQRERGRBg7nafdVDecRERER5RbMRBEREZGGvP7cO11gEEVERESaGENpxeE8IiIioixgJoqIiIg0MBGlHYMoIiIi0sC787TjcB4RERFRFjATRURERBp4d552DKKIiIhIE2MorRhEERERkQbGUNpxThQRERFRFjATRURERBp4d552DKKIiIhIAyeWa8fhPCIiIqIsYCaKiIiINHA4TztmooiIiIiygEEUERERURZwOI+IiIg0cDhPOwZRREREpIF352nH4TwiIiKiLGAmioiIiDRwOE87BlFERESkgTGUdgyiiIiISBOjKK04J4qIiIgoC5iJIiIiIg28O087BlFERESkgRPLteNwHhEREVEWMBNFREREGpiI0o6ZKCIiItIk0dErCxYtWgQHBweYmJjAzc0NFy9e/Gz97du3o1y5cjAxMUHFihVx4MCBrDWcSQyiiIiIKNfYunUr/P39MW7cOFy5cgWVK1eGt7c3nj9/nmb9s2fPokOHDujZsyeuXr0KHx8f+Pj44J9//sn2vkqEECLbWyGt7kS+03cXKBd5l6TQdxcol+mz4bK+u0C5yMVR9bO9jffJutmP3DBz9d3c3FCjRg0sXLgQAKBUKlG8eHEMGjQII0eO1Kjfrl07vH37Fvv27VOV1apVC66urli6dOkX9V0bZqKIiIhIg0Sim1dmJCUl4fLly/Dy8lKVSaVSeHl54dy5c2luc+7cObX6AODt7Z1ufV3ixHIiIiLKNomJiUhMTFQrMzY2hrGxsUbdmJgYKBQK2NnZqZXb2dnh1q1bae4/MjIyzfqRkZFf2HPtGETlEmUK5dN3F/QuMTERQUFBCAgISPOPi749PCf+kxPDN7kdz4ecZaKjCGH85CBMmDBBrWzcuHEYP368bhrQIw7nUa6RmJiICRMmaHxjoW8Xzwn6GM+Hr1NAQADi4+PVXgEBAWnWtba2hkwmQ1RUlFp5VFQUChUqlOY2hQoVylR9XWIQRURERNnG2NgYpqamaq/0MolGRkaoVq0ajh07pipTKpU4duwY3N3d09zG3d1drT4AHD16NN36usThPCIiIso1/P394evri+rVq6NmzZqYO3cu3r59i+7duwMAunbtiqJFiyIoKAgA4Ofnh3r16mHWrFlo1qwZtmzZgr///hvLly/P9r4yiCIiIqJco127doiOjsbYsWMRGRkJV1dXHDp0SDV5/OHDh5BK/xtIq127NjZt2oQxY8Zg1KhRKF26NHbv3o0KFSpke1+5ThTlGpw0Sp/iOUEf4/lAuQ2DKCIiIqIs4MRyIiIioixgEEVERESUBQyiiIiIiLKAQRQRERFRFjCIolyD9zgQEdHXhEEU6d379++RmJiIR48eISEhQd/doTyKQToR6RqDKNKrmzdvonPnzqhevTqcnJzg7u6OkSNH6rtb9JWLiIhAYGAg+vXrh/nz5wMAJBIJAyki0imuE0V6c/36dXh6eqJz586oUqUKLC0tsXbtWhw6dAjfffcddu3aBUNDQ313k74yISEhaNKkCZydnfH06VPcuXMHvXv3xrJly/TdNdKT6Oho3Lt3D1KpFC4uLsiXL5++u0R5BIMo0ovo6Gh4e3vD29tb9fyj1PJt27ZhxIgR+OGHH7BlyxY99pK+NtevX0etWrUwdOhQTJgwAc+fP8e2bdswdOhQbN26Ff/73//03UXKYdevX0f79u0hhMCtW7fQvXt3DB8+HM7OzvruGuUBHM4jvXj8+DGSk5PRqVMnKBQKAB+e1G1jY4POnTsjICAAv//+O3bv3q3fjtJXIy4uDm3atEHp0qUxefJkyGQyFC5cGE2aNIGdnR2Sk5P13UXKYaGhoahVqxZatGiBHTt2YPHixVizZg2OHDmi765RHsEgivQiJCQEd+/eRYUKFSCTySCEUD1Q0szMDB07doShoSHu3r2r557S10KhUKBjx454/PixWnZToVAgPj4eFhYWeuwd5bSbN2+iRo0a6NevH4KCguDi4oIOHTqgUKFCOHDggOrLWyqlUqmnntLXzEDfHaBvU6lSpQAAO3fuxI8//giJRKL2c0dHR5QsWRJPnjzRR/foK3Lnzh3Ex8ejRo0aGDBgAExMTBAUFAQzMzP06NEDjRs3Ru/evdG0aVN9d5Vy0OXLl5GcnIwSJUrg1atXMDU1xaJFi/Ds2TOULFkSw4cPR6VKlVCyZEnUq1dP4xpElBEMokgvHBwcYGpqinXr1qF69eqwt7cH8OHboFQqRVxcHORyOapVq6bnnlJuFhISgipVqmDevHmoUaMGrK2t0bNnTwghMGrUKAwePBhDhw7FjBkzIIRQy3hS3ta5c2fExMRgyJAhMDY2RlRUFObMmYNVq1bB3Nwcd+7cwerVq3H//n1YW1ujc+fOGDp0KIMpyhQGUaQXxYoVw5IlS9ChQwcEBgZixIgRKF++vOoDbvbs2Xj69Ck8PT313FPKrUJCQlC7dm2MGjUKgwYNUpVbW1uja9euMDQ0xNSpU1XDeBKJhEM235ghQ4ZAoVCgX79+kEgkCA4ORosWLVQ/79GjBx4/foy5c+eiefPmDKAo0xhEkd60bNkS8+fPx8CBA3Hx4kV4eHigcOHCuHfvHg4ePIhjx46pMlREHwsNDYWHhweGDBmCyZMnq8o3b96MDh06oEiRIujSpQtSUlIwdepUGBgY4JdffoFMJtNjryk7PXr0CH/88QeUSiWcnZ1Ru3ZtAMCwYcNgamqKvn37Ijw8HPHx8TAzMwMAWFlZwcbGBmvXrtVn1+krxiCK9EYmk6Fv375wdXXF9OnTceHCBZibm6Ny5co4e/YsypUrp+8uUi709OlTuLq6olevXpgyZYqqfNq0aQgICEDp0qVRvXp12NraomfPnpBKpRgxYgSMjIwwZMgQ/XWcsk1oaChatGgBW1tbREREwMLCAtOnT0erVq0AAL1798br168xbNgwJCcn46effoKpqakq8y2EYBaKsoRBFOmdm5sbtm3bBqlUqhpy4bwVSo+xsTEqV66MS5cuISwsDKVLl8avv/6KmTNn4vDhw6hevbrqQ9Ha2ho9evSAoaEhJ5bnUaGhoXB3d8egQYMQGBiIM2fOoFu3bli9ejWaNGkCAwMDGBoawt/fHwAQEBCA9+/fw9/fHwULFgQABlCUZVxsk3KFj78J8lshpSf13Hjx4gWaNm2K5ORkNGrUCGvXrsWmTZvw3XffqdW/ceMGypYtC5lMxnMqD3r06BGqVq2KRo0aqS3MW716dbx9+xYXL15UBUqpJkyYgHnz5iEsLAxWVlY53WXKY/h1n3KFjz/g+GFH2lhZWeHQoUMwNTXF7NmzMX36dI0AasSIEWjfvj1ev37NcyqPCgsLg4ODAxISEnDu3DkAQFBQEK5cuQJTU1N07twZvXr1wpIlS/Do0SMoFAqMGzeOARTpDDNRRJSrPXv2DOfPn8fhw4chhEDTpk1Ro0YNFC1aFLGxsfjhhx/w6tUr7Ny5E2XLlgUAjBs3DjNmzMDJkydRs2ZNPR8BZaf9+/djwYIFkMvlsLW1RXBwMBYtWoSaNWvi77//xo0bN7Bw4UKYmJigcuXK2Lt3L5e6IJ1hEEVEudaNGzfQrVs3mJmZITIyEnK5HJcvX0bLli3h7+8PT09PxMXFwdvbG2/evMEff/yBFStWICgoCGfOnOE6Y3lYSkoKDAw+TOv9/fffMX/+fPz111+YMmWKav5TqujoaJw4cQJVq1ZVLfRLpAucWE5EuVJISAjq1auHPn36oG/fvnBycgIArF69GsOGDcP79+8xZcoUVKtWDQcPHkTz5s1RrFgxyOVyBlB5VHR0NORyOQoUKAADAwMkJyfD0NAQzZs3h6GhISQSCY4fPw53d3e4u7sDAJKTk2FjY4O2bdvqufeUFzGfSUS5zo0bN1CnTh34+flh+vTpcHBwAPBhRfvu3btj4cKFOHHiBLZu3QrgwxypPXv2oF27dgyg8qiHDx/C3t4efn5+6N+/P968eaM2161JkyYYOHAgUlJSMHHiRJw/fx4AYGhoqK8u0zeAw3lElKskJCSgZs2aiI+Px/Hjx1UZqNRLVeoH54gRI7B48WLcunULRYsWBQAuj5GHnT59GvXq1cPcuXNx/vx5XL9+HU2aNEGzZs1Qv359Vb3du3djxYoViIuLUz0OiCi78GpDRLmKiYkJpk2bBkNDQ0yePBkhISEA/gueFAoFAMDLywtSqRTR0dGqbRlA5V2enp4YMGAAHjx4gE2bNsHPzw8ymQwtWrRA//79sW7dOgCAj48PBg0aBDs7O9jZ2em515TXMRNFRLlK6lpQhw8fRp8+fdCwYUP4+/ujYsWKAP7LNq1atQpz5szBn3/+CUtLSz33mrJT6jmxYcMG/Pbbb9izZ49q/adatWrh6dOnMDAwQKFChdC+fXv07NkTMpkMJiYmeu455XX82kZEuYpEIoEQAt7e3li+fDmOHz+OWbNmITQ0FMCHbFNycjIuXLgANzc35M+fX889puyWmoXs3LkzIiMj8euvvwIAfH198fjxY2zduhUnTpyAlZUVNm/ejLi4OAZQlCOYiSKiXCf1risAOHToEPr27YsGDRrA398flSpVwvjx41UBFp+x+G1QKBSQyWTYtm0btm/fjtjYWPz777/Yt2+f2o0EcXFxsLCw0GNP6VvCIIqI9CatR/ykflg+fvwY8fHxKF++vGpo7/vvv0dKSgo2bNiAM2fOoGrVqnrqOeWk1HMC+HDnZrNmzfDmzRtcu3YNxYoVA8CbCkg/eMYRkV5ERERg5syZ6NGjB2JjYwF8+CCUyWR48OABypYtiy1btqiG9lasWIH169dj48aNOHv2LAOoPCit7/SpAdSTJ09w48YNlC9fHiNHjoSFhQVSUlJU9RhAkT7wrCOiHHf9+nV4e3vj3r17MDMzQ758+QB8+CB8/vw5atWqha5du2LChAmqOVKNGzfGkSNHEBISgipVquj5CEjXtAXVZcqUUT1kuGbNmrC0tMSpU6f02WUiDucRUc4KCwuDh4cHevTogSlTpqiGaVKH9o4fP45//vkHgwYNUg31pTXsR3nH9evX0bp1a3z33XcwNjZGUFCQamL48+fPUblyZfj4+GDRokWqjFOTJk3w4sULnDt3TvX4F6KcxiCKiHKMQqGAn58fYmNjsWrVKsjlcn13ifQss0F16vBeREQEhBCqxViJ9IFBFBHlqOrVq6Nx48aYOnWqxs9SPyBTJwlzsnDexqCavna8OhFRjhBC4M2bN3j8+DGsrKxUZR9LDaD8/Pzw8uVLBlB5nEwmw/nz5+Hg4JBmAJW6Or1SqVT7f6LcglcoIsoRQgiYmJjA1dUVwcHBiIiIUJvzlCoiIgKXL1/Gs2fP9NVVygEMqikv4BlJRDlCKpXCwMAAPj4+OHv2LFatWoWnT58CgNqk8fXr10Mul6NQoUL66irlAAbVlBfwlgYiyhbh4eHYtGkTLly4AGNjYxQvXhxjxozBTz/9hPDwcAQFBSE+Ph6dO3dGrVq1EBoaitWrV2Pt2rU4deoUV53O46RSKaRSKXx8fNC/f3+sWrUKAwYMQJEiRRhU01eDQRQR6VxoaCgaN26MGjVqwNzcHJGRkTh+/Dj27duHefPmYdq0aZDL5Zg3bx5WrFgBS0tLmJubw9jYGMePH1c9bJjyDgbVlBfx7jwi0qkHDx6gTp066NKlCyZNmgSZTIaUlBSEhoaiT58+iIyMxM6dO+Hm5oZTp04hKioKERERcHNzg7OzMzMOedDHQbWZmRkiIyNx+fJlWFlZYd68eWjatCnGjx+PefPmISEhQS2oXrNmDVxdXfV9CERpYhBFRDq1YMEC7N+/H8HBwZDL5WoLZYaHh6NFixYwMzPD2bNn9dxTygkMqikvYxBFRDrVo0cPPHr0CEePHtVYaVypVGLlypXw8/PjA4S/EQyqKS/j3XlEpFMSiUTtrruPv6dJpVJ4eXkhMTERL1680FcXKQddvXoVCoVCI4ACAEdHR/j5+eHq1au4cuWKHntJlDUMoohIJ1KDJU9PT0RGRmLx4sUAPgRSKSkpAD5komJiYuDi4gJHR0e99ZVyDoNqyssYRBFRliUkJGiUNWvWDMWKFcOsWbOwadMmAFA9IFYqlSI4OBj58+eHpaVljvaVchaDavoWMIgioix58uQJunbtihMnTgD478PRxsYGu3fvRlJSEkaOHIl+/frhxo0b2Lt3L4YPH46FCxdi2bJlDKLyIAbV9K3hxHIiypKIiAh07twZlpaWCAgIgIeHBwAgOTkZhoaGePjwIUaMGIGTJ08iOjoajo6OKFGiBObMmYNKlSrpufeka0+ePMHQoUPRr18/NGjQAACQkpICAwMD3Lt3D3Xr1oVEIkGzZs0wcOBAhIeH488//8SyZctw+vRpLmNAXyUGUUSUZWFhYRg8eDCEEAgMDFQFUklJSTAyMsL79++RmJiIw4cPo3nz5khOToaZmZmee03ZgUE1fYsYRBHRF0kvkFIoFEhJSUFgYCAiIiKwefNmGBoa6rm3lJ0YVNO3hkEUEX2xtD48k5KSMGzYMCxZsgSXLl1ClSpV9N1NygEMqulbwiCKiHTi4w/PkSNH4uDBg1iwYAHOnDnDAOobw6CavhUMoohIZ8LCwuDv748zZ87g7du3OHfuHFcl/0YxqKZvAYMoItKp27dv45dffsHUqVNRvnx5fXeH9IhBNeV1DKKISOdS78giYlBNeRmDKCIiylYMqimvYhBFRERElAV87AsRERFRFjCIIiIiIsoCBlFEREREWcAgioiIiCgLGEQRERERZQGDKCLKlG7dusHHx0f17/r162PIkCE53o+TJ09CIpHg5cuX6daRSCTYvXt3hvc5fvx4uLq6flG/7t+/D4lEgmvXrn3Rfogo92MQRZQHdOvWDRKJBBKJBEZGRihVqhQmTpyIlJSUbG97165dmDRpUobqZiTwISL6WhjouwNEpBtNmjTB6tWrkZiYiAMHDmDAgAEwNDREQECARt2kpCQYGRnppF1LS0ud7IeI6GvDTBRRHmFsbIxChQrB3t4e/fr1g5eXF/bu3QvgvyG4KVOmoEiRIihbtiwA4NGjR2jbti3Mzc1haWmJli1b4v79+6p9KhQK+Pv7w9zcHFZWVvjll1/w6fq8nw7nJSYmYsSIEShevDiMjY1RqlQprFq1Cvfv30eDBg0AABYWFpBIJOjWrRsAQKlUIigoCI6OjpDL5ahcuTJ27Nih1s6BAwdQpkwZyOVyNGjQQK2fGTVixAiUKVMG+fLlQ8mSJREYGIjk5GSNesuWLUPx4sWRL18+tG3bFvHx8Wo/X7lyJZydnWFiYoJy5cph8eLF6bYZFxeHTp06wcbGBnK5HKVLl8bq1asz3Xciyn2YiSLKo+RyOV68eKH697Fjx2BqaoqjR48C+PAoDm9vb7i7u+P06dMwMDDA5MmT0aRJE4SGhsLIyAizZs3CmjVr8Ntvv8HZ2RmzZs1CcHAwGjZsmG67Xbt2xblz5zB//nxUrlwZ9+7dQ0xMDIoXL46dO3fixx9/xO3bt2Fqagq5XA4ACAoKwoYNG7B06VKULl0af/75Jzp37gwbGxvUq1cPjx49QuvWrTFgwAD06dMHf//9N4YNG5bp96RgwYJYs2YNihQpguvXr6N3794oWLAgfvnlF1Wdu3fvYtu2bfj999/x6tUr9OzZE/3798fGjRsBABs3bsTYsWOxcOFCVKlSBVevXkXv3r2RP39++Pr6arQZGBiIf//9FwcPHoS1tTXu3r2L9+/fZ7rvRJQLCSL66vn6+oqWLVsKIYRQKpXi6NGjwtjYWAwfPlz1czs7O5GYmKjaZv369aJs2bJCqVSqyhITE4VcLheHDx8WQghRuHBhMX36dNXPk5OTRbFixVRtCSFEvXr1hJ+fnxBCiNu3bwsA4ujRo2n288SJEwKAiIuLU5UlJCSIfPnyibNnz6rV7dmzp+jQoYMQQoiAgADh4uKi9vMRI0Zo7OtTAERwcHC6P58xY4aoVq2a6t/jxo0TMplMPH78WFV28OBBIZVKxbNnz4QQQjg5OYlNmzap7WfSpEnC3d1dCCHEvXv3BABx9epVIYQQzZs3F927d0+3D0T09WImiiiP2LdvHwoUKIDk5GQolUp07NgR48ePV/28YsWKavOgQkJCcPfuXRQsWFBtPwkJCQgPD0d8fDyePXsGNzc31c8MDAxQvXp1jSG9VNeuXYNMJkO9evUy3O+7d+/i3bt3+O6779TKk5KSUKVKFQDAzZs31foBAO7u7hluI9XWrVsxf/58hIeH482bN0hJSYGpqalanRIlSqBo0aJq7SiVSty+fRsFCxZEeHg4evbsid69e6vqpKSkwMzMLM02+/Xrhx9//BFXrlxB48aN4ePjg9q1a2e670SU+zCIIsojGjRogCVLlsDIyAhFihSBgYH6n3f+/PnV/v3mzRtUq1ZNNUz1MRsbmyz1IXV4LjPevHkDANi/f79a8AJ8mOelK+fOnUOnTp0wYcIEeHt7w8zMDFu2bMGsWbMy3dcVK1ZoBHUymSzNbZo2bYoHDx7gwIEDOHr0KBo1aoQBAwZg5syZWT8YIsoVGEQR5RH58+dHqVKlMly/atWq2Lp1K2xtbTWyMakKFy6MCxcuoG7dugA+ZFwuX76MqlWrplm/YsWKUCqVOHXqFLy8vDR+npoJUygUqjIXFxcYGxvj4cOH6WawnJ2dVZPkU50/f177QX7k7NmzsLe3x+jRo1VlDx480Kj38OFDPH36FEWKFFG1I5VKUbZsWdjZ2aFIkSKIiIhAp06dMty2jY0NfH194evrC09PT/z8888MoojyAN6dR/SN6tSpE6ytrdGyZUucPn0a9+7dw8mTJzF48GA8fvwYAODn54dff/0Vu3fvxq1bt9C/f//PrvHk4OAAX19f9OjRA7t371btc9u2bQAAe3t7SCQS7Nu3D9HR0Xjz5g0KFiyI4cOHY+jQoVi7di3Cw8Nx5coVLFiwAGvXrgUA/PTTTwgLC8PPP/+M27dvY9OmTVizZk2mjrd06dJ4+PAhtmzZgvDwcMyfPx/BwcEa9UxMTODr64uQkBCcPn0agwcPRtu2bVGoUCEAwIQJExAUFIT58+fjzp07uH79OlavXo3Zs2en2e7YsWOxZ88e3L17Fzdu3MC+ffvg7Oycqb4TUe7EIIroG5UvXz78+eefKFGiBFq3bg1nZ2f07NkTCQkJqszUsGHD0KVLF/j6+sLd3R0FCxZEq1atPrvfJUuWoE2bNujfvz/KlSuH3r174+3btwCAokWLYsKECRg5ciTs7OwwcOBAAMCkSZMQGBiIoKAgODs7o0mTJti/fz8cHR0BfJintHPnTuzevRuVK1fG0qVLMXXq1Ewdb4sWLTB06FAMHDgQrq6uOHv2LAIDAzXqlSpVCq1bt8b333+Pxo0bo1KlSmpLGPTq1QsrV67E6tWrUbFiRdSrVw9r1qxR9fVTRkZGCAgIQKVKlVC3bl3IZDJs2bIlU30notxJItKbIUpERERE6WImioiIiCgLGEQRERERZQGDKCIiIqIsYBBFRERElAUMooiIiIiygEEUERERURYwiCIiIiLKAgZRRERERFnAIIqIiIgoCxhEEREREWUBgygiIiKiLGAQRURERJQF/wdtaRWlZALydAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    padded_eval_short_labels,\n",
    "    blstm_eval_short_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for BiLSTM on evaluation (short sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    padded_eval_unlabeled_labels,\n",
    "    blstm_eval_unlabeled_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for BiLSTM on evaluation (unlabeled sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    padded_eval_large_labels,\n",
    "    blstm_eval_large_predictions,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for BiLSTM on evaluation (large sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n",
      "Je: O\n",
      "pensais: O\n",
      "partir: O\n",
      "de: O\n",
      "Bangkok: LOC-DEP\n",
      ".: O\n",
      "Mais: O\n",
      "finalement: O\n",
      ",: O\n",
      "je: O\n",
      "vais: O\n",
      "devoir: O\n",
      "trouver: O\n",
      "un: O\n",
      "voyage: O\n",
      "de: O\n",
      "Tokyo: O\n",
      "vers: O\n",
      "Osaka: O\n",
      ".: O\n"
     ]
    }
   ],
   "source": [
    "test_sentence = \"Je pensais partir de Bangkok. Mais finalement, je vais devoir trouver un voyage de Tokyo vers Osaka.\"\n",
    "\n",
    "\n",
    "p = predict(test_sentence, bilstm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### BiLSTM with POS\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "bilstm_word_input = tf.keras.layers.Input(shape=(MAX_LEN,), name=\"bilstm_word_input\")\n",
    "bilstm_pos_input = tf.keras.layers.Input(shape=(MAX_LEN,), name=\"bilstm_pos_input\")\n",
    "\n",
    "emb_size = 64\n",
    "\n",
    "bilstm_word_embedding = tf.keras.layers.Embedding(\n",
    "    len(vocab), emb_size, name=\"bilstm_word_embedding\"\n",
    ")(bilstm_word_input)\n",
    "\n",
    "bilstm_pos_embedding = tf.keras.layers.Embedding(\n",
    "    len(unique_pos_tags),\n",
    "    emb_size,\n",
    "    name=\"bilstm_pos_embedding\",\n",
    ")(bilstm_pos_input)\n",
    "\n",
    "bilstm_concatenated = tf.keras.layers.Concatenate()(\n",
    "    [bilstm_word_embedding, bilstm_pos_embedding]\n",
    ")\n",
    "\n",
    "bilstm_masked_cat = tf.keras.layers.Masking(mask_value=0)(bilstm_concatenated)\n",
    "\n",
    "bilstm_layer_with_pos = tf.keras.layers.Bidirectional(\n",
    "    tf.keras.layers.LSTM(MAX_LEN, return_sequences=True), name=\"bilstm_layer\"\n",
    ")(bilstm_masked_cat)\n",
    "\n",
    "bilstm_dropout = tf.keras.layers.Dropout(0.2)(bilstm_layer_with_pos)\n",
    "\n",
    "bilstm_output = tf.keras.layers.Dense(len(unique_labels), activation=tf.nn.log_softmax)(\n",
    "    bilstm_dropout\n",
    ")\n",
    "\n",
    "bilstm_with_pos = tf.keras.Model(\n",
    "    inputs=[bilstm_word_input, bilstm_pos_input], outputs=bilstm_output\n",
    ")\n",
    "\n",
    "bilstm_with_pos.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(0.01),\n",
    "    loss=masked_loss,\n",
    "    metrics=[entity_accuracy],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m297/297\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m27s\u001b[0m 83ms/step - entity_accuracy: 0.7734 - loss: 0.1357 - val_entity_accuracy: 0.9912 - val_loss: 0.0055\n",
      "Epoch 2/10\n",
      "\u001b[1m297/297\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m29s\u001b[0m 97ms/step - entity_accuracy: 0.9930 - loss: 0.0052 - val_entity_accuracy: 0.9991 - val_loss: 0.0011\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x2d9ee9f70>"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bilstm_with_pos.fit(\n",
    "    pos_train_dataset.batch(32),\n",
    "    validation_data=pos_test_dataset.batch(32),\n",
    "    epochs=10,\n",
    "    callbacks=[\n",
    "        tf.keras.callbacks.EarlyStopping(\n",
    "            monitor=\"val_loss\", min_delta=0.01, restore_best_weights=True\n",
    "        )\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [00:24<00:00,  1.22it/s]\n"
     ]
    }
   ],
   "source": [
    "bilstm_with_pos_results = bootstrap_evaluation(\n",
    "    bilstm_with_pos, pos_sentences=bootstrap_eval_sentences_pos\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparison\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "blue"
         },
         "name": "BiLSTM",
         "type": "box",
         "y": [
          0.7837041007772715,
          0.7847896440129449,
          0.783288409703504,
          0.7917320774463632,
          0.777387070597782,
          0.7896145610278372,
          0.8009721847150959,
          0.771006711409396,
          0.782981000802783,
          0.7776008492569002,
          0.7727520435967302,
          0.7713310580204779,
          0.7712166987091459,
          0.7819879187259747,
          0.7635494155154091,
          0.7698154180238871,
          0.7941888619854721,
          0.7898471615720524,
          0.7717303005686433,
          0.7771550573792367,
          0.7904061052057781,
          0.7846735024284943,
          0.778675282714055,
          0.7859838274932615,
          0.7787222374554428,
          0.7543668122270742,
          0.7617787188988883,
          0.7648316408337787,
          0.7718264595607928,
          0.7906602254428341
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "red"
         },
         "name": "BiLSTM with POS",
         "type": "box",
         "y": [
          0.7505422993492408,
          0.7371727748691099,
          0.7534713125491224,
          0.7374358974358974,
          0.7519484009674818,
          0.7457803168008309,
          0.7453954496208017,
          0.7434864356701585,
          0.7594465141032464,
          0.7537154989384289,
          0.7647058823529411,
          0.7359580052493438,
          0.7411920529801325,
          0.7587661481676773,
          0.7640449438202247,
          0.7588614393125671,
          0.7492749802267334,
          0.7354427014091997,
          0.7323498419388831,
          0.7607947019867549,
          0.7554734898443682,
          0.7655579399141631,
          0.7630606860158311,
          0.7455811462238886,
          0.7641160949868074,
          0.7471895424836601,
          0.7432072456046883,
          0.7355507088331515,
          0.7689778021375719,
          0.747867803837953
         ]
        }
       ],
       "layout": {
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "title": {
         "text": "F1-scores for BiLSTM vs BiLSTM with POS"
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import plotly.express as px\n",
    "import plotly.graph_objects as go\n",
    "import pandas as pd\n",
    "\n",
    "df_bilstm_results = pd.DataFrame(bilstm_results)\n",
    "df_bilstm_with_pos_results = pd.DataFrame(bilstm_with_pos_results)\n",
    "\n",
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=df_bilstm_results[\"f1_scores\"],\n",
    "        name=\"BiLSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"blue\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=df_bilstm_with_pos_results[\"f1_scores\"],\n",
    "        name=\"BiLSTM with POS\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"red\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.update_layout(title=\"F1-scores for BiLSTM vs BiLSTM with POS\")\n",
    "\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## CamemBERT\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model used will be the fine-tuned version of [CamemBERT](https://huggingface.co/almanach/camembert-base) which is a state-of-the-art language model for French text.\n",
    "\n",
    "The model has been fine-tuned for our task [in here](./camemBERT_finetuning.ipynb) and saved on Hugging Face for convenience.\n",
    "\n",
    "[Checkout the model on Huggingface Hub](https://huggingface.co/Az-r-ow/CamemBERT-NER-Travel)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Transformers is only compatible with Keras 2, but you have explicitly set `TF_USE_LEGACY_KERAS` to `0`. This may result in unexpected behaviour or errors if Keras 3 objects are passed to Transformers models.\n",
      "Some layers from the model checkpoint at Az-r-ow/CamemBERT-NER-Travel were not used when initializing TFCamembertForTokenClassification: ['dropout_39']\n",
      "- This IS expected if you are initializing TFCamembertForTokenClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing TFCamembertForTokenClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "All the layers of TFCamembertForTokenClassification were initialized from the model checkpoint at Az-r-ow/CamemBERT-NER-Travel.\n",
      "If your task is similar to the task the model of the checkpoint was trained on, you can already use TFCamembertForTokenClassification for predictions without further training.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from transformers import TFCamembertForTokenClassification, CamembertTokenizerFast\n",
    "import numpy as np\n",
    "\n",
    "tokenizer = CamembertTokenizerFast.from_pretrained(\"cmarkea/distilcamembert-base\")\n",
    "\n",
    "camembert = TFCamembertForTokenClassification.from_pretrained(\n",
    "    \"Az-r-ow/CamemBERT-NER-Travel\", num_labels=len(unique_labels)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenized_eval_short_sentences = tokenizer(\n",
    "    eval_sentences,\n",
    "    return_tensors=\"tf\",\n",
    "    padding=\"max_length\",\n",
    "    max_length=150,\n",
    ")\n",
    "\n",
    "tokenized_eval_unlabeled_sentences = tokenizer(\n",
    "    eval_unlabeled, return_tensors=\"tf\", padding=\"max_length\", max_length=150\n",
    ")\n",
    "\n",
    "tokenized_eval_large_sentences = tokenizer(\n",
    "    eval_large,\n",
    "    return_tensors=\"tf\",\n",
    "    truncation=True,\n",
    "    padding=\"max_length\",\n",
    "    max_length=150,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 34s 1s/step\n"
     ]
    }
   ],
   "source": [
    "camembert_short_sentences_predictions = camembert.predict(\n",
    "    tokenized_eval_short_sentences\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(150,), dtype=int64, numpy=\n",
       "array([0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 2, 2, 2, 0, 0,\n",
       "       0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])>"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.argmax(camembert_short_sentences_predictions.logits, axis=-1)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 1s 631ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 100), dtype=int64, numpy=\n",
       "array([[0, 0, 0, 0, 0, 1, 0, 2, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])>"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_sentence = \"Je veux partir de Paris à Montpellier.\"\n",
    "\n",
    "tokenized_test_sentence = tokenizer(\n",
    "    test_sentence, return_tensors=\"tf\", padding=\"max_length\", max_length=MAX_LEN\n",
    ")\n",
    "\n",
    "camembert_test_sentence_predictions = camembert.predict(tokenized_test_sentence)\n",
    "\n",
    "tf.math.argmax(camembert_test_sentence_predictions.logits, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "def align_labels_with_tokens(encodings, labels):\n",
    "    \"\"\"\n",
    "    Aligns the labels to match the tokenized outputs.\n",
    "\n",
    "    Args:\n",
    "        encodings (BatchEncoding): Tokenized outputs from the Hugging Face tokenizer (must use a fast tokenizer).\n",
    "        labels (List[List[int]]): Original labels for each sentence before tokenization. Each inner list corresponds to one sentence.\n",
    "\n",
    "    Returns:\n",
    "        List[List[int]]: Aligned labels, where each inner list corresponds to the aligned labels for the tokenized sentence.\n",
    "                         Special tokens and padding are assigned a value of -100.\n",
    "    \"\"\"\n",
    "    adapted_labels = []\n",
    "\n",
    "    for i, label in enumerate(labels):\n",
    "        word_ids = encodings.word_ids(\n",
    "            batch_index=i\n",
    "        )  # Get word IDs for the i-th sentence\n",
    "        aligned_labels = []\n",
    "        previous_word_id = None\n",
    "\n",
    "        for word_id in word_ids:\n",
    "            if word_id is None:\n",
    "                # Special tokens (e.g., [CLS], [SEP], or padding)\n",
    "                aligned_labels.append(-100)\n",
    "            elif word_id != previous_word_id:\n",
    "                # New word\n",
    "                aligned_labels.append(label[word_id])\n",
    "            else:\n",
    "                # Subword token (same word)\n",
    "                aligned_labels.append(\n",
    "                    label[word_id]\n",
    "                )  # Or append -100 to ignore subwords\n",
    "            previous_word_id = word_id\n",
    "\n",
    "        adapted_labels.append(aligned_labels)\n",
    "\n",
    "    return adapted_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "aligned_eval_short_labels = align_labels_with_tokens(\n",
    "    tokenized_eval_short_sentences, padded_eval_short_labels\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "aligned_eval_long_labels = align_labels_with_tokens(\n",
    "    tokenized_eval_large_sentences, padded_eval_large_labels\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "aligned_eval_unlabeled_labels = align_labels_with_tokens(\n",
    "    tokenized_eval_unlabeled_sentences, padded_eval_unlabeled_labels\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(758, 150)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    aligned_eval_short_labels,\n",
    "    camembert_short_sentences_predictions.logits,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for CamemBERT on evaluation (short_sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 4s 1s/step\n"
     ]
    }
   ],
   "source": [
    "camembert_large_sentences_predictions = camembert.predict(\n",
    "    tokenized_eval_large_sentences\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96, 150)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    aligned_eval_long_labels,\n",
    "    camembert_large_sentences_predictions.logits,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for CamemBERT on evaluation (long sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 5s 1s/step\n"
     ]
    }
   ],
   "source": [
    "camembert_unlabeled_sentence_predictions = camembert.predict(\n",
    "    tokenized_eval_unlabeled_sentences\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 150)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(\n",
    "    aligned_eval_unlabeled_labels,\n",
    "    camembert_unlabeled_sentence_predictions.logits,\n",
    "    unique_labels,\n",
    "    title=\"Confusion matrix for CamemBERT on evaluation (unlabeled sentences) dataset\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "camembert_eval_sentences = eval_sentences + eval_unlabeled + eval_large\n",
    "camembert_eval_labels = np.concatenate(\n",
    "    [padded_eval_short_labels, padded_eval_unlabeled_labels, padded_eval_large_labels]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "camembert_eval_sentences_tokenized = tokenizer(\n",
    "    camembert_eval_sentences, return_tensors=\"tf\", padding=\"max_length\", max_length=150\n",
    ")\n",
    "\n",
    "camembert_eval_labels_aligned = align_labels_with_tokens(\n",
    "    camembert_eval_sentences_tokenized, camembert_eval_labels\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "954"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(camembert_eval_labels_aligned)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [18:42<00:00, 37.42s/it]\n"
     ]
    }
   ],
   "source": [
    "camembert_results = bootstrap_evaluation(\n",
    "    camembert,\n",
    "    sentences=camembert_eval_sentences_tokenized,\n",
    "    labels=camembert_eval_labels_aligned,\n",
    "    from_logits=True,\n",
    "    has_mask=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "with open(\"./results/camembert_results.pickle\", \"wb\") as f:\n",
    "    pickle.dump(camembert_results, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction Analysis\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "def mask_predictions(predictions: np.ndarray, labels: np.ndarray, mask=-1):\n",
    "    \"\"\"\n",
    "    Mask the predictions based on the labels.\n",
    "\n",
    "    Parameters:\n",
    "    predictions (tensor): Predicted logits.\n",
    "    labels (tensor): True labels.\n",
    "    mask (int): Mask value.\n",
    "\n",
    "    Returns:\n",
    "    masked_predictions (tensor): Masked predictions.\n",
    "    \"\"\"\n",
    "\n",
    "    # Perform argmax to get the predicted values\n",
    "    predictions_class = np.argmax(predictions, axis=-1)\n",
    "\n",
    "    # Create the mask based on the mask in the labels\n",
    "    mask = np.where(labels == mask, mask, 0)\n",
    "    return predictions_class + mask\n",
    "\n",
    "\n",
    "def get_mistakes_positions(predictions, labels):\n",
    "    masked_predictions = mask_predictions(predictions, labels)\n",
    "    wrong_predictions = np.where(masked_predictions != labels, masked_predictions, -2)\n",
    "    return wrong_predictions\n",
    "\n",
    "\n",
    "def plot_mistakes_positions(\n",
    "    predictions, labels, title=\"Mistakes density in sentences\", max_len=None\n",
    "):\n",
    "    mistakes_positions = get_mistakes_positions(predictions, labels)\n",
    "    mistakes_positions = np.where(mistakes_positions >= 0, 1, 0)\n",
    "    mistakes_positions = np.sum(mistakes_positions, axis=0)\n",
    "    mistakes_position_density = mistakes_positions / np.max(mistakes_positions)\n",
    "\n",
    "    plt.figure(figsize=(20, 5))\n",
    "\n",
    "    if max_len:\n",
    "        mistakes_position_density = mistakes_position_density[:max_len]\n",
    "\n",
    "    sns.heatmap(mistakes_position_density.reshape(1, -1), cmap=\"Blues\")\n",
    "\n",
    "    plt.title(title)\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "def label_pos_freq_in_sentences(\n",
    "    labels,\n",
    "    unique_labels,\n",
    "    title=\"Sentences labels heatmap\",\n",
    "    skip_zero_positions=False,\n",
    "    max_len=None,\n",
    "):\n",
    "    \"\"\"\n",
    "    Generate a heatmap of the sentences and their labels.\n",
    "\n",
    "    Parameters:\n",
    "    labels (list): List of labels in a sentence (it could be predictions)\n",
    "    unique_labels (list): List of unique labels.\n",
    "    title (str): Title of the plot.\n",
    "    \"\"\"\n",
    "    positions_label_counts = {}\n",
    "\n",
    "    for label in tqdm(labels):\n",
    "        length = max_len if max_len else len(label)\n",
    "        for i in range(length):\n",
    "            # skip padding\n",
    "            if label[i] not in unique_labels.values() and not label[i] == -2:\n",
    "                continue\n",
    "            if i not in positions_label_counts:\n",
    "                positions_label_counts[i] = {}\n",
    "                for l in unique_labels.values():\n",
    "                    positions_label_counts[i][l] = 0\n",
    "            if label[i] == -2:\n",
    "                continue\n",
    "            positions_label_counts[i][label[i]] += 1\n",
    "\n",
    "    # Filter out positions where the sum of label counts is zero\n",
    "    hm_data = {\n",
    "        k: v\n",
    "        for k, v in positions_label_counts.items()\n",
    "        if skip_zero_positions and sum(list(v.values())) > 0 or not skip_zero_positions\n",
    "    }\n",
    "    hm_data = pd.DataFrame(hm_data)\n",
    "\n",
    "    # dtype conversion\n",
    "    hm_data = hm_data.astype(float)\n",
    "    hm_data.set_index(pd.Index(unique_labels.keys()), inplace=True)\n",
    "\n",
    "    # normalizing the data per column\n",
    "    hm_data = hm_data.div(hm_data.sum(axis=0), axis=1)\n",
    "\n",
    "    plt.figure(figsize=(20, 10))\n",
    "\n",
    "    sns.heatmap(hm_data, cmap=\"Blues\")\n",
    "\n",
    "    plt.title(title)\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 757/757 [00:00<00:00, 3743.50it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    padded_eval_short_labels, unique_labels, title=\"Short sentences labels\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 757/757 [00:00<00:00, 36883.60it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    mask_predictions(blstm_eval_short_predictions, padded_eval_short_labels),\n",
    "    unique_labels,\n",
    "    title=\"Short sentences predictions\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 757/757 [00:00<00:00, 117731.03it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    get_mistakes_positions(blstm_eval_short_predictions, padded_eval_short_labels),\n",
    "    unique_labels,\n",
    "    title=\"Short sentences mistakes\",\n",
    "    max_len=25,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 95/95 [00:00<00:00, 3243.09it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    padded_eval_large_labels, unique_labels, title=\"Large sentences labels\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 95/95 [00:00<00:00, 29150.55it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    mask_predictions(blstm_eval_large_predictions, padded_eval_large_labels),\n",
    "    unique_labels,\n",
    "    title=\"Large sentences predictions\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_mistakes_positions(blstm_eval_large_predictions, padded_eval_large_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_mistakes_positions(lstm_with_pos_eval_large_predictions, padded_eval_large_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_mistakes_positions(\n",
    "    lstm_with_pos_eval_short_predictions, padded_eval_short_labels, max_len=25\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_mistakes_positions(\n",
    "    blstm_eval_short_predictions, padded_eval_short_labels, max_len=25\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 95/95 [00:00<00:00, 27883.76it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_pos_freq_in_sentences(\n",
    "    get_mistakes_positions(blstm_eval_large_predictions, padded_eval_large_labels),\n",
    "    unique_labels,\n",
    "    title=\"Large sentences mistakes\",\n",
    "    skip_zero_positions=False,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Overall Results\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "hmm_results = pd.read_pickle(\"./results/hmm_f1_results\")\n",
    "lstm_results = pd.read_pickle(\"./results/lstm_results.pickle\")\n",
    "lstm_with_pos_results = pd.read_pickle(\"./results/lstm_with_pos_results.pickle\")\n",
    "bilstm_results = pd.read_pickle(\"./results/bilstm_results.pickle\")\n",
    "camembert_results = pd.read_pickle(\"./results/camembert_results.pickle\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "hmm_results_df = pd.DataFrame({\"f1_scores\": hmm_results})\n",
    "lstm_results_df = pd.DataFrame(lstm_results)\n",
    "lstm_with_pos_results_df = pd.DataFrame(lstm_with_pos_results)\n",
    "bilstm_results_df = pd.DataFrame(bilstm_results)\n",
    "camembert_results_df = pd.DataFrame(camembert_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "blue"
         },
         "name": "HMM",
         "type": "box",
         "y": [
          0.1787280701754386,
          0.17358490566037735,
          0.21916357861302277,
          0.15047021943573669,
          0.2246104245029554,
          0.15633423180592992,
          0.2132390096008085,
          0.2222222222222222,
          0.21583459404942007,
          0.2234214786832164,
          0.22330097087378642,
          0.20961145194274028,
          0.15492200107584722,
          0.26423200859291085,
          0.24883720930232558,
          0.21973333333333334,
          0.15724286483575659,
          0.22066738428417654,
          0.1799163179916318,
          0.23738532110091742,
          0.17857142857142858,
          0.1553398058252427,
          0.23135964912280702,
          0.22306034482758622,
          0.15951539626451286,
          0.19616685456595265,
          0.22860298177802318,
          0.22660098522167488,
          0.18969298245614036,
          0.16020671834625322
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "red"
         },
         "name": "LSTM",
         "type": "box",
         "y": [
          0.7618257261410788,
          0.7644217107132708,
          0.7775585001409642,
          0.774390243902439,
          0.7608447488584474,
          0.7607361963190185,
          0.7933425797503467,
          0.7749303621169916,
          0.7679465776293823,
          0.7556066705002875,
          0.7724331926863572,
          0.7743201569946734,
          0.8006952491309386,
          0.7796135536264351,
          0.7630090497737556,
          0.7779024116657319,
          0.7739227756015669,
          0.7729957805907173,
          0.7790178571428571,
          0.7672242699177771,
          0.7670149675232985,
          0.780650542118432,
          0.7711670480549199,
          0.7820829898323716,
          0.7791563275434243,
          0.7692307692307693,
          0.7648757943385326,
          0.7610810810810811,
          0.7864849517319704,
          0.7941501103752759
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "green"
         },
         "name": "LSTM with POS",
         "type": "box",
         "y": [
          0.8164736164736165,
          0.7985812009120852,
          0.784516765285996,
          0.8030188679245283,
          0.8226950354609929,
          0.8065573770491803,
          0.8091872791519434,
          0.8013398608606029,
          0.7972406996797241,
          0.801345059493016,
          0.8103008669046405,
          0.820815175596001,
          0.802,
          0.8129829984544049,
          0.8045859872611465,
          0.8225765628954695,
          0.8179299924452279,
          0.8139656070870245,
          0.8120261437908497,
          0.8071500503524672,
          0.8032786885245902,
          0.8046324269889225,
          0.8132375577219086,
          0.8171772978402813,
          0.8084672277480235,
          0.8151873271310033,
          0.8107396683337721,
          0.8050568900126422,
          0.8066098631551769,
          0.8025316455696202
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "orange"
         },
         "name": "BiLSTM",
         "type": "box",
         "y": [
          0.8619989852866565,
          0.8681347837622712,
          0.853715775749674,
          0.8695187165775401,
          0.8717029449423815,
          0.8714934544483035,
          0.8753206772703951,
          0.8563927540036755,
          0.8748035620743845,
          0.8718487394957983,
          0.8780860269788751,
          0.8719262295081968,
          0.8841447539703202,
          0.8628571428571429,
          0.8853388301984024,
          0.8711751092825919,
          0.8743589743589744,
          0.8676699532952776,
          0.8754977435625166,
          0.8647551382897741,
          0.8802456499488229,
          0.873336881319851,
          0.8622816032887975,
          0.8749016522423289,
          0.8817584638706417,
          0.8745762711864407,
          0.8757606490872211,
          0.88,
          0.8573635427394438,
          0.8709338929695698
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "purple"
         },
         "name": "CamemBERT",
         "type": "box",
         "y": [
          0.9662855179859645,
          0.9660926509822945,
          0.963169164882227,
          0.9690846537771963,
          0.9700561145510835,
          0.9691146190803019,
          0.9700717889018238,
          0.9696671262556628,
          0.9682887622387841,
          0.9667264878498326,
          0.968851762707065,
          0.9701566282919116,
          0.9683338982230184,
          0.9628297362110312,
          0.969846833437365,
          0.9671995723796103,
          0.9672492219296146,
          0.9675979262672811,
          0.9689582936583508,
          0.9660683320054365,
          0.9632474477394264,
          0.970085259474644,
          0.9646859903381643,
          0.9734946842388907,
          0.9673554520760574,
          0.9648621625482074,
          0.974221214511041,
          0.9705854844743733,
          0.9684483856743913,
          0.9642486557871899
         ]
        }
       ],
       "layout": {
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "yaxis": {
         "title": {
          "text": "F1-scores"
         }
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import plotly.graph_objects as go\n",
    "\n",
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=hmm_results_df[\"f1_scores\"],\n",
    "        name=\"HMM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"blue\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=lstm_results_df[\"f1_scores\"],\n",
    "        name=\"LSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"red\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=lstm_with_pos_results_df[\"f1_scores\"],\n",
    "        name=\"LSTM with POS\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"green\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=bilstm_results_df[\"f1_scores\"],\n",
    "        name=\"BiLSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"orange\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=camembert_results_df[\"f1_scores\"],\n",
    "        name=\"CamemBERT\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"purple\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "\n",
    "fig.update_layout(yaxis_title=\"F1-scores\")\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "blue"
         },
         "name": "LSTM",
         "type": "box",
         "y": [
          0.7389743328094482,
          0.7424870729446411,
          0.7421249151229858,
          0.7397189140319824,
          0.7368146181106567,
          0.7379094958305359,
          0.7438764572143555,
          0.7052023410797119,
          0.7781836986541748,
          0.7354166507720947,
          0.7414684295654297,
          0.7362755537033081,
          0.7496054768562317,
          0.7684938907623291,
          0.729968786239624,
          0.7328205108642578,
          0.7568585276603699,
          0.7441006898880005,
          0.7421875,
          0.7382833361625671,
          0.752598762512207,
          0.7518286108970642,
          0.7472867965698242,
          0.7437797784805298,
          0.7225103974342346,
          0.7497420310974121,
          0.7191069722175598,
          0.7411764860153198,
          0.7272727489471436,
          0.7470129728317261
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "red"
         },
         "name": "BiLSTM",
         "type": "box",
         "y": [
          0.8762248754501343,
          0.8772118091583252,
          0.8661375641822815,
          0.883695662021637,
          0.8873826861381531,
          0.8759398460388184,
          0.8899322152137756,
          0.8661710023880005,
          0.8887706398963928,
          0.890080451965332,
          0.8878023624420166,
          0.8869202733039856,
          0.8927444815635681,
          0.8793011903762817,
          0.8975966572761536,
          0.8786306977272034,
          0.8880208134651184,
          0.8827877640724182,
          0.8932827711105347,
          0.8824443221092224,
          0.8986415863037109,
          0.8784796595573425,
          0.8780742883682251,
          0.883474588394165,
          0.8957905769348145,
          0.8901273608207703,
          0.895282506942749,
          0.8931377530097961,
          0.869905948638916,
          0.8778424263000488
         ]
        },
        {
         "boxmean": true,
         "boxpoints": "all",
         "marker": {
          "color": "green"
         },
         "name": "CamemBERT",
         "type": "box",
         "y": [
          0.9495962262153625,
          0.9493274092674255,
          0.9514999389648438,
          0.9535566568374634,
          0.9554597735404968,
          0.9603195190429688,
          0.955078661441803,
          0.9598649144172668,
          0.9541693925857544,
          0.9559488892555237,
          0.958782970905304,
          0.9597527384757996,
          0.9568421244621277,
          0.949177086353302,
          0.9561835527420044,
          0.9508851170539856,
          0.9558723568916321,
          0.9551921486854553,
          0.9567764401435852,
          0.9546648859977722,
          0.9534945487976074,
          0.9541284441947937,
          0.9559615850448608,
          0.9591890573501587,
          0.9482483863830566,
          0.9509974718093872,
          0.9585822820663452,
          0.9563137292861938,
          0.952602207660675,
          0.9516522884368896
         ]
        }
       ],
       "layout": {
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "title": {
         "text": "Entity accuracy distribution for LSTM and BiLSTM models"
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=lstm_results_df[\"entity_accuracies\"],\n",
    "        name=\"LSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"blue\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=bilstm_results_df[\"entity_accuracies\"],\n",
    "        name=\"BiLSTM\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"red\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Box(\n",
    "        y=camembert_results_df[\"entity_accuracies\"],\n",
    "        name=\"CamemBERT\",\n",
    "        boxmean=True,\n",
    "        boxpoints=\"all\",\n",
    "        marker=dict(color=\"green\"),\n",
    "    )\n",
    ")\n",
    "\n",
    "fig.update_layout(title=\"Entity accuracy distribution for LSTM and BiLSTM models\")\n",
    "\n",
    "fig.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}