03.2_LearningNLP_Tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "artificial-passing",
   "metadata": {
    "papermill": {
     "duration": 0.031712,
     "end_time": "2021-05-20T16:18:33.025672",
     "exception": false,
     "start_time": "2021-05-20T16:18:32.993960",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# CommonLit Readability Challenge - Word2Vec + UMAP 🗺️\n",
    "## Introduction \n",
    "\n",
    "In this notebook we are going to see how to apply a pretrained [Word2Vec](https://jalammar.github.io/illustrated-word2vec/) model (from gensim) to a text corpus of our choice and,after that, we are going to see if our embedding has captured readability by employing a dimensionality reduction algorithm named [UMAP](https://umap-learn.readthedocs.io/en/latest/how_umap_works.html).\n",
    "\n",
    "``` If you're trying to locally install umap, pay attention to the fact that the correct command is \"pip install umap-learn\", indeed it is \"import umap\" while importing it```.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "mighty-decimal",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:33.088627Z",
     "iopub.status.busy": "2021-05-20T16:18:33.087667Z",
     "iopub.status.idle": "2021-05-20T16:18:54.890454Z",
     "shell.execute_reply": "2021-05-20T16:18:54.890924Z",
     "shell.execute_reply.started": "2021-05-20T15:42:16.237988Z"
    },
    "papermill": {
     "duration": 21.835899,
     "end_time": "2021-05-20T16:18:54.891224",
     "exception": false,
     "start_time": "2021-05-20T16:18:33.055325",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np\n",
    "import re \n",
    "\n",
    "import nltk\n",
    "#nltk.download('punkt')\n",
    "\n",
    "# model imports\n",
    "from gensim.models.word2vec import Word2Vec\n",
    "import gensim.downloader as api # we need an internet connection for this one \n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# visualization imports\n",
    "import umap # nonlinear dimensionality reduction\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ready-macro",
   "metadata": {
    "papermill": {
     "duration": 0.029973,
     "end_time": "2021-05-20T16:18:54.951796",
     "exception": false,
     "start_time": "2021-05-20T16:18:54.921823",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Utilities\n",
    "Here some utility functions: \n",
    "* ```remove_ascii_words``` it might be useless in this case but it's quite useful for general purpose preprocessing, where we don't know whether there are non-ascii characters around\n",
    "\n",
    "* ```get_good_tokens``` removes useless punctuation\n",
    "\n",
    "* ``` w2v_preprocessing``` all the necessary preprocessing for gensim Word2Vec model, we basically divide each document into individual sentences, remove punctuation by using ```get_good_tokens```, tokenize every sentence and remove empty lists. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "unauthorized-recall",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:55.016792Z",
     "iopub.status.busy": "2021-05-20T16:18:55.016243Z",
     "iopub.status.idle": "2021-05-20T16:18:55.021510Z",
     "shell.execute_reply": "2021-05-20T16:18:55.021927Z",
     "shell.execute_reply.started": "2021-05-20T15:42:42.386351Z"
    },
    "papermill": {
     "duration": 0.039086,
     "end_time": "2021-05-20T16:18:55.022118",
     "exception": false,
     "start_time": "2021-05-20T16:18:54.983032",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "our_special_word = 'qwerty'\n",
    "\n",
    "def remove_ascii_words(df):\n",
    "    \"\"\" removes non-ascii characters from the 'texts' column in df.\n",
    "    It returns the words containig non-ascii characers.\n",
    "    \"\"\"\n",
    "    non_ascii_words = []\n",
    "    for i in range(len(df)):\n",
    "        for word in df.loc[i, 'excerpt'].split(' '):\n",
    "            if any([ord(character) >= 128 for character in word]):\n",
    "                non_ascii_words.append(word)\n",
    "                df.loc[i, 'excerpt'] = df.loc[i, 'excerpt'].replace(word, our_special_word)\n",
    "    return non_ascii_words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "moderate-devon",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:55.084502Z",
     "iopub.status.busy": "2021-05-20T16:18:55.083942Z",
     "iopub.status.idle": "2021-05-20T16:18:55.088810Z",
     "shell.execute_reply": "2021-05-20T16:18:55.088384Z",
     "shell.execute_reply.started": "2021-05-20T15:42:42.396641Z"
    },
    "papermill": {
     "duration": 0.03665,
     "end_time": "2021-05-20T16:18:55.088957",
     "exception": false,
     "start_time": "2021-05-20T16:18:55.052307",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def get_good_tokens(sentence):\n",
    "    replaced_punctation = list(map(lambda token: re.sub('[^0-9A-Za-z!?]+', '', token), sentence))\n",
    "    removed_punctation = list(filter(lambda token: token, replaced_punctation))\n",
    "    return removed_punctation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "immune-vegetable",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:55.155436Z",
     "iopub.status.busy": "2021-05-20T16:18:55.154716Z",
     "iopub.status.idle": "2021-05-20T16:18:55.157772Z",
     "shell.execute_reply": "2021-05-20T16:18:55.157332Z",
     "shell.execute_reply.started": "2021-05-20T15:42:42.41343Z"
    },
    "papermill": {
     "duration": 0.039231,
     "end_time": "2021-05-20T16:18:55.157917",
     "exception": false,
     "start_time": "2021-05-20T16:18:55.118686",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def w2v_preprocessing(df):\n",
    "    \"\"\" All the preprocessing steps for word2vec are done in this function.\n",
    "    All mutations are done on the dataframe itself. So this function returns\n",
    "    nothing.\n",
    "    \"\"\"\n",
    "    df['excerpt'] = df.excerpt.str.lower()\n",
    "    df['document_sentences'] = df.excerpt.str.split('.')  # split texts into individual sentences\n",
    "    df['tokenized_sentences'] = list(map(lambda sentences:\n",
    "                                         list(map(nltk.word_tokenize, sentences)),\n",
    "                                         df.document_sentences))  # tokenize sentences\n",
    "   # df['tokenized_sentences'] = list(map(lambda sentences: list(map( ,sentences)), df.tokenized_sentences))\n",
    "    df['tokenized_sentences'] = list(map(lambda sentences:\n",
    "                                         list(map(get_good_tokens, sentences)),\n",
    "                                         df.tokenized_sentences))  # remove unwanted characters\n",
    "    df['tokenized_sentences'] = list(map(lambda sentences:\n",
    "                                         list(filter(lambda lst: lst, sentences)),\n",
    "                                         df.tokenized_sentences))  # remove empty lists"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "plain-colorado",
   "metadata": {
    "papermill": {
     "duration": 0.029648,
     "end_time": "2021-05-20T16:18:55.217256",
     "exception": false,
     "start_time": "2021-05-20T16:18:55.187608",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Data Import\n",
    "Let's load training data as a ```pandas``` dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "identical-permission",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:55.282336Z",
     "iopub.status.busy": "2021-05-20T16:18:55.281700Z",
     "iopub.status.idle": "2021-05-20T16:18:55.461707Z",
     "shell.execute_reply": "2021-05-20T16:18:55.462198Z",
     "shell.execute_reply.started": "2021-05-20T15:42:42.428878Z"
    },
    "papermill": {
     "duration": 0.21498,
     "end_time": "2021-05-20T16:18:55.462367",
     "exception": false,
     "start_time": "2021-05-20T16:18:55.247387",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>url_legal</th>\n",
       "      <th>license</th>\n",
       "      <th>excerpt</th>\n",
       "      <th>target</th>\n",
       "      <th>standard_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>c12129c31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>When the young people returned to the ballroom...</td>\n",
       "      <td>-0.340259</td>\n",
       "      <td>0.464009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>85aa80a4c</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>All through dinner time, Mrs. Fayre was somewh...</td>\n",
       "      <td>-0.315372</td>\n",
       "      <td>0.480805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>b69ac6792</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>As Roger had predicted, the snow departed as q...</td>\n",
       "      <td>-0.580118</td>\n",
       "      <td>0.476676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>dd1000b26</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>And outside before the palace a great garden w...</td>\n",
       "      <td>-1.054013</td>\n",
       "      <td>0.450007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>37c1b32fb</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Once upon a time there were Three Bears who li...</td>\n",
       "      <td>0.247197</td>\n",
       "      <td>0.510845</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          id url_legal license  \\\n",
       "0  c12129c31       NaN     NaN   \n",
       "1  85aa80a4c       NaN     NaN   \n",
       "2  b69ac6792       NaN     NaN   \n",
       "3  dd1000b26       NaN     NaN   \n",
       "4  37c1b32fb       NaN     NaN   \n",
       "\n",
       "                                             excerpt    target  standard_error  \n",
       "0  When the young people returned to the ballroom... -0.340259        0.464009  \n",
       "1  All through dinner time, Mrs. Fayre was somewh... -0.315372        0.480805  \n",
       "2  As Roger had predicted, the snow departed as q... -0.580118        0.476676  \n",
       "3  And outside before the palace a great garden w... -1.054013        0.450007  \n",
       "4  Once upon a time there were Three Bears who li...  0.247197        0.510845  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data =  pd.read_csv(\"../input/commonlitreadabilityprize/train.csv\")\n",
    "train_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "rough-champagne",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:55.528570Z",
     "iopub.status.busy": "2021-05-20T16:18:55.527874Z",
     "iopub.status.idle": "2021-05-20T16:18:56.391393Z",
     "shell.execute_reply": "2021-05-20T16:18:56.391782Z",
     "shell.execute_reply.started": "2021-05-20T15:42:42.61489Z"
    },
    "papermill": {
     "duration": 0.89892,
     "end_time": "2021-05-20T16:18:56.391958",
     "exception": false,
     "start_time": "2021-05-20T16:18:55.493038",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Replaced 1327 words with characters with an ordinal >= 128 in the test data.\n"
     ]
    }
   ],
   "source": [
    "train_data.excerpt = train_data['excerpt'].apply(str)\n",
    "non_ascii_words = remove_ascii_words(train_data)\n",
    "\n",
    "print(\"Replaced {} words with characters with an ordinal >= 128 in the test data.\".format(\n",
    "    len(non_ascii_words)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "industrial-vulnerability",
   "metadata": {
    "papermill": {
     "duration": 0.030954,
     "end_time": "2021-05-20T16:18:56.453537",
     "exception": false,
     "start_time": "2021-05-20T16:18:56.422583",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Pretrained Word2Vec\n",
    "Here we can load a pretrained version of Word2Vec (on Google News) from ```Gensim```'s API. Each vector has 300 components"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "sacred-wesley",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:18:56.523999Z",
     "iopub.status.busy": "2021-05-20T16:18:56.523439Z",
     "iopub.status.idle": "2021-05-20T16:23:27.979021Z",
     "shell.execute_reply": "2021-05-20T16:23:27.979662Z"
    },
    "papermill": {
     "duration": 271.495786,
     "end_time": "2021-05-20T16:23:27.980140",
     "exception": false,
     "start_time": "2021-05-20T16:18:56.484354",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[==================================================] 100.0% 1662.8/1662.8MB downloaded\n"
     ]
    }
   ],
   "source": [
    "W2Vmodel = api.load('word2vec-google-news-300')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "animated-treatment",
   "metadata": {
    "papermill": {
     "duration": 34.384335,
     "end_time": "2021-05-20T16:24:36.889466",
     "exception": false,
     "start_time": "2021-05-20T16:24:02.505131",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Preprocessing\n",
    "\n",
    "Here we apply our utility functions for preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "frequent-practice",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:25:45.688222Z",
     "iopub.status.busy": "2021-05-20T16:25:45.687331Z",
     "iopub.status.idle": "2021-05-20T16:25:50.495422Z",
     "shell.execute_reply": "2021-05-20T16:25:50.494933Z",
     "shell.execute_reply.started": "2021-05-20T15:34:03.334503Z"
    },
    "papermill": {
     "duration": 39.26338,
     "end_time": "2021-05-20T16:25:50.495583",
     "exception": false,
     "start_time": "2021-05-20T16:25:11.232203",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "w2v_preprocessing(train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "painted-receiver",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:26:58.416917Z",
     "iopub.status.busy": "2021-05-20T16:26:58.416370Z",
     "iopub.status.idle": "2021-05-20T16:26:58.451583Z",
     "shell.execute_reply": "2021-05-20T16:26:58.451043Z",
     "shell.execute_reply.started": "2021-05-20T15:34:11.258124Z"
    },
    "papermill": {
     "duration": 33.868456,
     "end_time": "2021-05-20T16:26:58.451774",
     "exception": false,
     "start_time": "2021-05-20T16:26:24.583318",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_data.drop(train_data[train_data.tokenized_sentences.str.len() == 0].index, inplace= True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "advisory-azerbaijan",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:28:06.641319Z",
     "iopub.status.busy": "2021-05-20T16:28:06.640488Z",
     "iopub.status.idle": "2021-05-20T16:28:06.644757Z",
     "shell.execute_reply": "2021-05-20T16:28:06.644291Z",
     "shell.execute_reply.started": "2021-05-20T15:34:11.602256Z"
    },
    "papermill": {
     "duration": 34.128773,
     "end_time": "2021-05-20T16:28:06.644889",
     "exception": false,
     "start_time": "2021-05-20T16:27:32.516116",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of sentences: 25633.\n",
      "Number of texts: 2834.\n"
     ]
    }
   ],
   "source": [
    "#create dictionary with all sentences\n",
    "sentences = []\n",
    "for sentence_group in train_data.tokenized_sentences:\n",
    "    sentences.extend(sentence_group)\n",
    "\n",
    "print(\"Number of sentences: {}.\".format(len(sentences)))\n",
    "print(\"Number of texts: {}.\".format(len(train_data)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "promotional-present",
   "metadata": {
    "papermill": {
     "duration": 34.211213,
     "end_time": "2021-05-20T16:29:14.968822",
     "exception": false,
     "start_time": "2021-05-20T16:28:40.757609",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Feature Extraction\n",
    "\n",
    "The following function, ```get_w2v_features``` transforms each sentence into a feature vector by averaging words vectors. In this way we can take into account the different length of each sentence, on the other side, by averaging on different words we will lose some word-specific information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "random-torture",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:30:23.181258Z",
     "iopub.status.busy": "2021-05-20T16:30:23.180300Z",
     "iopub.status.idle": "2021-05-20T16:30:23.182413Z",
     "shell.execute_reply": "2021-05-20T16:30:23.182823Z",
     "shell.execute_reply.started": "2021-05-20T15:34:13.995186Z"
    },
    "papermill": {
     "duration": 34.250759,
     "end_time": "2021-05-20T16:30:23.182986",
     "exception": false,
     "start_time": "2021-05-20T16:29:48.932227",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def get_w2v_features(w2v_model, sentence_group):\n",
    "    \"\"\" Transform a sentence_group (containing multiple lists\n",
    "    of words) into a feature vector. It averages out all the\n",
    "    word vectors of the sentence_group.\n",
    "    \"\"\"\n",
    "    words = np.concatenate(sentence_group)  # words in text\n",
    "    index2word_set = set(w2v_model.index_to_key) # set(w2v_model.wv.vocab.keys())  # words known to model\n",
    "    \n",
    "    featureVec = np.zeros(w2v_model.vector_size, dtype=\"float32\")\n",
    "    \n",
    "    # Initialize a counter for number of words in a review\n",
    "    nwords = 0\n",
    "    # Loop over each word in the comment and, if it is in the model's vocabulary, add its feature vector to the total\n",
    "    for word in words:\n",
    "        if word in index2word_set: \n",
    "            featureVec = np.add(featureVec, w2v_model[word])\n",
    "            nwords += 1.\n",
    "\n",
    "    # Divide the result by the number of words to get the average\n",
    "    if nwords > 0:\n",
    "        featureVec = np.divide(featureVec, nwords)\n",
    "    return featureVec"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "laughing-hunter",
   "metadata": {
    "papermill": {
     "duration": 34.329459,
     "end_time": "2021-05-20T16:31:31.653282",
     "exception": false,
     "start_time": "2021-05-20T16:30:57.323823",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Extracting Word2Vec features as a dataframe column. This cell takes quite a lot of time (~1 hr), it might be useful to parallelize the feature extraction process, you know nested ```for``` loops are always problematic in Python 🐍"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "middle-hughes",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T16:32:40.718875Z",
     "iopub.status.busy": "2021-05-20T16:32:40.717995Z",
     "iopub.status.idle": "2021-05-20T17:04:13.501150Z",
     "shell.execute_reply": "2021-05-20T17:04:13.501641Z",
     "shell.execute_reply.started": "2021-05-20T15:34:16.132014Z"
    },
    "papermill": {
     "duration": 1927.523999,
     "end_time": "2021-05-20T17:04:13.501945",
     "exception": false,
     "start_time": "2021-05-20T16:32:05.977946",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_data['w2v_features'] = list(map(lambda sen_group:\n",
    "                                     get_w2v_features(W2Vmodel, sen_group),\n",
    "                                     train_data.tokenized_sentences))\n",
    "\n",
    "train_data[\"w2v_resh_features\"] = train_data[\"w2v_features\"].apply(lambda x : x.reshape(1,-1) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "economic-uruguay",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:05:22.481359Z",
     "iopub.status.busy": "2021-05-20T17:05:22.480756Z",
     "iopub.status.idle": "2021-05-20T17:05:22.483540Z",
     "shell.execute_reply": "2021-05-20T17:05:22.483097Z"
    },
    "papermill": {
     "duration": 34.032743,
     "end_time": "2021-05-20T17:05:22.483671",
     "exception": false,
     "start_time": "2021-05-20T17:04:48.450928",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# save w2v features\n",
    "#train_data.to_csv(\"w2v_features.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "nutritional-medication",
   "metadata": {
    "papermill": {
     "duration": 34.324791,
     "end_time": "2021-05-20T17:06:30.968505",
     "exception": false,
     "start_time": "2021-05-20T17:05:56.643714",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Converting Word2Vec features to ```numpy.ndarray``` for visualization purposes with ```UMAP```\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "historic-occasions",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:07:42.005256Z",
     "iopub.status.busy": "2021-05-20T17:07:42.004608Z",
     "iopub.status.idle": "2021-05-20T17:07:42.511600Z",
     "shell.execute_reply": "2021-05-20T17:07:42.512246Z"
    },
    "papermill": {
     "duration": 34.744561,
     "end_time": "2021-05-20T17:07:42.512480",
     "exception": false,
     "start_time": "2021-05-20T17:07:07.767919",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "arr_w2v = train_data.w2v_resh_features[0]\n",
    "for i in range(1, len(train_data)):\n",
    "    arr_w2v = np.vstack((arr_w2v, train_data.w2v_resh_features[i]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "twenty-binary",
   "metadata": {
    "papermill": {
     "duration": 34.230927,
     "end_time": "2021-05-20T17:08:50.932013",
     "exception": false,
     "start_time": "2021-05-20T17:08:16.701086",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Dimensionality Reduction and Visualization via UMAP\n",
    "\n",
    "```UMAP``` is a nonlinear dimensionality reduction technique, something like ```PCA``` but more fancy! \n",
    "It can be used for both *unsupervised* and *supervised* problems and also both for *regression* and *classification* problems.   \n",
    "\n",
    "Since we are facing a regression problem, i.e. our target variable is continuous, we have to impose ```target_metric = 'l1'``` (thanks Leland McInnes, here a [github issue ](https://github.com/lmcinnes/umap/issues/257) as motivation). \n",
    "We opted for a 2-dimensional representation (```n_components = 2```) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "extended-apartment",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:09:58.835006Z",
     "iopub.status.busy": "2021-05-20T17:09:58.834355Z",
     "iopub.status.idle": "2021-05-20T17:10:22.095121Z",
     "shell.execute_reply": "2021-05-20T17:10:22.094488Z"
    },
    "papermill": {
     "duration": 57.23889,
     "end_time": "2021-05-20T17:10:22.095278",
     "exception": false,
     "start_time": "2021-05-20T17:09:24.856388",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "umap_emb = umap.UMAP(n_neighbors= 15, n_components = 2, target_metric = 'l1' , n_epochs = 500).fit_transform(arr_w2v, y=train_data.target)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "attempted-wellington",
   "metadata": {
    "papermill": {
     "duration": 34.9744,
     "end_time": "2021-05-20T17:11:33.949840",
     "exception": false,
     "start_time": "2021-05-20T17:10:58.975440",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Let's take a look at what we get by plotting this new 2-d space."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "swedish-exhibit",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:12:42.727734Z",
     "iopub.status.busy": "2021-05-20T17:12:42.726830Z",
     "iopub.status.idle": "2021-05-20T17:12:42.989198Z",
     "shell.execute_reply": "2021-05-20T17:12:42.988654Z"
    },
    "papermill": {
     "duration": 34.477967,
     "end_time": "2021-05-20T17:12:42.989341",
     "exception": false,
     "start_time": "2021-05-20T17:12:08.511374",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtoAAAIxCAYAAABzbR9yAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAACSNUlEQVR4nOzdd3jTVd/H8fdJ0t3SQtl77w2CKCg4cSLuhXuL4sAtOBAnejtRESc4UQQFREBBBWXI3nvvWbrbJOf5IxDgERWwzS9JP6/7ymX275PctP3m5HvOMdZaRERERESkaLmcDiAiIiIiEo1UaIuIiIiIFAMV2iIiIiIixUCFtoiIiIhIMVChLSIiIiJSDFRoi4iIiIgUA8/R3NkYo7UARURERCLPDmttOadDHE6VCs1tfn5mSI61M2PNj9bariE5GEdZaIuIiIhIRFrrdIC/k5+fyTmdnw7JsT4ZeU3ZkBxoHxXaIiIiIuIYawx+l3E6RrFQj7aIiIiISDHQiLaIiIiIOMpqRFtERERERI6URrRFRERExDkG/G6NaIuIiIiIyBHSiLaIiIiIOMaCVh0REREREZEjpxFtEREREXGO0Yi2iIiIiIgcBY1oi4iIiIiDjNbRFhERERGRI6dCW0RERESkGKh1REREREQcY7VhjYiIiIiIHA2NaIuIiIiIo7S8n4iIiIiIHDGNaIuIiIiIY6wBvys6x36j81WJiIiIiDhMI9oiIiIi4iBtWCMiIiIiIkdBhbaIRCSPx80pp7SiZcs6fPLJI2ze/DUjRz7DnDnvUVg4Hp9vAj7fBLze8QwYcBvGGAYO7MUvv/yPatXK8/77vdm2bTgfffSQ0y9FRKRk27eOdihOIX9p1tojv7MxR35nEZEiVqtWJfr0uZrCQi8333zuX27f//vMmEN/mfp8Ppo3v5lZs97B4/Hw/fe/063biQc9DmbOXMqZZz7Enj1ZxfsiREScMdNa29bpEIeTVr6u7XjRSyE51uh3Lgzp+6AebREJO3FxMbRr14g+fa6mQYNqZGbmUr16eebNW0WHDo0xxmCMCRbW1lqMMeTm5pOVlUuZMqUAcLsDX9qNGDGFFSs28uefy6hXrwojRkw5pNA2Blq2rMs55xzPp59OOCRLSkoimZk5IXrlIiIljyV619FWoS0ijktKiqdVq3pMm7aY//3vDu64o9tf7uP3W/LyCvD7LQsWrCIpKYHU1ESGD/+Np58eQnJyAsuXb/jH43Tq1Ct4fvbs5bz+ek/atWuEx+Nm1apNTJgw85D7P/vsTTz44GV8/fWvDBs2iS5dWvHww++RlZVbNC9cRESimlpHRCSkkpMT6NChCYsXr+Xxx6/mtNPaUFjopUaNCowYMYULL+xEXFwMB/9u2rJlF/Pnr+LGGweQlZVLRkY2R/O761jNnPkOrVrVY+fOvZQpk7LvWsuKlRtpUP+6Yj++iEgRCtvWkdTyde0Jl70ckmONffMCtY6ISHSpUaMCI0f0o3nzOkCg1SMzK5eUlARcLhfZ2Xm4XC4qVizDwIEjueuu7uzatZfJkxcwe/ZynnlmqCO5b7jhJR544DK++eZXvv76ycCVBurUqUzb4xrw54yljuQSEZHIoEJbRIpNzZoVeffdezn9tMDgwf6+amMMMR4PH330I507t+Shh96lUqV0vvxyEtu27ea++wY6nDxg7tyVXH31swCcfPK9/DD2WRIS4rAW1qzeAgRWP3G5XBQUFDoZVUQkchn1aIuIHLUF8z8gMTEueNlai7UwatTvPPb4ByxYsNrBdEdn8uT5pCSfR4UKpcnJySczM4fy5dNYuOgDUlOTmDlzGR2Ov8vpmCIiEka0jraIFKkBL91GTvYPrFr5WXDVD2stGzfu4KabB+D2nEq3C/pEVJF9sK1bdwdXIWnQoBqpqUm4XIbjjmvA2nWfEROj8QsREQnQXwQRKTL161fj7rsvxONxU6VKOrff8SrVq5fnww/HsnbtVqfjFbnJkxcwZ/ZyWrepD0CVKmXZuOlLTj/tQebOXelwOhGRyGBxZjOZUNCItogUmfXrt7Fu3Tb8fsvEiXMYOnQCTz75cVQW2RAYqW/Xrievv/5tcBWUMmVSmDnrbZo1q+lsOBERcZwKbRE5KgkJcX97W25uPnXrXU1c/Bl0PeuhEjNB8L5736Z2rR7k5RUEr/tz5jvcdPPZDqYSEYkQ+yZDhuIUaiq0ReSI3HvPxSxe9BGZe0cxY/rbf9nm/GA+nz+EycLD+vXbaNb0JrKzA5vZuN0u3n67F02a1HQ2mIiIOEaFtoj8owYNqnHjjWfz8su306BBNVwuFy1b1v3Hke2SavXqLZQrezErlm8CAssZTp32BjfdpJFtEZF/Yl0mJKdQU6EtIodVoUJp/N7xLF74Aa/97/bg9fn5hdx88wBycvIcTBe+CgoKadjwOiZMmAUEWm3eefce6tSp5HAyEREJNa06IiJBLpeL2rUr8cPoZ6lTpzIQGJVNTIznjjtfIy01iZdfGUZhodfhpOGv65kPk5s3Jrjc3+9/vEHlSpfh8/kcTiYiEl6sNqwRkWhWvnxp2ratz4jhT+PxHPiia/9OjsO+nsQ773znYMLIVLtWD1avGYrH4yY9vRRffvU4F1/0lNOxREQkRFRoi5RQHo+bL794nO4XdAQCLSExMe7gMnXWQk5OHk8+9QkDXv7KyagRa/PmnVx15bN8+VUfAM4+u53DiUREwpNGtEUkKtSpXZmRI5+mdq1KxMfHBlcPiYuLYcuWXWzavJMLL3qCdeu2B4tuOXbffPMbs2evoEWLOqxYucnpOCIiEkIqtEVKkCGfPMxVV54KHGgL2V9MvzVwJHfd/aaT8aLW8e17ctxxDZg3LzK3nRcRKU7WRO/OkCq0RaLYww9eRlJyAs8+9zm5ufmcecZxhxTYLw74kldfHc6WLbucjhrVfD4/U6cudjqGiIiEmAptkShTqVI6n3z0ICd0aExCQiwANWqU55prX+SqHs9y263nMm7cTEZ+97sK7BArWzaV3g9cwsSJc/lx7Ayn44iIhA0n1rgOBRXaIlGiRYvavPzSrZzSpeVfbsvJDqx5PX78TMaPnxniZLLfhx8/QNeux3F/70toUO86Vq3a7HQkEREpRiq0RSKYMYZfJr1M29Z1iY//606N27bt5qthv9Lr3rcdSCf/X+benOD5P2cNpEK5S7QmuYgI0bvqiHaGFIlAcXEx9Lj6VKb/8TodT2hCfHxccPUQay2bN+/iiqufo2KVy7n7noFaPSRM3HzTK8HzKSmJtGvXwME0IiJS3DSiLRJhGjaszsJ5g4KTGvfLyyvg40/GcV/vQdoePUxlZ+eRnZ1HUlI8APf3voQpUxY6nEpERIqLCm2RCHNyp6bB0Wu/3/LnzKXkZOdxzvl9yMsrcDid/Jvnnv2MZ/rfAECHE5o4nEZExHnRvAW7WkdEIsz7H/7IxElzWLVqEyefcj8dTuzFqWc8pCI7Qrzw/JdAoL++bNlSxMRovENEJFqp0BYJE9WqlWPk8L5488ZQmDua009vfdj7eb0+Tj39Ieo2uJ7ff18U4pRSFHJy8oNtP90u6OBwGhER51m3Cckp1FRoi4SIy+WiYsXSDHr7bkZ99zTvDLyb1NQkHnvkcnZtG8balZ9w3rkdcLkMbreLV1++zenIUkzu3TdB1VqLy0Tn16UiIqIebZGQqF27EssXvx/srYZAf/W6ddt4sPcllCqV9JeVQQa88nWoY0qIDB3yE+8OugeAfv2v56uvfnU2kIiIk4xRj7aIHLvRI5/CGHNIoW2t5Y+pi+n71BDWrNnK+vXbyc7O4+n+Q/HEn8OHH41zMLEUJ5/PR0GBF2stS5asdzqOiIgUE41oi4TA/AWradCgGtZa/H7L/17/hief+pTs7DwmTprLa6+PcDqihJDP56dl81s5rl0Dhn8z2ek4IiLOi9IRbRXaIiFw+VUv0Oz5L1myZD35+YVOx5EwsHz5RpYv3+h0DBERKUYqtEVCwO/3M3fuKqdjSBgpUyaFhYvfJybGTaMGN7J9+x6nI4mIOMOAyxWdOxirR1tExAEffHQ/6ekplCqVyMzZA52OIyIixUAj2iIiDti4YUfwvDtKexNFRI6EweJya0RbRESKyJ13vMmSxWvZsWMvx7W9y+k4IiJSDFRoi4g44PY7z6NhoxqULVuKC7prd0gRKcH29WiH4hRqKrRFRBzQvXvH4PlTTm3tYBIRESku6tEWEXFAuXKlgufnz13pYBIREedp1RERESkytWpVCp4fMeJ3B5OIiEhxUaEtIuIAv9+/77+W+fPXOBtGRESKhVpHJCoZYzipU1N27Mhg4aJ1TscR+Qu/34/FsnHjjmDRLSJSEhlD1C7vp0Jboo4xhj3bvyQlJQmAu+55m7feHuVwKpEDatWqQEqpRIwxxMbp17CISLRS64hEHW/uKFJSkjAmsAnItT1OcziRyKEqVCgNgLWWrMwch9OIiDgvXJb3M8Z8YIzZZoxZUBSvS0MpEjXatqlHg/pVMSYwqm1t4AfqpltfcziZyKHq1qsSPD9nzioHk4iIyP/zEfAm8ElRPJkKbYl4xhim//4qrVvVJT+/kGXLNlCvXhVmz1lJ2+N7OR1P5C8mTpxLQYEXv9/Pow9/4HQcERFHGZzZTOZwrLW/GmNqFtXzqdCWiJaamsS6lR+RkrKv3zXWw0uvDOf9D390OprI35ox801iYwO/ftu2rcfq1VscTiQiUmKUNcb8edDlQdbaQcV1MBXaErHq1a3MhB+eJSUlEQj0u372+UQV2RL2SpdODp53ezRVRkRKOBPSDWt2WGvbhupgKrQlYt179wVUrZIOgM/n58ZbX+WTIT85nErkn1WqVCZ4vqCgkC8+/8XBNCIiUpw0lCIR67MvfyEjI5uffppFXHI3FdkSEbZt28Oc2SspLPRywflPOh1HRCQsuNw2JKdQ04i2RIy4uBiqVinLylWbAZg8ZSFlKl3ucCqRo+Pz+enQXpN0RUTCkTHmc6AzgV7uDcAT1tr3j/X5VGhLRGjatCY/fv806WVS6P/8l/R77gunI4mIiEgRMKHt0f5H1torivL51DoiYe/yS09i3ow3qFSxNDExblq0qO10JBEREZF/pRFtCWvfDnuMbuceH7y8a1cmd/Ya6GAiERERKWrhMqJd1FRoS9h67plrDymys7LzaNjidnbu3OtgKhEREZEjo9YRCVsP3X9x8Hx2Th6p5S5VkS0R7cKLOpJbMIrMnO9o2LCa03FERMKCMYGdIUNxCjUV2hJ2jDHExcWQn18IwJ492aSkX4K10fm1kpQcb7zVE7fbTWysh0GD73U6joiIFDO1jkhYKVMmmbVL3ycxMZ47736bjKxcvhz2m9OxRIpEXl5B8ANjs+Y1nQ0jIiLFToW2hJXLLjqJpKR4AC7sfgJnnNvX4UQiRaNmzQpUrVoWYwzWWib+PNfpSCIiYcPtwGYyoaDWEQkbo77ty1uv3YrX62PPnmzue2iw05FEiswXXz0KgLUWr9fHhRc85XCiksflcvHIY5eRmf8NWQXDmbd4IG63/gyKSPHRiLaEhW+/fJSzz2yDMQaPx02Zylc6HUmkyJQpk0LjJjWw1mItnHHaQ05HKpG27v6MhITY4OXadSrSuUszfpqgbxdEnBROG9YUNX2UF8cZYzjvnHZAYLTv18kLHE4kUrRuv/NcYmM9gGX69MVMmbzI6UglStNmNTjv/HYkJMRijAle7/f7GfrFA9x065kOphORaKZCWxz3cO+LcLkCf/y8Xh+ndO3jcCKRolO7dkX69L0KgMJCL/fc9Y7DiUqW1m3rMmnyC3ww5D72ZuRgrWXhgjWc2K43LpeLUqlJvPDS9U7HFCnxonV5P7WOiKOMMTzz5NXBCWIDXv0Wv9/vdCyRIvPigJuC59et3cbs2SscTFNyeDxuvF4fSYlxgZYd4KrLXmTiz/OC91m0cB2Nm1Rn2tSlzgUVkaimQlscdVKnJod8ldvnqU8dTCNS9ObNXc0557bHGMMtN73qdJwSYcz4p+h0UhMm/Tyf8856iqsvfwnjch1SZAO0a3UPZcuWYscObYQl4iRjwBWlq46o0BZHeVxu/H6LywVffzsFn0+j2RJdnnv2S1at2sKiRWuZNVOj2cXJ5XKxeccQEpPiADi5SzOMMfw4dtbfPkZFtogUJxXa4qiff5nH089+QVpaEo/2HeJ0HJEiV1joZeiQn5yOEfWMMWzP+JS4uMCqItZahn7ys3aUFYkQriidNahCWxyTlBTP3OmvUblSOl26PkZeXoHTkUQkQo0a9ySxsTFAoMj+8P3x3H3Huw6nEpGSLko/P0gkaFC/KtWrlSM21k2Xk5o5HUdEItQ1151Cx46NAfbtujlPRbZIJDGhWXHEiVVHVGiLY1q3rA0YrIWZWolBRI5B4ybVeOvdO4JLhD728Cecf9bTDqcSEQlQ64g4ZsGitXi9XvILvMxfuNbpOCISge7oeW7w/JzZq3j9f985mEZEjoVBO0OKFLny5VL5bcpCzr+4H1u27HY6johEoKuu6Ywxhr17czjrtL5OxxEROYRGtMUxnwy+h9RSiZROS6JdpwecjiMiEaTTyU0Y+nlvduzYS/nyaTz20CdkZeU5HUtEjoXW0RYpWo0bVqNUSgLWWlav2ep0HBGJIC6Xi5Gj+xIT4wagZZO7WLlis8OpRET+Sq0j4ogRwx4Jnj/j1JbOBRGRiLNg6cBgke3z+Vm9Sh/WRSQ8qdAWR8yfvy54fv9qASLRonmL2pQpk+J0jKhVtVpZjAn83uh6ah/8fu0oKxLJ9k+G1PJ+IkXk8msH4PX6MMaQkpJInToVnY4k8p/FxcWwZdsXTP/zdZav+pC4uBinI0WV2FgPbw26gxXLN2KtZeofS5n6x1JHslSvXpaYGHVfisg/028JcURhoZepM5bRsUMjABbPepOU8leQn1/ocDKRYzf9z9dJK52MMYakpHji4mL0b7oITZ/7KrVrVwBgztyVnN75MUdy7Nj7BfHxga3eK5S+kuxsTcIU+a+0vJ9IERs46IfgeY/HzdefPehgGpH/rnadyhhjsNaSl1fA3r05TkeKKtWrl8MYgzGGFi1qB9tHQi0+PjaYo269So5kEJHIoEJbHPPFV7/x9HNfAoFtk+Ni9TW7RLZtW3djbWBUpknDmx1OE31ef2Vk8P31+23wfKjl5xVircXv9zN3zmpHMohEFW3BLlI8nnzmCxq0uJNevd/nyutfcTqOyH8yYfxsCgq8/DJpHhs37nQ6TtR5ss9npMZfykXn9admpesdy1G+9BXUqXYDKXEXOZZBRCKDOZoRAWNMdDbQiIgUAWMMDRpWZcXyTXi9PqfjiIgcbKa1tq3TIQ6nXIPa9oJB/UNyrMGdrwzp+6ARbQlLZcqkcOVlJ5GeriXSJDL0uvcC1m0cQqdOTVVki4gIoEJbwtSYb/sweGBPfhjR1+koIkfk/t4XU658Gvf2VjuBiMhRMVpHWySkAiOCFq/XT+eTmtKwQVWnI4n8rS6ntKBMejJer48Xnv3S6TgiIhImtI62hKWzu/fjlM7NKFumFOO+ewKPx8XmzbuoUk8rOUj4GfbNY7jdgS3BZ81a4XAaEZHIYgB3lG4SrRFtCUt79+Yw4rtp+Px+PB43xhgqVSrD3q2fOh1N5BDNm9ciKTkhuKbz/Hla7k1ERAJUaEtY+2jIz2Tn5GGtxRhDclI8Lpf+2Ur4qF6jPBBYCz4jI9vhNCIikcllQnMK+esK/SFFjpy1lpTyVzJl6lIKCgp5Z/CP+P1+p2OJAOB2u7jiqi6sW7uVWTOXU7v6dU5HEhGRMKIebYkInU579LDXP/7wJdx5S1fe/2gCffp94dhOcVIyPdmvBxdedCIAH74/npycPIcTiYhEHvVoi4SpR3tfRMUKpXn0wYvxZX7NpHH9nI4kJchdd58fPD9/gXqzRUTkUCq0JaK99tb3wVFsYwwnndCYt1+71eFUUhJUrFiauLhYjDH4/ZbR3093OlKJV7FSaZatHcSS1e84HUVEjoYJjGiH4hRqKrQloj3yxKdcf9ubAMGCu3rVsk5GkhKi173dgcC/u+FfT3Y4jQB8O+pxKlYsTeXK6Wzc8YnTcUREVGhL5Pt46ERcyRfxxtuj+XzYb1zSY8Bh75eYGKct3aXI3H3PBcHR7Ntued3pOAJs2bwzeD4+PtbBJCIiAZoMKVGj1wMf/O1tZcoks2zuWyQlxvFQnyFkZecx5LNfKCz0hjChRItH+1yOa986UQUFhWRnaxJkOOh+7rPMWfwGlSqV4fSTH3M6jogcoWieDKlCW0qEShXLkJgQh8vl4n8v3IDBMvjN29m4aSfVGqinW47OjTd2BSzWwkcfjHc6jhykZaO7nI4gIhKk1hEpERYuWsfdvQcz7NvfMSbw6dkAVSqV4czTWwKByZRVKpdxMqZEgIcfvYxKVUoDkJ2dR+/7BjmcSEQk8mnDGpEIN/ijCVx9w6vcce97+K1l/4rbS5ZuBGD8yD6sX/Iu/r3DWLPwbeeCSljrcd2pBD6mwbfDJ+PzaQMlERE5PLWOSInzzntjeee9sbQ/rh5r121ny9Y9ALRtUwcIjGxXr1aWunUqsmLlFgeTSjgqXToJCHxQ27NbW66LiPxX0dyjrRFtKbGmzVgeLLIBLunxP/x+i7UWv9+ycdMu58JJWEpJSSAhIQ6AwkIv/Z76zOFEIiISzlRoi+wz/uc5eNIupXztG0kodyW5uQVOR5Iwc9HFHYmLiwFgzJg/2bs3x+FEcqwefvxi1mx5n8eevNTpKCIl3v4RbW1YI1IC7Nixt1iW/fN43DRsUIXy5dOK/LklNGrWrhA8v2LZRgeTyH/14KMXUaZMCg8+chGbdn1CUlK805FEJAqp0BYJkXEj+7BoxqtsWfEe/r3D8GV8xfSJz1KpUhpD3uvJw/d1o2qVdCaP68fH795JnYOKOgkP1rKvtcjPZ0MnOh1H/oOxY2ZhrcUYQ1JSHJt2fcycJa/v68EXkZAy4HKF5hRqKrRFQqRRw6pAYLLl/lPL5rXo88DFXHbhiTzx8CV89/mDnNC+Pj0uP4nls1/Ht+cLFdxh5M8Zyygs9LF2zTaWL9vkdBz5D668+CXeeHUU1gbWHzLGUKtWBdZs+YDrbjzV4XQiEi1UaIuEyFnd+7N5yy6sDUy4zMsr4JuRU/ly+O/4/H42bd7N3IVrgvffX4wvn/06G5a8Q2pqonPhheTkBD7/6mFiYz3Uql2RlJQEpyPJf/T4Q0OoU/Vmli3deEjB/fDjlwBQvnwqXc9p5WREkRIhmnu0tbyfSIjMmbeaKvUPvwtlqcrX4PX6sdYya+4arrzkRNq3rQcE/vBXrlSaXWs/YOfuLC66cgC//bEklNEF6HhSEzwed/Dy7t1ZDqaRorJj+16Oa34fp53egq9GPoTL5eZ/L40gNTWJ5esHYYzBWsvVl7/M999OdzquiEQYjWiLhIHCQl9wRO2Nd36gw6mP40q9jEuueYXde7KDvaRly6Twy9inOP/stg4nLnkKC3xA4IOPRJ8J4+dSJvFK0uIv492BY6lc5dBdYt/7UFu7ixSnaB3RVqEtEsa+GTmNMtWv59pb3wpeZ4zhw7fvcDBVyTRv7iogMBlyb4aW9Yt2ixetZ+3abcEPwFN/17dIInL01DoiEgGGfvkbcxesZfjQ+6ldqyLDv5tG34cv5t47z+HtweN47OkvggWBFI/t2zNIjuuOy2Xwen1Ox5EQaF7/LlwuFxUrpbJp4+5DbqtbrxI33Hw6E3+ex/ixc5wJKBIlDOCK0i8LzdH8cTbG6C+5iMP294xuWzWYcmVLYa1lweL1dO3en02bd//7E4jIf/brtOdp0bIWAAvmr6XnLe8we9Yqh1OJ/KOZ1tqw7Dus2ri2vfOzZ0JyrEdbXRXS90GtIyIRZv+H4/sf+wS/3w9A00bVWL/4HU7t3MzJaCIlxtzZq4PnmzarwcTfn2Xb3qFcec1JDqYSkXCjQlskQg35/FdOPL0POfu2ijcGfhj+CPOnvqQJeyLF7O7bB3Fjj9dYsngD1ga+aYqN9TBw0B18N/Zxnnr2CqcjikQOA25jQ3IKNRXaIhFs6ozllK99I8tWbMZaS4zbRZNG1cjbNkTrbosUI2stX3/1O+1b3s8l3Z7D5/MHbzupc1N63deNZWvfcTChiIQDFdoiES4np4CGbe7ho08nYQlMKomJcbNi9muHrPssIsVj/I9zSE+6khef/Yb589YGr08rncxd95zLxN+fpVfv8+l8SlMHU4qEr2jesEaTIUWiyOsvXkvPW7oect21tw1kyBe/OZRIpOQZ9NFdnNS5Mbdc/ybfjemDcZng3Ir8/ELatbiPtWu2O5xSSqCwnQxZrUlte8/n/UJyrN4trtZkSBE5Nnc/+DHtujxGVlZe8LoPB96uNhKRELrlujdoWPN2fp24kC8++xW//8D27nFxMcxZ/IbDCUXCT7SOaKvQFokyf85eRedznw5e9vstlSqkORdIpAS77caBnHPak1hrg6PamqssUnJowxqRKDRrzmpSqlzHEw9dxPKVW1iybJPTkURKrN+nLKF0whW89Nr1XHp5J+676z2nI4mEFW1Ys//O6tEWERGJeK3a1Oahxy9k0MCx/Dx+gdNxJDTCtke7epPatveXoenR7tVMPdoiUgyaNq7Gd188wLldWzsdRUQcVKdeRX6a3I8zz2rNN98/yrhJT2oehzhOPdoiEtEmjX6Cc7u25rsvHqB/n0udjiMiDonxeIKbWhljOO74+jz57JW071CfDTs+4NfpzzmcUCR6qNAWKSF278kMnn/k/u7UrlnewTQi4pQlizfwyP2f4Pf7sdbi8/lZvnQjI394lKTkeJo1r8GOrKE0aVrN6ahSQpgQjWZrRFtEik3T4x8AApNOsJaLu7V3NI+IOOedt8aSnng11crewAltHmTg6z+wZcue4O1uj4vPhvfm+BMaOBdSJAqo0BYpIfLzvfQf8A1+v5/de7L4YvgfTkcSEYdlZeWxbElgVaKWDe9hyEcTg7dVrVqWH35+gulzBzgVT0oQlwnNKeSvK/SHFBGn9Hnma9xlriK91i2sW7/D6TgiEmbuvu09SsdfySO9P8HlMhhjqNegMmu3DQ72dYvIkVOhLSIiIod4582xjBo5HWstxhhSU5Podf95TseSKGVQj7aIiIiUID0ue5XxP84J7mp59nlt6Xp2S6djiUQUFdoiIiJyWJd2e5G2ze4nOzuf49rX5fPhD1KrTgWnY4lEDBXaIiIi8rdWLt9MXm5B8PLbg293MI1EI7WOiIiISIl1XLP7yc7Kx+v1sWtn5r8/QEQA8DgdQERERMLb7t1ZNKt/F+071OeXnxeQmpZEQX4huQeNdIv8F06MNoeCRrRFRETkX+3elcXY0bM446xWrN4yiI27P6RDx/pOxxIJayq0RURE5Iid261t8PxnX/cmJsbtYBqJBiZEm9VowxoREREJa489MASfzw9AaloSX3//sMOJRMKXCm0RERE5Ytu27eXGHm8EL3fq3Jhrru/sWB6JDlp1REREooa205b/4rvh0xn60cTg5Sf6X+FgGpHwpUJbRKQEaXNcXbILviUrfzivv3Wb03EkgvW6/X1ycvIBiI+Pof0JDRxOJJFK62iLiEjE6/3wRUz67YXg5UsvP9nBNBLprLW0bnQvfr8lMSmeMT/1pVRqotOxRMKK1tEWEYlyMTEedu79Arf70NUhnuo71KFEEi22bc3A5/PjdrswBk47oznDh011OpZEICdWBAkFjWiLiES5TduHHlJkZ2flUa1iD95+a7SDqSRazJy+HGstAFf00LckIgdToS0iEuUOXuc4P7+AyuV7sGtXloOJJJo80Ouj4PnS6UnOBZGIZQCXsSE5hZoKbRGRKHfOmX3Jyspj6ZL1lEm5DK/X63QkiSKNmlYLntdqNiKHUo+2iEiUmzJ5MRXKaPk1KR7720YA8rILHEwiEcuhFUFCQSPaIiIicszaHV8veH7T5t0OJhEJPyq0RURE5JjVrFUheP79dyc4mESkaBhjuhpjlhpjVhhjHv4vz6XWERERETlmZdKTMcbg91tmTF3mdByJQAZnJioejjHGDbwFnA5sAGYYY76z1i46lufTiLaIiIgcE7fbRc1a5fH7/Yz5/k98Pr/TkUT+q3bACmvtKmttAfAF0O1Yn0wj2iIiInJMYmM9lEpNwu+3bN+e4XQciWAh3LCmrDHmz4MuD7LWDjrochVg/UGXNwDtj/VgKrRFRETkmLRtXy+4TnthvpaNlIiww1rbNlQHU6EtIiIix+T8C9sCgd7aFq1qOppFIpcB3GHSow1sBKoddLnqvuuOiXq0RURE5Jhcc30XrLVYC26P+98fIBL+ZgD1jDG1jDGxwOXAd8f6ZBrRFhERkaOWmpaIx+PGGIO1lgd6feh0JIlgIezR/kfWWq8xpifwI+AGPrDWLjzW51OhLSIiIkelTr2K/DHrBawN7Ay5eNF65sxa43QskSJhrR0DjCmK51KhLSIiIkfl59/74fYEuk8/ev9n7u+p0Ww5dsYQNutoFzX1aIuIiMgRu/bGU0hOjg9sUuPz07/vMKcjiYQtjWiLiIjIEUlNS8Tv9+P3W1wu6Pvo5+zaleV0LIkC7jDp0S5qKrRFRETkX/W892yeevYKfD4/t143kOXLNjN/7lqnY4mENRXaIiIi8o9uv6srd993LhDYdn37jr0qsqXIGMJn1ZGipkJbRERE/taqLe+SmpoIQF5uAT9PmM9vExc5nEokMqjQFhERkcN6dsDVpKYmBtfK/mzIrzzQ62OnY0kU0qojIiIiUqIsWbwBCKyVvXtXpopskaOkEW0RERE5rE/en8TOHZmULpPC0A8nOR1HJOKo0BYREZG/NXrkTKcjSJQzRO/yfmodERERKcGat6jBtLkvMG/5qzRsVMXpOCJRRSPaAkDVSmmUT09m1oINTkcREZEQ8HhcTJ3zArXqVAxe1+2i9ix5ZriDqaREMjZqJ0Oq0I4yp3Wsj8/vZ+LvK47o/v16d+XRnqcB4Pf7cbndWGt5/q2f2JuVz5QZq5ny5+rijCwiIiH26BMXc3uvriQmxAW+t7dQUODly88mOx1NJKqo0I4Cj951Gv3u7wqA1+un0OvjjKvf5fc/1/zrY2+6vH1w2SaXy4UBjDE8fOdpGGPw+y2PvTiaPr3OICE+ht9nrub0K98hP99bvC9KRESKVNmyKZx0ahOuu/EUTujUEGMMWMBCbm4+Dav3JCsrz+mYUkJpwxoJO1UrpTF/fG9KJccHr/N4XBR6fRQW+o7oOW5+6CtGDL4BYwyzFm6gdZNqGAM5eQUkJcRhsVxzUVsSE2IBOLFtbbKWvMB1931G+bIpDPlmBjt2ZRfL6xMRkaLx5LOX0/PeswHw+QJ/H6y1zJy+gkd7f8rMGSudjCcStYy1R94TY0yUNtBEoLO7NOT7D28CCI5IAwx4dyJfjZrLzPnH3mvtdrtITIjhxsuPx++3LFq+hR+H3ha83W8tfp/F4wnMpd22M4tKrfv+h1cj4ah+3Uqc0L4+n3/9O/n5hU7HEZGjZIzhvofOp1qNslx2VUdiYtwA5OYWMHb0bL4dNpUx32lFkRJkprW2rdMhDqdxy5p26ITQ1BFtyt0Y0vdBI9oR6s5rTwyet9YyYNBEHn9p7BGPZP8Tn89PZlY+rw7+JXhdTK37eenx82lUpwLbdmZx6Xkt8eDCGEP59GTGfHILx7eqwYW3fMikP46sP1zCV0yMm0XTXsLlMrzS/2rK1LzZ6UgichSmz3+R2nUrYq0lP99Lbm4Bbnc827dlcEX3V5g3Z43TEUVKBBXaEeqxl36gfasa5OYVcsZV77J4xbZiPZ7fb7n/6ZHByy++8xPfvHs99WuXZ9XaHZzWqQEuY7iyexsm/bGCT9+8hlM71mPBks3c0Ptz1m3YXaz5pGhZCy6XwRhDWmoijRpUYfHSjU7HEpF/kFIqgYce787GjbuoXbdioAcbAEu30/uzauVWsrPzHc0o8neitUdbrSNSJPo9cDadO9Tlpge+wO12Mf+nhw+5vdsN7zF/yWbWrt/lUEI5WjmbPyI+PgaAzKw8Uqvd6HAiETmcjic14t2PbychIZbExLh97X1+4hNi2bhhJye0fFgFtkCYt4589lNoWkdalVXriESgPi+NCZ5PTIilsNAX7AcEGPlBoPUgKzuf9KaP4PX6Q55Rjs4bg37kgbsDk6dSkuNwuQKr0IhIeEgplcDQYfccWEGEQOvf7u2ZtGv6AF6vX/MrJCIYiNp1tLUzpBS5nNwCEur05qV3JpCxN5eMzFwgMDEnOSmOJb885nBCORJffPP7QZesimyRMFCmTDLlK5QC4IqrO9H+hAbBItvr9XH2qf1o1eA+srPzVWSLhAGNaIeRlOR4Rr53LXv25nHDA1+xZ2+u05GOmbWWh/uP4uH+o4iJcbNx5tOkl04CYMNm9WtHgtnz1mKtDa5qk5QUp6+fRRxSsWIqU+e/QFJSYDnXHpe+xh9TllJY4CUrq5D77vyA74bPcDilyLGL1pFfFdphokaV0qya/Ejw8hMbT+fep79zMFHRKSz0Ub75Y/S+vQuVypU6ZFKlhLfFyzbRqH5lfD4/Pp/afURCrULFNJ7/Xw9OPaMpCQlxwetP7NiQPg9/To1yN+vbJpEwpkI7DLhchhnf3X3IdbMXHvs62OFqwNsTnY4gR+m4Ln254eqTcLtcJCbEkZenr6JFQqHNcbV57d0bqd+gCub/rcawetVW+j/5DYCKbIkKxoA7Snu0VWiHgU7tapGWlhi8/ObHU/jkm1kOJhIJyM0t4I6bTqNhvcq83P8qYstdx9GsVCQiRy6tdBKPPXkRV/ToRGyc56Dl+SAvr5BH7h/Kz+Pms2mjVm8SiRQqtMNAXr4XYwJ9zT37fss7Q6c6HUkkqH6dwHq8Lhfcen0X3vngZ6cjiUQNYwxNmlXjwkvb0/Pesw8priGwpv1Hg3+m70Of6xslkQikQjsMnHdqI4zfYoF1G/c4HScslUtPJiMzj4ICr9NRSpzvxszignMDS46+/sI1KrRFikjfZy6lxw0nk5wUh9vjxuVyYa3FWsuyJRt56N6hzJqxitzcAqejihS7aN2wRoV2GOh1XUeMCawjmZQY63ScsHNl97Z88vo1GAMzZq+lY/f/aR3uELromtfx7foEYwxut4tLLmjPsBHTnI4lEpE8Hjc33HIqTz532SF7DeTmFJCVmcvPExbwzBPD2LJpj3MhRaTIqNAOA4kJMcEl1Mb/ttzpOGGnU/s6gQ8ixnBcqxrkrX6Vd4b8Rs9HhzkdrcQYNXY253ZtBcDH79yiQlvkGLz02jVce9Mpf5ncuHNHJhef+xIL5q1zJpiIw7RhjYRMRmae0xHCzuMvjGL7zqzgJDxj4PZrOpG5fACVK6U6nK5k6H71a+yfAxkX46F1ixrOBhKJME2aVeW6mw8tsnNzCujXZxgNq9+lIlskSqnQDgMr1u4IFpFP3Xuaw2nCz87d2VRs8Sh9XxoVLPaMMSQmxLJ+Rj8GvXQFsbH6cqY4+f1+nnt5BPj9WL/l1utO+cukLRE5oHHTqixY/Sprd77LeRe0ZdPGPcG16DMycrjxqjepXvYWXn95tMNJRcKD24TmFGrmaJbqMiZKx/UdNmdML5o1qASA1+cnrv6jDicKXy6Xi6mj7qNN8+r7Cj2D31o+/moadWqWpd//fuCn35Y6HTNq9X/8Yh7qdQ7GGJ5/bTSP9VP7jsjB6jeszPAxD1K2fClcLoMxhmm/L+PcU5+lVGoi5SqUYuWyLU7HlJJpprW2rdMhDqdZq5p25C+PheRYdVJvCen7oBHtMHDPk4EdII0xuKN12m0R8fv9tDt7AFVaP8b6jbvx28DOkxd0bcFJx9fj7ecvdzpiVPtqxHSMCRQPD919NnVrV3A6kojjKlUpzYrNA1m5ZSCvvnM95Sum4nIZ/H4/e/fm8OyTwwHYm5GjIlvkMIyxuEJ0CjUV2mFg0vTVDPl2Fj6fnyHDZzodJyJs2ZZJjXZ9qdTiYdqc+QKDP5uC1+tjyNfT/3Lf7me35PorOjiQMvrMXbCOr0ZMx1qLMYbF055X246UWA0aVqZ23QrMWjKAUqkJpJRKoEHDKlgLe/Zk06X9E9SpcAe//7bE6agi4hC1jkhU635WS74efBMAefmFXHH7B3z343yHU0W29DIpbFv2RnClnL7PfUP/l793OpZISNSsXZ6PvryLho2q4HIZCgt9eNwuzL5vI38cM5sbrniLwkKfw0lF/iJsW0eat65hR/0amtaRGim3qnVEpKjExwdGW40xxMfF8O0Ht5K35jUa1avocLLItXNXJv1f+Q6/P/C5u+8D3XG79atEolPpMklceNnxfDHiPl549Wp+mPQ4jZtUDf6bd7kMfR76jHmz1/DSsyPocfHrKrJFJEjf+UpU+/zbPznlxPpce+nxuN0ujDHExLhZMOlxRoyZw0U3D3Y6YkTq++xwtu/IZMDTV7B56x6aNa7KnPlansxpaaUTyc0pID9fO6gWhZffvI6rrjsJ177Rauu37NyZGThvLdnZ+Tz/5DcMemsCg96a4GRUkYhmAHeUNk1oGEqi3s29PyO2+t1M+G3JQWtxG847szlb5z/PBV1bOJwwMr0xaDyDPv6Z8mVTmD7hCerWLu90pBLtyus7Mn/dy6zY8QYLN7xCuxPqOh0pYpzWtRljf32c7pe0C16XkBBLjxtOPuTbGr+1XHf5mzz+wGe0a/ogtcrdxrtvjXcisohECI1oS4lx5uVv0rpZVaaOeSg487hcegr33noKk6evpM+9ZzHm54X8OHGRw0kjx8bNe4iLiwFg6fQXcZe9ztlAJVjXc1sG1zYvlZrANz/eD9aycN567rnlI9av20V2Vr7DKcNHi1Y1eH3wjTRsVCV43dsf3sKoETMpLPSRm1vAu2+O45zz2zDqu5mMGzOH+XPWkrEnh+l/aAdfkaIWrYuuaTKklFi97ziN6y/rQM9Hv+TlJy+iRZNqAKxYtY2+L33PlyO1AsyR8O34KFjgVW58N1u2ZjicqGQqXymVGUueC7Y5GDhkF0JrLb//upw+D3zJ0kWbnAnpsMTEWH6a9iQ1apYLjlTv/7drrSUrM4+6FXtyNH8XRSJI2E6GbNG6hh07OTR7iFROui2k74MKbRHgl2/vo2P7uof80d2zJ4f0xr0dThb+hr57G5dfeDxbt+2hSpN7nI5T4iUmxfHBV7dzQqf6h12Xf+XyrdSoVQ6/38+qFdv4YeRsBvQf5UDS0Lj8mo6c3rU5p3dtTmycJ7gO/MF/+/bsyeaGy99i3px1ZO7NdTCtSLEK40K7pv1xcmhWHamUpA1rRELu1Ete48Mv/sDvt8E1otPSEhny5vVORwt7V9/6Dp5y1/HW4J/I2jCIpx+9yOlIJVpOdj6Xn/Mq1UvdwcVdX2Hc6Lls2byH3JwCrLX4vH48HhdxcTE0bFyZex85hxlL+3PqmU15oM+5xMVFZkdh/wFXsGLrmzzy5IXB694cfCOvvXM9517Qhrj4GFyuwJ88ay07t2dy3x0fUSn5JhpUvpspvy5VkS0iRU4j2iIHKZWSwK4lA4Ij29+OmcPFNw1yOFVkyFj7LilJsVgLHc58mumzVjsdSf6f5JR4KlUpzQdf3EbN2uUOrKax78PlftZahg39g3tvH+JU1L+4/5HzuPG2U3B73CQlx/PJ4Ik8ev/nALRqW4uxvz4evG/tcneQnZ3PmEmP0qZdnUOeZ+L4BQx4diR/TlsV0vwiYUAj2oR+RFuFtshhjPm0J4VeH92vfye4XrT8s/HfPsgpHRthDEyZvpxTur2g9YTDmDGG+x87l8uuOp5JPy3iyus6Bq7HBBq8Laxcvo2unfrz0ps9OKdbKwywft0uxo2ey7Q/VjD2+zmccFIDWrSqwbTfl9PjxpPYtiWDwkIfY0fNIWtvHj6/n9Lpyezdk4O1sHb1dgDqN6xE1Rrp1K5bgZM6N6LzqU1xe1w8/djXvPvGuEOy9nnmYu68t+sh2a21VEwKbEZVt35FpszpH7y9csrN+Hx+kpPj+GbsgyTEexjywa98PPgXCgq09KGUWGFdaI+b8vi/37EIVEy8WYW2iESeuLgYxn51Hyed0AAAn89PbMWbHE4lR+qJ5y7ihtu74PG4A6PbFvLzvFx01suMmvjwIfMXALxeH6++MJp7HzoXt+fQiYV/p7DAyyP3fgYYnn/1amJi3IfcboyhsMBL1bRbD7l+6Dd3c1rX5odct3rVNjo0OzB5qk272txwaxf6Pf4NWzbvOZa3QCTaqdAm9IW2erRFpEjk5xfSpdsLQKBg8njc9LzpVIdTyZF66pFvqJHWkyrJt9Op5RNMGLuAL4b+zpyZa1i2ZDPW2kNX47Dg9foDo9/8e5EN4HK5aNikyiFtKwez1jJ96l+XzrvzxsFMnbKcLZszeOyBz6hboechRTbAzOmruPPG91Vki0QgAxhcITmFWmTOehGRsDV/0QaaNa4KQId2dXlz8E8OJ5KjtXL5Vq65+M3g5c5tn8TtdnHxVcdTkO8jLtbNimVb+HPaKubPXUeb42qzdMkmbrr9VPx+S3x8DNYCBjL35pKcHMeiBRvYtGE3gwf+jM/rY+eOLNqfUI9SaYngt2zduodPP/qNKb8s/UuejD05XHDGCyF8B0REioZaR0SkSLndLr4dchcVypai66WvsHtPttORREQkjFtHWrapacdPeSIkxyqfcENI3weNaItIkfL5/Jx/5WuUTkvi9M5NmPDLQnbtVrEtIiIljwptESkWo7+4lzYtarJh0y7qtHnQ6TgiIhK2jCP906EQna9KRBxXsXwqHo+LmtXLctv1XZyOIyIiEnIqtEWkWAwe+kvw/OvPXeVgEhERCXcmRP8LNRXaIlIsXnhtDBBY9s3tdvF47/MdTiQiIhJa6tEWkWLh8/lZu34nNaqlA+DWx3oRETksgzHR+UciOl+ViISF+u0eprDQh7XQsllNp+OIiIiElEa0RaTYeL0+YmICW3q3b1Pb6TgiIhKGAjtDhr5/OhQ0oi0ixer7H+dgrSUtNYFKFdOcjiMiIhIyKrRFpFjt2JkJQFxsDEv+6O9wGhERCUfGuEJyCjUV2iJSrKb+uSJ43u/3O5hEREQktFRoi0ixGjzkN+bMX4PP66XPc986HUdERCRkVGiLSLFr2bQ6Ho+bl/td7nQUEREJQ9qwRkTkP3K7onNWuYgc0OnUJkxbMYCJ857lhJMbOR1HxFFa3k9Eit3wUbO46Lw2fPfjHKejiEgx8Xjc9HnpCi6/rhMAaenw4Yh7AJj66xKu7fY/B9NJeDOYKB37jc5XJSJhpXmTqlhrOfmEBk5HEZFi0q5jfc6/9Pi/XG+M4fiTGjL+z6dZvvtdvp/ch3IVUh1IKBJ6KrRFpNjVqVkOYwxppRKdjiIixWTx/PXszcghP89LdlYeSxdswO/zY61l+uSl1KpbAZfL0KhpVf5Y+iLX3X6q05ElTBgCH8hCcQo1tY6ISAioN1sk2u3emcXJjR887G0xMW5+XfgC5cqnAIGi6qIrO/DR2z+FMqJIyEXViHaDGmXwTn0Y37SHGdznLKfjiMg+Y3+ah9/vZ/7CdU5HEREHFBb66FC/N1eeM4A//1jB3JmrefKBL5yOJWHE4ArJKdSiqtC++/LjMAaMy8X157cia/IDnHdSPadjiZR4Z53aFJeBZo2r0u2slk7HERGHTJ+ynMvPeokLT3mOmVNX/PsDRCJcVLWO9Bv8G7dd2ApL4IvqhLgYRrx8Cfe9PJ5aVdPwev3ExrjpN3gy23fnOB1XpERISzu0Lzsnt8ChJCISKSpUSuPnOc8QG+vh6Qc+Z8jgX5yOJMXKmTWuQyGqCu0tO3OI6fACs4beSLN65YPXP3nbSZRKigtevuyMxuzam0e96mUwLsPiVTu4e8BY/ly4mb1ZKgJEilKD2hUAsNayYdNuxk9a5HAiEQl3L719HXFxMQA8MeBKtm/PZOzIWQ6nEjl6UdU6AuD3W1peOZhON3zMzoxcVqzfxZS5G/DbA/eJj/NQf1+RbYyhUe2yjB94Nbt/eYAt4+7hjYfOpMc5zQBoWCudpIQYh16NSOTr0vHAkn7LVm51MImIRIrPP/7tkMsvD7rBoSQSKsa4QnIK+euy1v77vfbf2Zgjv3MYcbkM9WuUoUubGrRrXJmXP5vGqFcvo1qlQ9fx3L/si8/vJ7/Ax+BvZ3HLhW2IjfVgXAavz88tT4/i4+/mOvEyRCJSi6bVmD7uMQzQrceb/PDTAqcjiUgEqF2/IqMm9yEmxkN2Vh7jR8/h9ee/Z/2aHU5Hi1QzrbVtnQ5xOK3b1LWTpw0IybGSYrqH9H0oEYX2P6laIYVbL2rNA9d2wON24bcWvx9y8738OGUFF57aEHeMB2MM+9+rTtd/yO9zNjicXCRytGpWnbg4D1P/XOV0FBGJILGxHjqf2Yw3P7oFl9uFt9BHy+r3kKe5HscirAvtKdNeDsmxEmMuCOn7EFU92sdiw9ZM+gz8hT4DD0y0qFwuhZy8QoyBhau20/vaE0lOigVrwRi0JrDI0Zk9X8v6icjRKyjwMu772fj9FpcbPDFuvhr7AOef3N/paCJHpMSPaB+N+3p0YPXG3Xz78xKno4hEjJTkeG6/vjN/zlnDz7/pZ0dEjl6LNrX4/IfexMYGxgc/HTyJJ3p/7nCqiBO2I9pt2tazU6a9EpJjJXjO14h2uHplyB9ORxCJOBNH9KZVs+r4/Zb0+vewNzPX6UgiEmHmzlzN0w9+zjOv9gCgWs1yDicSOTJRt+qIiISXVs2qY4zB5TKUKZ347w8QETmMLz6azItPfMPsGSs54eT6zFv/P+LjNV4o4U2FtogUq9zc/H0TiS1tmld3Oo6IRLBBr40jLtZNbKyH5FIJTFv+ktORpIhoC3YRkWPwx58rwfrBWgY8ebHTcUQkwv0wIrBxjTGGlFIJwaV5RcKRCm0RKVY33TsE/75J19WqpBMT43Y4kUj0O+HkRgwd8yC33X+201GK3Nv/+5GDF3I4mkUdJHyZEP0v1FRoS4n13vMXkbekPx+9dKnTUaLa2vU7eeO9n4HA6phLfn/K4UQi0W3ysgG8/+29tDm+Lnc90o1Gzao5HanIvfLMd/j9lgVztXSohDcV2lJiXXdxW2I9bnp0b03Own5cfm4LpyNFrcGfTgYCX/VWrVzG4TQi0WvomAdJL1vqQDuFhQ1ro28nxbcG/ECdtNs4/6RnnY4iRcJE7RbsKrSlxBr/6zIsge2H4mI9DHnlMpISY52OFZUWLd3M/MUbsdbidhlaNqnqdCSRqFKlWjotj6tN6r6Vfay1WAsd6t5D5l4tqSniFBXaUmKdfcOH/DJtJdZajDFYCzm5hU7Hilq/TV0GBEa1f/3uAYfTiESHt7/oyYId7zJuznN8OOI+Phs8iZ9/mMvLT31D4/Rb2JsRnUV2TIybF9/swfvDepKYFOd0HCkC6tEWiUKnXPUeGZl5AGRl52tSTTEaOXbOgQtaJEDkP6lUpQxjZvTjpNOb43IF/pQblwu3y8WdV73F4Nd+dDhh8brt3jO5uMeJdDmjKSMmPeJ0HJG/pUJbSrxlq3dgrSU1JZ7mDSs5HSdqTfhlKe8N+Y3V63Zw/FnPOR1HJGJddHVHxs7sT43aFYBAm8ie3dk8ed8QPv9gkrPhQuSEzg2D51NKJTiYRIqCAfVoi0SrsmUO7FZ49sn1HEwS/R57bgTJiXHM+PEx2ras4XQckYjUtFUNjMvg91vmz1rNAze/x4l172XE53/g8/mdjhcStetWCJ6/s8e7DiYR+Wfau1RKvBj3gc+b3c9szvPv/upgmuhWp2Y5kpPjMcbQvnUt/pyz1ulIImHN7XbT89Hz+P6LaaxavhmAV58ZQVZmHjP/WM6kH+c5nDD0GjapQrkKqcHLi+avdzCNFBUn+qdDQSPaUuItXXVg6atre3/pYJLoN33WGp577Qf8Ph+v97+Mwg1vaQMbkcMwxtCsdU3mbXubW+45m++nPkV6uRQAMnZn8/KT35TIIhvggsvbB0uyjet2kqdJ7BLGVGhLiZaSHEeXDnUA2J2Ry5JV2x1OFP36/28MCQmxGMDldtEqCjfTEPkvzr6wHT/Pf4EhYx465Hq3Rx9KAVofVwtjwGXgvdfHOR1HioTB4ArJKdTUOiIl2tLx9+NyBcZGvD6fw2lKDrt/AXNg/uKNjmYRCQeJSbGccX4bej3enTLpKcTEevD7LT6fH2MME3+YzbbNe5yOGRaa7Zvf4fdbRnw13eE0Iv9MhbaUaNkHfeX4ybezHUxSsiTW7MkzD53PiwPHk6uvfaWE+3jMQ7RuXxeD3denaiks9DLpx3kMfOF7li7c4HTEsPHR13eRsG9jsUXz1kXtOuElUXA30yijQltKtNbnv87g5y5i1oKNvPDuL07HKTEKCnw82O9bp2OIOOqci9vT4/bTadwiMEJrMfgKvXzw5jhee0Y/H4fTsUvDfRuMWRbN17dhElrGmEuAJ4FGQDtr7Z//9hgV2lLiVCqXwuxRvYjxuGl/4ZtcdtdnTkcSkRIiPiGWL396nFr1K2GtxeVy4ff7MRh8Xi/dTnyStSu3Oh0zLHk8btweV3B7+T736Xd3VImM/eIWABcCR7ympAptKVHcbhcjB11LuTJJANx8eTseeuEHh1OJSEnw4cjetDmxAXDga/KC/EIGPDGM0V9NZW9GjpPxwt6wH+8Pvm/ZWXnk53sdTiQljbV2MRxdm4sKbSkxPB430765nSb7Njrwev0M+kITaUSk+L0/IlBk7297sNayfNEGLjrpKaejRYyWbWsH37+h76vVT45ZWWPMwS0fg6y1g4rrYCq0pcTImN2X+NjAP/nsnALaXvAGK9fudDiViJQEzQ8qEv1+yz093mLSj3OdjhVRcnPySUyKo7DQx/N91cMedWzIdjXdYa1t+3c3GmMmABUPc9Nj1tqRR3swFdpSIrhchvhYT/AP3Tk3fsiy1Tv+/YEiIkXgqXs/4a7HuvPn70t57I4PnI4TkU5p8wSnntWc0d/OdDqKRDFr7WlF+XwqtKVE8PstD734A316nsKwsQv47c81TkcSkRJk1LCpjBo21ekYEW3Lpj18+v6vTseQYmFDOaIdUiq0JWo1a1CRmlXSGDVxKdZaBrw/mQHvT3Y6loiIiEQgY0x34A2gHDDaGDPHWnvmPz1GW7BLVOrQqjp/fHUbn/3vcm67op3TcURE5Bh9Mepelm1/g5t6nup0FCkulsCIdihO/yWmtd9aa6taa+OstRX+rcgGFdoSZdxuFzNH3Mlvn99CQnwMEGgbERGRyNPvpUs48eQGJMTH8MjTFzodR+SoqXVEokabJpWYPrxn8LLfb7nl8W/5/HvN7BcRiUSX9jgxeD42RmOD0St6e7T1r1aiQlysh6nf3HnIdX1fn6AiW0Qkgr3+8g/s6yvQt5MSkVRoS0QzxnDLZcdx/YWtOXifppE/Lea5tyc5FUtEoshlN5zMvB3vMW/He9z1WHen45Qob730I6e1e4pvPp9Kx+Z9nY4jxcnvD80pxIy1R/4J0Rijj5MSNjweF7NH9KRRnXIUFHh55u1JdGhVg8f/N465S7Y4HU9EIpgxhhvvOYsLr+5E1Rplg9fv2pFJ54b3OZhM5JjN/KeNWpzUtk09++fvr4bkWCb+3JC+D+rRlohUr2Y6CXEe6tcogwFcLhdfjpnPs+9oW14R+W9OPKUJb33RC5fLHHK9tfD0fZ84lEokykVpj7YKbYk49153Iv3vPZ0tOzJ5/5uZHNe0Cnc+PYqV63Y5HU1EIlC1WuW48JpOzJ22kklj59Krz4WHFNmZGTl8Pngibz43wrmQIhKRVGhLxDjvlIbk53tpUq88brehUrkU7n32B/ILvE5HE5EIY4zhqltP4/RurWlxXF0A/Hf4ufrMZxnyznieefMGCvK9PP/wp4z8Yiper8/hxCJRzEbvqiMqtCUijP/wOk45vg4Avfp/z5qNe5g8c42KbBE5KsYYbul9Lhdd04nylUoHr7PWYowhOzOP77+cyvdfart0EfnvVGhLWLv+otbcd92JNK5bPnhdzSrp9H7hBwdTiUik8HhcVKlZjstv7MJlN3QhN6cAT4yb2LgDf/68Xh8F+V5uv+QV1q7c6mBakRJMI9oioZMQH8NnAy7m/FMaA4HNZ4yB+cu2qsgWkb/ldrvo0KUpDZtX4+SuzWnWunbwNmMMSSnx5OcVUpDvZc2KLcyaupwXHv7cwcQiEs1UaEvYueTMJnw24NJDJiPtzcqj5ikvk5md72AyEQknsXEeKlRJ5/iTG/Hz6Fk89foNtO/cGJfLhcsEPpzvbwvZv5Tt7p2ZXHry02TtzSE3p8DhVyAi0U6FtoSFhPhY3n7iXLqf3ojkhLjg9T6fj7WbMqh35v8cTCciTiuVlsTlN3eh25WdKFs+FZfL4Pa4AfAW+uh+dScaNK2O2fcB3VpDYDdBHz6fnwdveIffJizAW6hJjSLhxzqymUwoqNAWx/XrdRqP3NwJYw5ds3bD1r2cdv2HLF+706FkIhIqpdKSOOOCtnTvcTLTflnI8E9+5a1h91O+Uhr/e+Irbrn/XMqUKwVwyO+K/ZMY9+zKYuDzIzj5zBbMm7WK5Qs3MO7bGeTm6FswEXGOCm1xVKnkOB7+f0V2Zk4BZ938MX/MWe9gMhH5L+LiY6hQpQz1m1TjxNOaU7dJNUZ9/hs9H7uIHdsyuKXbi7z+xT0kJMaRlBxP6bIpwcc2aFaddic3pkqNsoEdGu89l7j42ODtB+9ovHtHJk/1+pBpvy4mP6+Q9/83OqSvU0SKiCZDihy7CmWTufeaExj/xwp++mNV8Pq9WfmM/GkxZ5xYF6/Xz8AvpvP4qxMcTCoi/8YT46FFuzqkpCWRm5PPgj9XkZ2ZS+NWtRjwyV2USksiJvbAnxcT6OKgQf8rcbsM1ZLiuP3hC6heuwLGQGxcTLCXer91K7dRoXIZkpLjebP/N8ydtpwLrj6JZQvXUbZ8GmOHT2PH1gwHXr2IyJFToS3Fpl6NdK44pykXnt6U3XtzObF1De659gSS2zyN13vgk+vFvb5wMKVIeHC5DPGJ8eRk5TodJah+0+pcf9+5lC5bihbt6gHw2pNfccej3YmLiwEC+0ysXbGZyzo+zv8+7UVa+oGR6f9fPPv8flzGhbWWz9+bQKnSSSQmxbN+9VZO7tqKWVOXUSY9hR+/ncb3n//+l01iXn/66xC8ahEJOYtGtEWOVM0qpZk/4g4S42PAGAyQ7/Xj91tWrd91SJEtIoER4p/WDMIYyMsr4JK2vcnYleVoprj4GD4c1ze4+s/+9q4rbz09WGQHrg/0VwMsnLWaE09vDkBBfiEulyE3p4D1q7aQk5nHQzcOxO1ykbEnG4De174VfJ7nHhgaktclIhJKKrRLqA4tqnLH5ccxfPwilqzZyc49uWzblX3Uz9P9tEaUSopj+ITFwaX3hr5wIYnxB74KtsYwf9lWLr7nc7Zsd7Z4EAlH5SqmBZeii4+PZeS8V9myfieP3/QmKxYe+1wFT4yb+PhYGrWqSXJKIr/+OAffEW4lblyHTk72+wNL5L3w4Cc0blmLy285nZycPH4ZM5uv3gu0e93f43Uq1yiHz+tly4Zdx5xbREqa6N2C3Rz8td6/3tmYI7+zFLu4WDc/DLyKdk0q89WPC+n5wg/Uq57Og9edwPY9OVgL/d/7jR27cw553Nh3rub0DrX/8ny79uaSlhwPBnbsySUhPoaCQj8uF4z4eQm79uRw44WtSUyMJS+vkLhYD3H7+jAnTV/NqTd+DMCA3mdy77UdAFi/JYMejwzntz/XFvO7IRLZvl/wOqllkoFDWy6GvjGa3Jx8tm3axZI5q1m7fPO/PldaegodTmnGY6/dGByJ9nm9DH3rBwY9/+0RZ6rfrDoXXtuF8SOmM3Py4mN4VSISRmZaa9s6HeJw2raqbf+c1D8kxzJpV4b0fVChHcEuPLURwwZcfPgbjcFnLe8Pn8Xt/Q6dhZ81/RES42P5u//vjTFYwLhcwfv4fIHNH1z/b5Rrv99mraXztR8GLzepW571WzLYm6WltUSOlMtleGLgbXQ5/7i/3GYMWK+PpfPXUqNupX33d/H4zQO5p/8VVK5ennnTl/Nqn8949/vHiIn14Ha7Dnyz5Pfz8Wvf896LI0P9skQkPIR1oT1j4tMhOZardI+Qvg+uUB1Iit7vcwNfKRtjgqNW/3992anzNvzlcbf1G01efiFZOQVMmb2Obbuy2JOZxyffzyUv30uh10dmVv4hvdTZeQV/mZjkt5ac3EK+GruAbj0P3cJ44YptKrJFjpLfb3nitrd57r4P2bs7i8WzVx1yu8vlon6zGiSlJJCUkkBCUhwvftKLarUq4na7aNWhAWdeeDyGwCYuu3bsxVpLbnY+Lz74Me8P+M6ZFyYiUkJpRDvC3dS9FW8/fk5wpHnrzixmLdnCO1/9yfQFm46p7/pgMR43FovX6ycpIZbUlDiMgYzMfLK0fbFIsTLGcO6VnShTPpX6TaoxYcR0zrvqJFqf2JCYmEDb1s6tGaRXSAUCH66v6PgobTs1xuNx8/UHP/3tN1ciUuKE74h2y1p2xs9PheRYrvRr1ToiIiL/rEKVMrRoX59xw6dSr1k1mraty6ihv1KoLcZF5PBUaBP6QlurjoiIRKCtG3cxbvhUAJbPX8/y+dpJVUQiWJSuOqIebRERERGRYqARbRERERFxUPSuo60RbRERERGRYqBCW0RERESkGKh1REREREScpdYRERERERE5UhrRFhEREREHaTKkiIiIiIgcBY1oi4iIiIhzLODXiLaIiIiIiBwhFdoiIiLyt5JTE2nQoqbTMSTaWX9oTiGm1hERERE5LLfHzXfzXyMmxoO1ljsveJb501c4HUskYmhEW0RERA7LE+PG43EHL7814lEeefVG4uJjHUwl0cdG7Yi2Cm0RERE5rPzcAhbMXIG1NnjdWZeeyI8rBtK4TW0Hk4lEBrWOiIiIyN+647xnadC8Brc8ehHHndQUAONyM/C7PmBh84Yd9Oz2DDu3ZjicVCKa1tEWERGRkmjpvLXcf/kr9L/7PTJ2Z2EtGGMwLkOlamX54o8BVKiS7nRMkbCjQltERESOyI9f/8F5Te7mhfveJz+3INhSYrGkpafwzui+TNr4IV/PfIUr7jzb4bThx+Vycdfz13PLE1ficqsEC7I2sI52KE4hZg7uu/rXOxtz5HcWERGRqFapRjlOOqsN2zbtYuJ305m04UOMywQL8O2bdvHpW2MY++Vv5OUWOJrVE+PG77f4fX7Sypbi1r5XsG3jTjas2sL4YZMPuW9cQiwvfP4QsfExfPfRBE7pfgLlKqdTqnQSy+ev4bcxf2IwZGVkccblJxMXF8vWDdtYv3ILKxeu5Z4XbiA5LQmv17Jm6Qa2rNtOs+MbULpcWvAYC2cs5aFLng3lWzDTWts2lAc8Um2bV7MzRt8fkmO5qt8b0vdBhbaIiIgUiXue7cH5Pbrgcpm/3OYt9PHhgOH88OUUdm3/7/3cMbEeGraqzfoVW+h0Tlsuu+NsPG43WzZsx1vg4+NXvuW5ofcTGx+HtVBY4CU7M5c7zu7Lp3+8gjko46evjuSTV74NXn7560dp2q5+8PLBxY+14Pf58fstbo8bc9BL9fstYHG5XFgMlkDrsd/vx+V2YcyBDyFb1m/npk69//P7cBTCu9AedW9IjuWqcX9I3wdNhhQREZEi8eqjQ3j98aH0ful62nZqTPl9fdvGGDwxbm5+5BKuubcbF7a8h+zMXABOOrsNF910JkvmrqRu4xoMenYYNepVomHL2uRk5fHdkJ959PVbqde0OgV5hfS7823iEuJ46v1ehxS5+6VXTMMYQ5nyqcTuW4bQAG6Pi4SkOGrWq3xIkQ2QnJZ0yOUtG7YHC20baEgP3GAP/Mdi8fv9uA9uAbGWXdv2ULZSmX33MmTszmT5vNU0bluP+IQ4XG4XBfkF3Hf+k//hnZZIoUJbpISqXjEFv9+yYVtW8LqUxFgskJXj7Fe8IhK5/H7Li/d/AECNepV59LWbaNCiVvB2j8dNXHws2Zm5GGN4clBPAJq1qwfAwFF98RZ6iYn1UFjgpWaDKjRvXx9jDInJCVxzbzdWL910SJFtrcUcdIW1lu+HTuSSW88ivWIZCvIL+WPcLDau3sqs3xbx4r2DuK3vVezYsouZv8zng+eHHfIaXrrnPRbPWklsbAy/jp5BjfqVSS+fRv1Wtfnjx1lsWrONuPhYMvdk0bR9g30FdgY7tuxm64Yd1GxYldi4GFbMX4O30Fdcb7VEALWOiES5Ti2rMviRM9m4fS/L1u3hxvOaH/IH6vI+3/P1xGV0aFaZiQOvwBjD2KmrKJeWSPUKKbw0dBqvfTnLuRcgIhEvLT2FLue1o1Gr2owdNoVZkxcFb/t5w4d/uX9BfiGxcTEUFnh57bEh9OrfA0+Mh8L8Qp6+fSBL56/hwZdvomGbOiyetZLRn02i9YlN2Ll1Nz98/gu7t+/d18YhBwnv1pHve4XkWK6aD6hHW0SKRtXyyaz+5pZAv+C+n/X9oz77Lw8du4jrnvmBL/qdx8WnNtx3G4cU4zf1/4GPRi845Lkb1CjDzoxcduzJDcErEZFoVbdxdS657UwqVStHjXqVeb3Pp6xfsYncnAJ278ggKyPH6YjRQoU2oS+01ToiEqXuvrgVA+7qfMjkG6/Xj8fjwlrYk51HZnYht704HoDB382je+f6wfsf3HdYs3LqIc996WkN+PSp8zDGsCczj853fs6CFTtC9+JEJGqsWLSO5+5+z+kY4qT9y/tFIRXaIlGmbtU0vup3Lk1qlsXtMvj9fvIK/SxcvYMz7xlGRtbh+68nzFhL2mmvUeAN9BNef04z7r/qOP5cvJVnPvjjkPs2q10+sFmFMaSlxDPzo+swQG6+l5ue+4EpczewaXt2cb9UERGRsFZiWkcql0nEGMu7t3ekQ8Py5Ht9DP1lFR9OWMY5bavyy/zNzFq1C596uiSClU6JZ+3wm0mM8+Dz+yko9PP1xGXc9MI4fL6i+7cdG+Nm3cjbKJuWGLzO/L/p/z6/5ZdZ63hy8GSmzN1IemoCe7LyijSHiIgcsfBtHWlW1c4YcVdIjuWq+7B6tI9VfKybj+8+iXOOq8aerAKy8guZsmgrQ39ZyajHT8fjduF2BUbh7EEr9ewvD7LzvezJKmDVlr1c+uLPbM/IAwK9qkfxNomEXEK8h7Vf30SppDgA3C4X67bupc0NQ9mTlV9sxy1dKp6zOtTi+CaVuePi1mBM8OdpfwvK9j25PDV4Mm/cfxrGGPLyvXw4ej49XxoffJ6zTqxN6/oVySvwMnnuBqYt2FRsmUVESigV2qjQPmaXdqzFkHtOxv3/1sb0+i19P5tJ38taEet2HVhE/zCL6cOB4uDxoTOZNH8zvz1/DsYY8r1+1u/M5uEPpzFi6rrifjkiRywpIYbto28n1uPGGMOujFye/mgq746cR0GIl5VqVrccT9zYkc5tqpEQ6yE21s3oKSvJK/ByySkND+kXt9ZigfxCPwnxMcHnsEDVs99gy061noiIFKHwLrSH3xmSY7nqP6rJkEfj9OaV+O7R04iJ9WA4dPembRm5TJy/hVe/W4i1sHNvHtXKJZOV56V+pVJMXLCZ8qXi6XleE2I8hk07c2jXoDyFXj8/zd1Ir/Oa4nIFJoTFxbipWymVrx89I7giwwfjlnDLm5MPm0skVNZ+czOxHjcQKF5f+XImb3w925Es81ds5+JHDuyuVq50Ijv25FC5bAqnHVeT0inxwdv2/2wlxLmCP7tm39dHCXER/6tJREQkcke0K5VOYMKTZ9Kgyr7VEPb90S7w+hj04xIeHjKTvIL/NprXuk46f7x0fmCU3JhD1jszJjDJ7IQHvmPi8+fjckH3Z8bx45/r/9MxRY7G3h97kphwYDS4z+ApPPfJdAcT/bNq5VP49OnzqZieSK0qabjMgZ+t/EIvMW4381dup/XVHzgdVUQk2oTviHbTKnbGN3eE5Fiuho+rdeRIbPvoCsokxwYve32WJ76czYARC4plQmPVsok80L0la7dn0v+a44jxuBg3az2lSyXSrkF5rLX4/ZYxM9Zx58DJbNTX3lLMbr+gBW/cd0qwHeP+N37htWGRtbGMy2WIjXHj9frx+qJzaScRkTChQpvQF9oR9f1sYpyHCmnxrNmWRYwnMLrs81tuGjiFIZNWFuuxN+zIodd7vwPwyoj5weuv7FyXT3qfAgSKhrPb1aD31kwycgpIiHXz8IfTNJFSisULt3cCv8W64OtJyyKuyIbAVs15+V6nY4iIiNOidNW3iCm0z2lTlffuOJG0pFh6vT+NDg+NpmOj8nwxeTVZec79of5s0gp+nLWe7h1qcce5TWhYrTTxMW7uurw1AHed35Re7/zOe2MXO5ZRos+zt3QEC16fn9e/nM2Db//mdCQRERH5fyKi0F759kXUKJcMBFpEmtUozXvjl7FkY4bDyQJ27s1n8I9LGPzjEgAu6VSbm89qBECsx83bPTtxeuuqXPrs+H96GpEj0uuiljx4ZeBbr29/XaEiW0REIpslaneGdP37XZyTEONh18dXBItsgNEz19HnM2dWVDhSw35bRf8vZ+Lf9zWIMYYLT6jF6KfP4vzjazicTiLdgJ6dg5vDvPXtHGfDiIiIyN8K20K7UeVSZAy5gpSD1tddvGEPF704iYycw28hHU6eGDKThO6DGT9rAzn5hezOzKdrm+oM73Mmu4ddT8/zmjgdUSKUz+8PrkE9afYGp+OIiIj8RzYwoh2KU4iFbaE9++Xzg6N21mfpNXgqze4Z6XCqo+PzWc7qO4aUCz/gha/nYK3FYEhJiuXV2zuS+e2NTkeUCNT02o/54qeltLz+E6ejiIiIyD8I20L7wMbo0PO9qbw1dqmDWf67Ad/M5dY3fsVnD7STJMR5WPnRVZzRuqrD6SSSrNiQwdX9fmDh6l1ORxEREZF/ELaFdtd+49iZmc97E5bx7oRlTscpEu//uISE89/D57fBbahrVkjhsSvbOJxMRERExEF+G5pTiIVtoT1p0VYq3vQld7w31ekoRcrnt8SdO4ikboN5/svZbNyRxYtfhffkTgkPz916IltG3sqFJ9V1OoqIiIgcgYjdGVKkJImNcZEz/m6MMWTlFpDa9S2nI4mISGQJ350hG1e2Mz4Nzbw1V+tnQvo+hO2Itogc0KpeeQCsteTkFTqcRkRERI5ERGxYI1LSXXRyveD5d0bMczCJiIhIUXOmfzoUNKItEubcbkPPC1sCYC28850KbRERkUigEW2RMHf7+c2JjXEDUOj1sXVXjsOJREREilAUb8GuQlskzHVpXTUwlA0M/n6Bw2lERETkSKnQFglz559QG0NgIuTj709xOo6IiEjRi9IRbfVoi4Q9g7UWC+zNLnA6jIiIiBwhjWiLhDG3y+D1+YnxuFi5cbfTcURERIrBgR2zo41GtEXC2Lkn1CLG48IYQ2JcrNNxRERE5ChoRFskjN17cWsg0J89cOQcZ8OIiIgUhyhedUQj2iJh7M9lWwHwW8uQcUsdTiMiIiJHQyPaImHs9m7NAXAZw8btWQ6nERERKSYa0RaRUIuLcWOMcTqGiIiIHAMV2iJhKjkhhry8Qqy1rN6U4XQcEREROUpqHREJUw2rl8bjNvh9Pj7/Wf3ZIiISrSz4o3N5PxXaImFq/bZMPO7Al05VyiY5nEZERESOlgptkTB1W7emwfN7s7QjpIiIRCkt7ycioXbHec2D57dn5DiYRERERI6FRrRFwlRKYlzw/NvfLXAwiYiISDHTiLaIhFJMjCu4tN/uzHyH04iIiMjR0oi2SJjy+/wYtwt/lM7EFhERAcBG76ojGtEWCVNrtmRifZY1m/c6HUVERESOgUa0RcKU24DLZUiIczsdRUREpHipR1tEQik1KQ5rLQmx+jwsIiISifQXXCRMuVyBiZAYZ3OISPgpVbYUt75wO2nlU1k4ZSF7duzBW+ClIL+Qs286l9QKabgwZO/NJik1mRk/TueDRwbhLfQSExdDYX6h0y9B5FBROqKtQlskTK3anEGzWumsVo+2iADNOjXnhn43klI6hZi4mOD1VetVC573HTSfzBhDqfRUjDEcf04HlkxbxClXnkb1hjUAWLdkLc9f/Qx52Xkhew0iJY0KbZEw1bBaaTwuQ/m0BKejiEgxSkpL4vGhj5NaNpW5v85l56ZdnHPzuWxYtp6+F/bB2kD1fOf/epKYkhh83P7rD5ablUtCckLw9v33sX7LhmXrqdagenDZ0OoNa/DcDy/xzBVPsXPjjuJ+mSJ/L0JWHTHGvAScBxQAK4HrrbV7/ukxKrRFwlCj6mkkxAV+PKukJzmcRkSKUvlq5anXph7dbutGucrl8Oz7WTfG0LFbRwoLvLg9bqo3qkF65XR27CuCVy9YTePjG4OF7Ru3M+LN4Zx907kYY5g88jf27shgysgplK9egS2rNwePV6ZiGfJz88nOyGb4q8PofvfFuNwuLJBSNo3nx/8Pv8/Pbc2vO2zxLiJB44FHrLVeY8wLwCPAQ//0gLAttCukxpMcH8PKrZlORxEJKWNg9tuXYQALFBT6nI4kIkcgvVI6XS7tTJtTW7F17TY8sR4GP/4Be7bvCd7H7XHzzIhniImLwRPjwRhzYNR53wj07ImzaXVKa9YsWhMssgFeuO45ylcvz67Nu/AWegGYPGLyX3IcXGQD7NqyK3h+9KDvGT3oe5JSkzjt6jM4984LMcbgcrvodHFnfh02sSjfEpEjFwE92tbacQddnApc/G+PCctCu0qZRFa/fTEGWLw+g+a9RzodSSRkln5wJR63K/BVmrWc12eU05EkisXExdDvkwdwud30v+VVMnYV3eBGk+MbkrUnm7VL1h/+2LExJCTH4fdbDAav10tBXiHW7w+7jZqSUpMoU6E0cfFxlKlchpVzV9H5kpM59fIupO7rgz5YjYY18Pv9nHr5KXzzxvDg9fuLaes/UFz7/X6wMPqD0UwYOuGQwvj/27ZuW5G8nuyMbEa+9S3NOreiRpNaYGH9krVF8twiYa6sMebPgy4PstYOOobnuQH48t/uFJaFdnpyHC5jwG9pVDWVj+84kWsHTnE6lkixe/jy1tSulBoc5fp0wjJ+mr3R6VgSxV76+jHqNKsFwJCZb7J9005+/W4qn7/6LQV/szJFbFwMpculkrU3B1+hj2sevoTTLz2J7IxcSpVNZeOqLayYv4ou3TsC8OGznzNy8A+HPEeDVnXp/8WjeGICf4a8hV58Xl9gkp/Lhc/rZ9umHYwd8hO/fDuZ1ic3Z8v67ZQul0pauTRWL1pLbmYOa5duCD5nqfRSuFyuQ0aQj1RqeinOuu5M4uPjSC6dTNtTW5OQnID1W95+5D2u79OD2KTAZb/fT0FeQbAXGjhkZBoCBXRhfiFzfp17yHH8Pj+Pnv8oNRrVYMHvCyhdvjRb12496rxF5ZlL+lK6QhmM27Br007HcoiE0A5rbdu/u9EYMwGoeJibHrPWjtx3n8cAL/Dpvx0sLAvteet2k5NXSGKsBwNc2ak2T349l9XbspyOJlKsKpYOTHSy1pKVW8i1A35yOJFEO7fnwIZIBihXOZ3zrz+D7Rt3MmboX//9uVyGt8Y/R7nK6RiXOWQkNzY+HoBqdSsTF39gVYw2nVv8tdBuXReX+8BWDsblwhNjMK5A77Db7aJitfJc+8jlnHllF9IrlsEd4wFrD8k846fZ1G9Zh9QypbDGYIDxX01k8+ptJKYksGz2Cm7s2wNPXAwutwtvvhe/z0e5quXw+/zc2rEXe7bv4eUfnieldMqBPPtel3Ebut1ybvADAYbgkpvW2uD9rLXkZucxcdgkdm/dzYTPfiI/N/+w7/nOTTvZua+odbLI3m/31sAIevtzO9DjqRuZ/+tc3r33DYdTSYliCZvWEWvtaf90uzHmOuBc4FR7BJMawrLQBki99jO2DrqU0slxWAvzBpzP2c9O4LclRfO1mUg4evSDqSxZv5ulG/YwcY5GsqX43X12X4bMfJ3U0qWwEFxfeeWiw7cReGI8lK1UBpfLddg13gMdT5bHr3qeASOewlvoZcDdb/3lfuO/nES1epVpcUJTdu3YzbzJC8nNzueSO88nPjmB/X+9/D4/udmBgtX6Lf+vQ4OWHZsSGx+77+CBOQ6nXXIy+XlewBBzm+eQDwP7R5739yW36dKCn776BU9szCHPe3Df9Jf/+xqX20XDNg3YsXUXu7bsYvH0JVSsWYHyVcszY9wMsvfm/PMbHQHOu6M78YnxtD2zHbPOPp4ZY6Y6HUkkrBhjugIPAidba4/oh94czQxjY0zIm+ZuPq0er1/XDrfLRaHPzzfT1nLNm3+d/CEiIseucq2K7N2VSUFeAZ5YDzmZuX9731admtLutFaUrVSGRm3qERsfw47NOxn8zOdsWrWFret3/KfVKxKS4yldLg23x0O5Kuks+GMRtZrUZM/2PdRsVJ3ElESatG9A1TpVmDbuTy69u/u+YjtQUC+asYTqDWrgiXGTlZFNmQqlg8/t9/nZtn47FWtWIHtvDrd16kVedh4ValSga4/Tyc3MYdLw38jYtZf87MOPSEerc27vxgV3HZjb9WCXu9m9dbeDiaSIzfynlgknta1b1k4f0C0kx3J3/+CY3wdjzAogDtjfZzXVWnvbPz4m3AttgCcuacGD5zclLsYdmEQCNLt/JEs3aSMPERH5q9j42ECrSEEhNRpWJy8nn23rt1NYoB0R/8nA2R8EN8NZ8Nt8Xrv1RYcTSRFSoc1/K7SPhevf7+K8p4bNpd/Xc4Nf9xlgwcvd+ODWE5yOJiIiYaggr4C87Dy8hT5Wzl/NxpWbVGQfgWcu6xNol8XQuGMz4pPjnY4kJYXfH5pTiEVEoQ3w/MgFjJq1Lvh1pAEu61CT89tU47OenWhQqZSzAUVERCLcpmUbyc8rBALfHtc/rpHTkUQiWsQU2gDdX/qFqrcNY9bKnRQU+Hh1zCK+uLsTFx9fg9eubed0PBERkYg3+ZtJ7O8qveWVu50NIyWDBeuzITmFWkQV2gBbM/Jo//gYEq/9jMe+msOvi7cFFvqfHVhLtWJaAue1qUpcTMS9NBEREcdNHfkbEFihxRProU3X9g4nEolcYbu835Hq+vwEEmLd5BYEtqme1u9sKqXFY4whI6eA8wdM5Pdl2x1OKSIiEhnWLVyD3+cPrnPe7pwOzBw7zeFUEvXCbDfYohIVw777i2yAWI8rMGHSGFITY/ml75mMfeRUB9OJiIhEDmstg3of2LAmPjnRwTQikS0qCu2Dte8zhk27cw5MmjSGUxpXol3tMmR9eAV5H1/JzafUpWJqwl82Poh0pzQoT7cWlWmsiaEiIvIfzB0/k8ydezHGUP+4hrgO2o1TpMhZC74QnUIs4ltH/r91O7Kpcddw2tVJ57krWlGrXAof/bKSHp3qEOsOjHa/dk07Xr8G1u7IYsOuHBZvzOCVMYtZvd3ZLd4vaF0Zaw2PndOESqUSeHj4HD6demB3tl6n1qdRpVI8PHwue3IK+aHXyXSqVw6AXVn5VEg5dBmm3AIvw2Zu4OahM0L6OkREJLJZa5n/6xxO6H4SWKjXpgFLpy1yOpZIxIm6Qnu/6St3cuozE4KXq5dN5JpOdUiIdbMnO5/0lHhql0+hdvkUTmpYgZu71MPlCmzNO2/Nbn5atIWFG/awYXcOhV5Ldn4ha3dksyu74KhyJMa6yS30kZ4Uy46swGNdxvDSZS05vXFFqpROICkuBq/XT4zHBRYwgeULX72sdbDQblAhhX4XNMdtDGt35vDaT0s5tVHF4HEqlEoI7oa8f73xhFgP13SoyWs/L2PBpoxjeh/vO7UBbauX5vdVO1i3O4fRCzbji9I+KhEROWDiZ+M5oftJGJfhxIs7q9CWYmMBG6W1RdQW2v/fuh05pN70BQCNKqdy95kNOalRBepVTAHA5QpshGOMoUX10rSoXpoCb2Bhc2MC32rszS2k5n3fUuD1c0vnury5b0nBQq8fv8/Pz4u3sGpbFld1qElSnAdrAz3j+63clkWjR0dxdvPK3NGlHq59vSvGGGL29ZYHVi4N/KN7f/LK4GM37Mlle2Ye5VPi+X3lDnIKfKzankWtskn4/Zate/NIT4ol1u06ZOtjr9+yZmf2Mb1n9csn88Q5TYjzuLiwVdXg9T6/5ZWfljJgwlIycrUBhIhINMrZkxk8X75aeQeTiESuElNoH2zxpgxu/zAwg7pmuSR6n92Es1tVpmqZpEPuZ0xgEqwLA8aSkhBDrMdFgdfPw+c13XefQJGMy3Bm08qAxe0KFNf7R5b3q1UuGYAlmzMo9Fvi3GbfWqWW31ds57ha6azYlslbE5bz/byNbMnICz42O99LvcdGE+txkVcYmPzZsM/ov32NNdOTSE3wsGpHNln53mN6nzbsyWXL3lxq7Htf9r8WtwvuO60BlVITuEltKSIiUWnnpp389PEPNDi+CR8+9I7TcSSaWRzpnw6FElloH2zN9mx6fjwdPj5w3SmNK1I5LYE/V++kdFIsMW4XLaqXZubqXWTlBYrWx4bN4uNbO2KtxW8hO7eQob+v5poTa5MUd6C43j+6nJVXyFMj5wOwYlsWFXsNJyHWzc6sI29F8VsbLLL/9XUd4yj2wXIKfDR86geql0lkXM+TqJEe+KBgLRT6LKt2HFlPu8vA4KvackKtdL6atZ5Ppq1lR04Be3I0Gi4iEs6ys3JJr1qetueewOi3hjsdRyTimIPbDP71zsZE58eNIhQf4ybGbcgt9FEjPYmV25ydYFkcSsV7qJmexLyNR9b33a5GGSbd0xnX/1vlZc3OHC7/8A9OrJlOgd/PB1PXqv9bRCSMDFwwJNDWaC0PdryDrIPaSSTizLTWtnU6xOG0qZVupz15VkiOFXPdpyF9H0r8iHZRyyv0kbdvoDYai2yAvXneIy6yARZt2UtmfiGp8TGHXF+9TAKT7upMfFxg2ajXLmrF3txCen09my/nbCzSzCIicmz2D8h1vup0RmlUW4qDBXx+p1MUCxXaUuyy8r1UfnQUF7WsQquqaZzRqAJ1y6WwflcOVUodWJLQZSAtIYaPe7TDZ6fTtHIq09fu4ofFWziKL15ERKQIxCbEkZ+TR1xiPD6vjxlj/nA6kkjEUaEtIeHzW76atYGvZm3gke8WBK+vUzaJty5tRf3yKVQ+aB3wNy9pRVpCYATcby3nvjuFKat34nZBTkF0fuoVEQknz018nbjEwO/l9YvWsHX1ZocTSfSyWt5PpDis3JFN14GTAaiVnsjwGzswZ0MG9con07ZaaQDcLhcPnFKPMfU7ArBlbx7tX/mZzHwvOQVHNjlURESOXPmalUhITtzXnw3j3h/ldCSRiKRCW8LG6p05tHrxJwDiPC5u7lCLB05tgPVbpq3dQ5f6FTDGULFUPCv7nkVmvpdmz41jx1FuIiQiIv/s4oev3re8rcXn8zPnpz+djiTRTMv7iYRWvtfPm7+t5M3fApv2uF2Ga9vXoFKpeHzWYq0lIcZN5dQEFdoiIkVs2fRFNOzQjC1rt9D/goecjiMSsVRoS0Tw+S21nvqB5DgPKXFubutYh/W7c5h3jFvLi4jI35vwwWj+GP4LORnZHM0ywCLHTD3aIs7LyveSle/liTGLnI4iIhLVsvdE5xK1IqGkQltEREREnGPBRmmPtsvpACJOM+bf7yMiIiJytDSiLSVWWkIMKx7vSlKsmx8Wb+G6z/5kb57X6VgiIiIljAV/dO6RoUJbSqzjqpcmKdaNMYaujSsy9raOlEmI4Zu5G3lv6hrW7MpxOqKIiIhEMBXaUmJNWrGd6Wt30bZ6GXzW0qpqGgbDfV3qc1+X+jz942JW7Mhi2JyNTkcVERGJXlpHWyT6FPosJ7/5KylxHqqVTmD87Z0okxgbvL3vmY0wxvDgKXsY8PMyhs/fRGGU/iIQERGRoqfJkFLiZeZ7WbQlk6pPjqHLm7+weNteZm/YE7y9aYVSfHR5W9Y+1pUHOtfD7dLsSRERkaJk/TYkp1DTiLbIPtbC1LW7af3SzwAM6NaMi5pXoXxiHC4DpRNjeeTUBpxUuywn1Eznoxlruf/7+Q6nFhERkXClEW2Rv9F75Hxq9RvLRR9P5fc1O8kt9GGM4bT65UmMdXPHibXZ0OcsLmxW2emoIiIiEoY0oi3yL8Yu2crYJVtJjY/B4zKseuxMYt2Bz6jpSbEMvfI4jIG9OYXcNGwW3y/e4nBiERGRCBLFkyE1oi1yhDLyCtmZU0Da499z33dzmbZ2F34b2PDGGEOpxBi+urodr53fzOmoIiIiEgY0oi1ylKyFt39fw9u/r+GM+uX56LI2lE6KhX2TLDrWLEtSrJsF951KhZR4Zm/cw7kf/sHu3EKHk4tISVW6YjruGDc71m9zOorIYdioHdFWoS3yH4xbto3K/X6gbplEbu1Qi/JJ8bz063I61y5LxZR4jDG0rpLGz7d0pNVrE/G4DVVS4lm7J9fp6CJSQjTr0pqbX+2Fz+fn9RueZfWc5U5HEikxVGiLFIEVu3J4YPTC4OV1u3PILvCSFBv4EbOAy8CWx84iMcbN7rxCvp63kb7jF5Ohbd9FpJiceVt3zr3zQgA8LhepZdOcDSRyOBZHlt4LBRXaIsVgb76Xsk+NoXWlVFpUSWXU4i0kxLhJjHEDUDo+hpvb1eSiZlW45etZNCiXxP+mrHY4tYhEm/1FNoDf72fuT386mEak5FGhLVKMZm3OYNbmjODlAb+t4JbjalIqPvCjlxLr4esrjwOg7ykNqfnSeI1wi0iReHj4c4dcfufOV7A2OkcNJQr4/E4nKBYqtEVCqO+4xfQdt5g2lVM5s0F5jIXHOtfHGEOc28XK+08j3uPig5nreGjsInK90fmLR0SKV1LpFCrWrYofA9by8hV9WbdQ35qJhJoKbREHzNyUwcxNgZHuFhVLcXb9CmQXekmJiwHgprY1uKltDQZOW02/n5eRka9RbhE5ctm7M9m6ZjMValZi2bRFKrIlrNko7tE2R/M1kjEmOt8FkTDgccGiXqdQNTUBCKzNvf/nc2d2Phd/NoNpGzP+6SlERET+zkxrbVunQxxO68ppdsotHUNyrMSnRof0fdCGNSJhwuuH+v/7mfbv/MrktTvx7utXM8aQnhTHzzeeyLdXhOXvSBERkf9g3zraoTiFmAptkTAzf0smZ3w4lRMG/UaB1x8Y1d73u+GMeuWZdftJvH2edp8UkcOr1boBXe+4kMTUZKejiJR46tEWCVPzt2SS9swPpMfH8Oo5TTm/UUVyC300KJdMg/RkujeoyGfzN3Lfj4ucjioiYaJyw2r0+uhxAJp1bsNLlz7mcCKRI2AJ7q4cbTSiLRLmduYV0uOb2aQ+8wNvT18THN0uFR/DrW1r8MZZTRzNJyLh4+4P+gTPlyqX6mASEQEV2iIR5amJy6j36s/k5HuDEyVvaF2dnQ+d4XAyEQkH8ckJGGMA+Oj+1x1OI3LkrM+G5BRqKrRFIsymzDwqvDSOFTuzwIKxEO92c3qtsk5HExEHxcTHAmCtxVfoY+WsZQ4nEhEV2iIRyA+0eOc31mXkBDpJrKVBWU18EinJCvMKWDZ1AYX5hXxw32tOxxERNBlSJKI1GfgLb53VlLKJsXy2YKPTcUTEQXWPa0Tt1g3wFnhZt2CV03FEjlwUT4ZUoS0SwfwWbh+zwOkYIhIGKtSshMvtIiYuhlJlU8ncqQ2uRJymQltERCQK1GrTEONyY1zgj9LRQYlWFvZt0hZt1KMtIiISBRqf1BJL4Fv4esc1djqOiKBCW0REJCpk78kKnm9z3gkOJhE5Shas34bkFGoqtEVERKLAWzf2D56vUKuKg0lEZD8V2iIiIlFgz+adzP/pT6zfsnLmEqfjiBwdnw3NKcQ0GVIkCnlchvMbVGDSmh3syvU6HUdEQuTD+14jtUIZ9m7b7XQUEUGFtkhUmndrJ2qkJQJw5idTmbxxj7OBRCRkMrbu4vrX7qVO6wYMffgtlkyZ73QkkX9mwUbnoiNqHRGJRpVTEoLnx159PMbBLCISWunVytOsc2sSSyVx81sPEpec8O8PEpFioUJbJArd8+OBTWwMUC1Vf2hFSordm3di97WiWuCGN+7HuPRxW8Kb9ZuQnEJNhbYUuwvqVeC8OuWcjlGifDR3Iz6vH/zgs5b1GblORxKREPF7fbxz2wsUen1YoFbL+vT+9gWMUbEtEmrG2iOfgWmM0VZTcsTuPa4GT53cANe+kZRnJq/g+d9XOpyqZIl1GwocmGUtIs6rWLcq93zeD3eMB2MMW1dvYkD3hziav/sSVWZaa9s6HeJwWpUrZSdd1D4kx0p7d0JI3weNaEuRi/e4WHBTJ/qd3ACXMcH+4E7VSjuaqyRSkS1Scm1ZsYFxbw8PXi5fsxK9hz/vYCKRkkeFthSJikmxfHp+C06pUYZaqQlUSYkL3GAtFvD5Ld2G/eloRhGRkubnD77n2+c+Du6IV75WZV6aM4SabRs6nEzkUNaakJxCTcv7yX/21Il1uL99bTDQrX4FKrz2E2/OXMvxVUrz/px1fLl4CxpXFRFxxu9fTiBz516ufvFOXK7A+NoNr95L3463OpxMJPqp0Jb/5KfLj6N9pVQMBGa5m8B/+/y63OloIiKyz/wJ0/mkt4/rXrkHgK0rN9Lz06dZ/scCfnzzK2fDiRC962ir0JZjNv3q9jQul4IxBovFWrhl9AJyvD6no4mIyP+z8OeZPNr+BirUrcaJV5xBq3NOpGrjWhx/ySn88vFoJn3wvdMRRaKOerTlmPRuW/NAkW0tm/bmkfK/CXy+eLPT0URE5G8U5heyYeEq/hz5K7l7swFITE2m612Xclz3zo5mE4lGKrTlmLhdBuu3WGuZuSWD+oMnOx1JRESO0MoZi3jq5NvIz8nDWosxhuMuONnpWFJCWRuazWqc2LBGrSNyTF6cvpocr489eYUMWaRRbBGRSPRkx1s4+74rqNG8HqNf+Yx6HZpy3eu9McbwzVODmfn9b05HFIloKrTlmFjgjVnrnI4hUiJUrN6emnVOYt70j8jJ3u50HIki1lpGv/xZ8PLN7z6C2+MG4JKnbyFr916WTp7rVDwpQfxROhlSrSPyr1Ji3Azv1pJ72tRwOopIiXNct/4cf8r9lKl3PF0uft3pOBLlxr3zzSGXqzSs6UwQkSihEW35R/EeFxtv74wxcHrNdH7bsJuZW/c6HUukZDAu6npq4NmdR2Gsm8y0WKcTSZRbO3sZz555N90fv4Fd67cy+bMfnY4kJYQT/dOhoEJb/tHy6ztiDvq3vz4zz7kwIiVMw9PuPOiSZeqopxzLIiXH3m27+fjul52OIRIVVGjL32peNpm0OA/GD9bAwDlr2ZZT4HQskRLDuFzsqJREXG4hU8Y8xe7NC52OJCJS9Kw2rJESaPmeHHbkFpKeEMMjvy7jrbkbnI4kUiIYl4c2NwwkLiWdNfPHsWvlDBXZIiIRSIW2/EW3OuV4+eQGDF20mVrva2knkVCr2eV64stWwVhIKF+DnRPedjqSiEixsQTW0o5GWnVEDpHocTH0rGZUSIzj/jY1qJIc53QkkZLFuKh8fHes2+B3GfzGOp1IRESOkUa05RB3tay+71zgj3uV5Dg2ZuU7F0ikhKna8TJg30+gC1b+qNFsEYl+0dqjrRFtOcQDbWoGzw9dvInpW7SUn0golWt5JtYYMIZdK2aQu0MbQ0l0aXp6e2798AlqtWnodBSRYqdCW4JKx3mIdbuw/sAny9t/XuJ0JJESZ93Ej7H48Vsf2+f/5HQckSJ3+XM9qd6iHje99zhPTfuI+JREpyOJ0yz4/SYkp1BToS1BTdKSAQMYfl670+k4IiVScrXG+N1uCgtz2bnoV6fjiBS5vMxsAIwxeGI8PDrhbYyJzolwIiq0JejpE+sGR7Nj3W6n44iUSOlNT8LvMrgTUzAe7QQp0eeFs+5m6ZQ5WBuYC+SOcXPhkzc7nEqctr/+KO5TqKnQliCPMez7vccrs9Y4mkWkpPLHesDjwocP69UGURJ9CvMK+OSuAaydtzx4Xc02jUlITXYwlUjxUKEtQU3KpADgt/DThl0OpxEpmTyJpQAwMTEOJxEpXoOue4o5Y6bgt5BWqSx3fNbf6UgiRU6FtgDQqHQSLlegR87a/Yv7iUio+YMn/RRK9PvuuY+AQL92aoUyzoYRx1gL1m9Ccgo1FdoCwI2Nq8K+AvvByVptRMQpa755idzt61j24SNORxEpdvnZuRTmF2CtxRhDSnqq05FEipQ2rBEAlu4OzAK3fkt2YZSuGi8SAXbNGseuWeOcjiESMltXrKdqkzoAJKQmk7kzw+FE4gRtWCNRLd4T+KdgjMGtVZZERCREfnr7a3yFXvbu2M2ezTucjiNSpDSiLQDc0rgqgTW0LYV+9YaKiEholK9dlcK8fH4fMoaC3Hyn44gjDNZG5yifRrQFgMopCfvOGTLyCx3NIlKSlTv+HMq2P9vpGCIh07HH2SSUSqbDlV2djiJS5DSiLQB4/RDrNlhrtbSfiEMqnHwx1c+9FQCXO4Ztv490OJFI8YpLSiAhLRlrLQsnTHc6jjjFgl892hLNtucGZn37LRRatY6IOKFa1x4YYwFL2banOR1HpNjd9EFfPPvWjC9Xq7LDaUSKnka0BYDZ2/eSHp/OLrWNiDjGEx+LyxX4oFuwd5vDaUSKl8vjpnydauyfH5S1e6/TkcQhluhddUSFtgBw+6QFnF2jHL9v3uN0FJESy5+fjSshsA31qs9fcjiNSPFqfOpxWPZtzmRhZL/3nY4k8o+MMf2AbgT2FdsGXGet3fRPj1HriACQVejjqxVb2JCd53QUkRIrc9V8jPFTmLkDf75+FiV6lapQhvMevREIjGSOev4jfIVeh1OJYyJnZ8iXrLXNrbUtgVFA3397gAptAaBjpdIMPb0F7StoVy4Rp7hiPOD3YrT9ukS5Bie1xu3x4M0vYMRT7zLj6wlORxL5V9bag/ubkuDff1mrdUQAeKdzUyonxdMgLYnjhv3udByREsnljgEMa0Z97HQUkWLT8ryTOPOeKzEuw4xhPzF/rP7mSEh7tMsaY/486PIga+2gI32wMaY/cA2QAXT5t/trRFsAWLE7B+uH1Rm5TkcRKZHcCUmk1muO8Xgo06Sd03FEikX5utU495Hr8cTGYv2WKZ+Mwuf1OR1LSpYd1tq2B50OKbKNMROMMQsOc+oGYK19zFpbDfgU6PlvB9OItgBwYqUygKFLlbJORxEpkXy52Vi/D+P2EJOsFi6JPjWOa8K1bz2My2WwfssPAz4hc8cep2NJmPD/9/7pImGtPdK1VT8FxgBP/NOdNKItALjMgf+Gxz91kZInZ+sGrM9L5pqlTkcRKVJ1TmjBNQMfwZrADITsjCxmjZzkdCyRo2KMqXfQxW7Akn97jEa0BYBCn8XjBiw0Lp3Mwt1ZTkcSKVGSa9QnpWoNAOLLlHM4jUjRaXfFWZx531XBy34Ln979ItavSb8ScZ43xjQgsLzfWuC2f3uARrQFgPt/X4jPb8n3+bVpjYgDyrY8AWPAGEOFdic7HUfkP0uvWZk2F5/OCdeex/7vSn1eL+9e+QibF692NpyEFWsDkyFDcfpvOe1F1tqm+5b4O89au/HfHqMRbQHg0+WbmbMzkz35Xjbn5DsdR6TE2TZjIjW6Xsq+PdJwxcTiLyxwOpbIMbvuvSeISYhjx5pNeGJj2DBnKV89+Ap+b5RuAShyGCq0JWjhLrWLiDglZ9NasjauIrlKLbA+FdoSkao2q8cF/e4gITWFvMxsPHGxbFywgsFXP+p0NAlz1kbnDDEV2iIiYWL2C/dT+aSuZK1fhTdHH3wlctQ/uQ1n9r6W5PRUXG43fp+fWZ98z/o5S1k/b5nT8UQco0JbRCRM+Avz2fTLGJrc9AC1zr2C+e88q4JbwlqpCmU5u8/N1DmuSfA6v8/P+jlLmDX8J3Iz9O9XjkwIN6wJKRXaIiJhJK1Bc8q16gDG0P7JgUx58BqnI4n8RUKpZK77qB+plQ9dIcfv8/P1w6+y/NdZDiUTCS8qtEVEwsje1UvBBHoV49LK0P7pd5jW919XkBIJicpN6/J/7d15dFXlvcbx77vPkIlMJCGBhCGQhMioDFVQZPAySF1UsdZVlx1QO9321tpir67rbette716a7WKva12cLZqi7aIA4oDisgkIIRRZswcMicnZ9jv/eNAkIICas4J4fmslZXknH2yf3uvrJNnv/nt953543nkDC7A8Xo6H3ct1O8t54/zfkKoLRDHCuW0ZMF2kwVrPmsK2iIi3Ui4rYWq1W+SO34SxkCvvgUMu+YHbP7T3fEuTc5gWYX5XPbf3ye7MP/oJ6ylfPNuFv/8d9TtKY9PcSLdmIK2iEg3U/bA7WSUjCAxIxOAfhOnseWhe7GRSJwrkzNJUkYqM2+6jn7DB5OWE/1dNMZgraW5uo4nr7+D2t0nnEZY5IQs4KpHW0REYmX5jV9h4v/8kaSsHILNjdie+ldIugXjGBLSUsgp7M/E6y6jYFQJjjcaETzmSMAOBTrY+MJyltzxJ63sKHISFLRFRLqpt2+6lrRBRbRVVUSXThP5lDw+Lwm9ksEYSqaMZ+fydbjhCF979DaS03thXRfH4+kM1gARCy0VNWxcvIzlf34WG9FFn3zGLLiRnvkep6AtxzV/9GC+M3wgKyrrufrV9fEuR+SM1bTn/XiXIKe5CddcxvDZF7JrxTqGnD+GXtmZNFXWkpqbxTlfmsmrv34YX2JCNFwfWirdWot1XVrrGln5yCLefXpJnI9C5PSkoC3HNX/0EDyOYUZBDjefPYTb1u+Md0kiZ7yEjCxGfutG2msr2fzne9ROIsfVd2Qx02/6BqG2Nnav2MDE6y4H4Jy503EPtXs4HgesJdDUwr61Zaxf+AqpfXpTsWkHu97egDFwcF9lPA9DzjA99e1MQVuOq6y+iZGZ6YDh+hGDFbRFuoG+E6eSXnQWGcXDyB45hnd+9gM6Gg7GuyyJo2Gfn0TeWUPIKMhj1cN/58C7W7hiwS14fNGp93yJ/s5trbUsvfNBeg/sx+pHn6P3wH5Ubt6Jjbi8ce/j8ToEkR5NQVuO66JFK6n6ygyMMRgD148YxG827Yl3WSJntJr1Kyn8/JfwJSeRkJ7B5LseZNn8awnU1cS7NImR0tkXMuHay3FDITIKcjHmyNzD+WeXcu+UrxNobiUlMw0MrHnseRLTkumV05sVf1pIqD3YuX1rXUMcjkDkWNZCpIfeXKugLR9p48EmRvZOA2u4+ewSllfUs6auMd5liZyxWsv3s/6eWxl/020AGGDynX8g2NTAml/9lOZ9e+Jan3x6xuN03mzYb3QpHr+X/as3MfXGayia+jkS03oBR88EclhjeTUAD175I/qNGkr5ph0Em1tjfxAi0snYU7iT3RjTMy835COt+sIFDExNxgI7GluZtGh5vEsSOeP1nTCFEdd+H8frPbSIpMVxINwR4K1bbqCtUguHnC48fh/nXjuXkukTSe2TBUBl2Q7euOsRLr/vFrDwwn/+hjm/+jFAZ7B2AKxLW30TlWXvs+6vS/hg/VbNCCIfZ621dly8izieYSkZ9tERF8ZkX2NXLYrpedCItnysL76yltWXTcIAfZMS4l2OiAAVK16nYsXrDPvatxkw7WIcJ7pquzchgSn/+1tW3f4zajetj3eZcgLepES+veSBzvaPw5/zhhUfGak2EA6GWfXgMxRN+RzvL1tD/qgS3rrvCao2694Z6TncHroel4K2fKx9re20hSIkeT3UBoInfoGIxMzmh37HniXPMfo7N5BZOCT6oLX0Lh1G7pix5E+4gPX3L6B63dr4FirHlVMy8Kge68Pheu1j/6Bm227+9t2f43i9VG7awYG1ZbzzwF/jVaqIfEIK2nJCUxevYFb/HJ7Zo6meRLqbtooDrPjJj0jMymHgRbPwp6Wx+8VFzPr9QxjHYcjsOQra3VTFhm1U79hDVmEBm/7xGmXPLKVu94HO56u37o5jdSKxYy2dU0/2NArackJ7Wtr43Za98S5DRD5GoK6GbU890vn9tr/9hYLzJ7N94dPH3d6fmorHn0B7XW2sSuzRHJ+H6T/9Hhn98yjfsI3SmRdQ9o9Xefu+j58278l5t8SoQhGJBwVtOWnpfi9D03uxpraBHnrhKdJjbF/4FNsXPnXc5xIzezP1zgU4Hg8r7/gFtWUbY1xdz5A5KJ8Rl8+g9OIL8Sb4ONwEkjW4P8YYhs+ZdsKgLSJR6tGWM5rHGLZeMQ2DoT0cpvDJpfEuSUQ+IX9aOo4nuqBJcm4elG1k5LxvMGDyVBr27mHl7b8k3KZp4T6s79mljLl6DslZmXQ0t5I7vAhjDI7HwTjOoa2iIxDB1naayqvZuPDl+BUsIt2CgraclN4JPgzRxWuSvF4m5fXmzUqtSCdyOmrau5v1v19AQnoG+994FX9qKoNnXowxkFNayiV/foT2uhpenX8Doba2Y15vPB5s5DQafjKG3BElZAzKZ+9bq8k5q5iGPQdoqaqj/7mj6H/e2RROGkfjgQowDk0fVLL8roc45+o5jLj0Ipoqa8kpGfRPP9JgXZdIOIIhguP1EA6F2PLc67y94DHc8Gl0fkTiTD3acsarCQRpC4VJ9kV/ZZ6cNp75Kzby+G7N1ytyOvrg7Tc7vw61tkb/0n1oBozk7GwKp89k+9+fOep1xV+Yw6ivfpX9y5ez6td3kVlcQuncy2mrq+Hgtm0MnjmTpgP72ffGMgL19bQ31ONxPHgSEwi1tpE5eBDB9gCR9nZaq6ILrDg+H6OuvhJjLB+sXkf+uWPZ8cLLtJRXUXDuOYTaAtRu3YF1XUZ8+VL8SUlseeZ5krN702f4UHrl9eG9J56lva6elD7ZFE2fRMnsKfTKzcENh3E84HijI/jn33BNNCRHImx/aRnFMy/Ae+h9Lal3OgB5I4qp3baHUV+ciT8lCV9K0lHnINTaTv2+crYsfp2mA1WUr9+C7aEhQUQ+HQVtOWmDn1rKreeU8M2zCnGM4c4JI9nXGuCtao1si5zOrOuyaN5XmH73ApIyo2HTjUSo3brlmG0LzjsPYwx9x44F4NwfzicpKwtjoOjiWQDkDCulaOb0ztcYInDoP2IfDvObn17Ixkef4LKH7ychNQWAsy6fgzEwbO4l1O/cRVZxIQAHd+7B6/eRMTAfgGGXXYwbCePx+wEYMHEcT1/1Xab+5HqySwoxjhNt7fB6cZwjIbhzOj3j4PX7OdzucdT5sJaabbt45/d/YeTcGWx9YRkjr7gY4xg2PLGYDX9Z/AnOsoh8HLeHrrWkoC2n5L/W7+DaoYPwONEBsN9OHM2lS1eyq/nYfy+LyOnDDbTx0reviX7jODiOgxsOH7PdugceYNiVV7L7leh9GnVbt1Bw/gVYN4IbCuFJOHphKwNgnA/n6055Z49i46NP4P+nEePDeuXldH6dkpMNHPlLbBwDH+rOCHdE5/lv+qCCrCEDjwr0gaYWQu0BOppa2LFkGfljRrDrtXfY/cZK9r29lrSCXPLHjmDr4tep2byTtoMNhNs7qC57n7JnXgFg/ePPfdSpExH5SFqCXU7ZqMw0npt+Hh7HEHYtLpb/27yb32zeRUdPvSQVkY+UlJ2NLymJ1qoqckaNZPiVV5I+cABYl9rNWwi2NBIJhnE8DjnDhxEJBgm2tLLyrntoqayi77gxjPvWPCrf20gk2EHRjKnUbt3Bmvsf4rx/u45gczMbHvkrwdZWpt16I47Py+rfPkS4I0j6wH74k5PZvngpobZ2jGPIHNSfhv3l2EhELR0iR3TbJdhLE9Pt/YMuiMm+Jm97PqbnQUFbPhGvMZzXJ5PHJo/D4xhcFw60tHPT2jKWVdXFuzwRERE5moI2sQ/azok3ETlW2FreqjrI15a9S0VrAINhQEoyj184nl1zZ5Dj88W7RBERETkNWCDixuYj1hS05VN5vbKWL7yyki0NzYceMfg8DqvnTOOecaPJTvDHtT4RERGReFHQlk+tvD3ArCXL2d7YgrUWLDg4XDqgH6tmTyP50MIYIiIiIsew4EZsTD5iTUFbPhMuMG3JW3zj7Xd5t7aew63/HhNtKbl5+FDemD6ZH5YWk+jo105ERER6Pk3vJ5+pF8urebG8mrGZ6fxyzAheqaimvL2dbxUPxgDXlxZxbdEgzn3xNVqOM3VYvDjAvNwi/qV3P5bWV/BWYzW7As3cUTiOkoQ0GsIhNrYeJMeXyLMH97GsuQqAkUmZTErNpcCXTIZJINeXzIKaMl5rqYzvAYmIiEjcadYRiYkHJ4xjam50TtywtUxf+iZz++eTl5TIre9tpjlGodshOvo+tXceTeEgG5rrWTpmJr0cH65ro0tXWGiPRNjQcpDxvbLBRucMj35EpzO8dNurtERCvDB0Bj4THaG3FiIRi7XQEXYJR1waw0Furl7FgUh7TI5PRETkI3TbWUdKEtLtff3Oj8m+Zux5IabnQSPaEhNfX7GG0tRe/OvQIbxTe5BUn5dvFhdigP2tbVw1cAB+x2H+uxtYWl3zqfblM4Yhyalk+xOZ1acvQ5MzKExOIWzBbx3qQ0F6ebwYDD/etoY0r+9Qq4vBYDvXiasOHr0Iz+FnOlyXdjdMxFoaI0GyvYnH1OB3HEzYkOUksSD3Ai4tf5k8J4krUgazuG0fuyLNx7xGREREehYFbYmZrc0tfH/NBgDSfV7qgyEyfD5m98ujT2I0rP7hvPFUtgUoa2zinZo6hqdn8FJlBQ2hEO8cPDI/t884zC8q5ZK++WxqauQHG9eSl5DIwvGTSTx086V1ARtdHe5wkHZdS4b1E7HROX72BFqoDwXJ9PkJE6Yq0MHd+8rYF2pjV6CZ91oaGOBPYW+whc1tjaR7/OwINNLuRpeku3rnMgb5UyhNzCDN+Jid3p+dgSZGJmbh4ACW4KF93Zk1gRTj5aKkfEIhl7+17uID205luI3dtonwcZaCFhERORO4PXRxKbWOSNwYojdL/rC0mO8UF3U+6ro2Orbs0nlTZXsowr9v2sDzlRX0T0pm0YQLSfIcuU58uaqSdxvqubFo2KGfHG3hwD0StC2wtbGRJXUVrGysoSUcZmd7140sj/JnMjohi2db99Lshng4ewppTnS6w1CHBSzGMUQi0BFw+Y/gSvbT0mX1iIjIGa1bt47cmzcxJvuate9FtY7ImcES7de+Y8t2tja2cPvZI/E4Dlsbmhieno7FHLWt91AvdGlqGh5jjvpZ+9paea7yAFcVDKJvYhItoTALdm0j2etla3MTlYEA29oaY3h08F6wnveC9Z3ff69uOV9KGcxFCfl4cHCMwRiDx2NxHPiyKeYOuy6mNYqIiMSbtXDoH8U9jka0pVvyGsOM3Fym5/ZlSVUFgYjLazXVnc/dUFRKSUoqXsfh6f17eb6mIs4Vn5oM4+eqhCFMTSggErF0BFx+EV7LDmJ7MSAiImeMbjuiXexPt/f0ic2I9uwPYjuiraAtIiIi0vN166B9d/aEmOzrkoqXYnoetHKIiIiIiEgXUI+2iIiIiMSNtRBx411F19CItoiIiIhIF9CItoiIiIjElRvpmbcBakRbRERERKQLaERbREREROLHgqsebREREREROVka0RYRERGRuLFY9WiLiIiIiMjJU9AWEREREekCah0RERERkfjRzZAiIiIiInIqNKItIiIiInFjAdfVzZAiIiIiInKSNKItIiIiIvFjIRKJdxFdQyPaIiIiIiJdQCPaIiIiIhI36tEWEREREZFTohFtEREREYkfC656tEVERERE5GRpRFtERERE4kY92iIiIiIickqMtSd/BWGMqQH2dl05IiIiItIFBlprc+JdxPEYY14EsmO0u1pr7awY7evUgraIiIiIiJwctY6IiIiIiHQBBW0RERERkS6goC0iIiIi0gUUtEVEREREuoCCtoiIiIhIF1DQFhERERHpAgraIiIiIiJdQEFbRERERKQLKGiLiIiIiHSB/wfjDDCcfJhK8gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1008x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, figsize=(14, 10))\n",
    "plt.scatter(*umap_emb.T, s=3, c=train_data.target, cmap='Spectral', alpha=1.0)\n",
    "plt.setp(ax, xticks=[], yticks=[])\n",
    "ax.patch.set_facecolor('black')\n",
    "fg_color = 'black'\n",
    "cbar = plt.colorbar()\n",
    "plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color=fg_color)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aggregate-jumping",
   "metadata": {
    "papermill": {
     "duration": 34.243913,
     "end_time": "2021-05-20T17:13:51.505881",
     "exception": false,
     "start_time": "2021-05-20T17:13:17.261968",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Okay, in a supervised setting ```UMAP``` is able to isolate this structure, i.e. it seems that there exist a specific subspace which is kinda parametrized by the **readability target**. Is it possible to find a similar structure in a semi-supervised or unsupervised setting? Let's dive in 🏊🤿"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "australian-malaysia",
   "metadata": {
    "papermill": {
     "duration": 34.468642,
     "end_time": "2021-05-20T17:15:00.233679",
     "exception": false,
     "start_time": "2021-05-20T17:14:25.765037",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Validation Set\n",
    "\n",
    "Let's use ```scikit-learn```'s ```train_test_split``` method in order to extract a validation set and try to see whether a semi-supervised ```UMAP``` can capture some structures."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "worse-article",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:16:08.711477Z",
     "iopub.status.busy": "2021-05-20T17:16:08.710567Z",
     "iopub.status.idle": "2021-05-20T17:16:08.728422Z",
     "shell.execute_reply": "2021-05-20T17:16:08.727796Z"
    },
    "papermill": {
     "duration": 34.151865,
     "end_time": "2021-05-20T17:16:08.728563",
     "exception": false,
     "start_time": "2021-05-20T17:15:34.576698",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(\n",
    "    arr_w2v, train_data[[\"target\", \"standard_error\"]].values, test_size=0.10, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "revolutionary-wrapping",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:17:17.510238Z",
     "iopub.status.busy": "2021-05-20T17:17:17.509336Z",
     "iopub.status.idle": "2021-05-20T17:17:34.631145Z",
     "shell.execute_reply": "2021-05-20T17:17:34.630637Z"
    },
    "papermill": {
     "duration": 51.395828,
     "end_time": "2021-05-20T17:17:34.631281",
     "exception": false,
     "start_time": "2021-05-20T17:16:43.235453",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "mapper = umap.UMAP(n_neighbors= 15, n_components = 2, target_metric = 'l1' , n_epochs = 1000).fit(X_train, y=y_train[:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "greatest-august",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:18:43.203448Z",
     "iopub.status.busy": "2021-05-20T17:18:43.202888Z",
     "iopub.status.idle": "2021-05-20T17:18:51.180735Z",
     "shell.execute_reply": "2021-05-20T17:18:51.179779Z"
    },
    "papermill": {
     "duration": 42.269173,
     "end_time": "2021-05-20T17:18:51.180874",
     "exception": false,
     "start_time": "2021-05-20T17:18:08.911701",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "val_embedding = mapper.transform(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "engaged-advertiser",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:19:59.745695Z",
     "iopub.status.busy": "2021-05-20T17:19:59.744842Z",
     "iopub.status.idle": "2021-05-20T17:19:59.922294Z",
     "shell.execute_reply": "2021-05-20T17:19:59.921740Z"
    },
    "papermill": {
     "duration": 34.342759,
     "end_time": "2021-05-20T17:19:59.922424",
     "exception": false,
     "start_time": "2021-05-20T17:19:25.579665",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, figsize=(14, 10))\n",
    "plt.scatter(*mapper.embedding_.T, s=3, c=y_train[:,0], cmap='Spectral', alpha=1.0)\n",
    "plt.setp(ax, xticks=[], yticks=[])\n",
    "ax.patch.set_facecolor('black')\n",
    "fg_color = 'black'\n",
    "cbar = plt.colorbar()\n",
    "plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color=fg_color)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "aware-address",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:21:08.725020Z",
     "iopub.status.busy": "2021-05-20T17:21:08.724145Z",
     "iopub.status.idle": "2021-05-20T17:21:08.870308Z",
     "shell.execute_reply": "2021-05-20T17:21:08.869845Z"
    },
    "papermill": {
     "duration": 34.43352,
     "end_time": "2021-05-20T17:21:08.870444",
     "exception": false,
     "start_time": "2021-05-20T17:20:34.436924",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, figsize=(14, 10))\n",
    "plt.scatter(*val_embedding.T, s=3, c=y_val[:,0], cmap='Spectral', alpha=1.0)\n",
    "plt.setp(ax, xticks=[], yticks=[])\n",
    "ax.patch.set_facecolor('black')\n",
    "fg_color = 'black'\n",
    "cbar = plt.colorbar()\n",
    "plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color=fg_color)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "stock-credit",
   "metadata": {
    "papermill": {
     "duration": 34.527118,
     "end_time": "2021-05-20T17:22:17.755351",
     "exception": false,
     "start_time": "2021-05-20T17:21:43.228233",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Unfortunately the result is not what we wanted it to be. Anyway, a further investigation of the supervised representation extracted by ```UMAP``` is worth to be done even by considering it for augmentation strategies"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "blocked-lease",
   "metadata": {
    "papermill": {
     "duration": 34.150745,
     "end_time": "2021-05-20T17:23:26.327435",
     "exception": false,
     "start_time": "2021-05-20T17:22:52.176690",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# XGBoost on W2V features\n",
    "\n",
    "Here we apply a ```XGBoost``` regressor on our Word2Vec features,  in order to be ready for submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "union-airplane",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:24:35.592438Z",
     "iopub.status.busy": "2021-05-20T17:24:35.591856Z",
     "iopub.status.idle": "2021-05-20T17:24:35.758271Z",
     "shell.execute_reply": "2021-05-20T17:24:35.757633Z"
    },
    "papermill": {
     "duration": 34.834909,
     "end_time": "2021-05-20T17:24:35.758417",
     "exception": false,
     "start_time": "2021-05-20T17:24:00.923508",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import xgboost as xgb\n",
    "from sklearn.metrics import mean_squared_error as mse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "original-persian",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:25:44.349806Z",
     "iopub.status.busy": "2021-05-20T17:25:44.349224Z",
     "iopub.status.idle": "2021-05-20T17:26:00.105455Z",
     "shell.execute_reply": "2021-05-20T17:26:00.105843Z"
    },
    "papermill": {
     "duration": 49.932804,
     "end_time": "2021-05-20T17:26:00.106020",
     "exception": false,
     "start_time": "2021-05-20T17:25:10.173216",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/xgboost/data.py:114: UserWarning: Use subset (sliced data) of np.ndarray is not recommended because it will generate extra copies and increase memory consumption\n",
      "  \"because it will generate extra copies and increase \" +\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBRegressor(alpha=10, base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "             colsample_bynode=1, colsample_bytree=0.5, gamma=0, gpu_id=-1,\n",
       "             importance_type='gain', interaction_constraints='',\n",
       "             learning_rate=0.1, max_delta_step=0, max_depth=5,\n",
       "             min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "             n_estimators=1000, n_jobs=4, num_parallel_tree=1, random_state=0,\n",
       "             reg_alpha=10, reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
       "             tree_method='exact', validate_parameters=1, verbosity=1)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xg_reg = xgb.XGBRegressor(objective ='reg:squarederror', colsample_bytree = 0.5, learning_rate = 0.1,\n",
    "                max_depth = 5, alpha = 10, n_estimators = 1000, verbosity = 1)\n",
    "\n",
    "xg_reg.fit(X_train,y_train[:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "identical-henry",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:27:08.503744Z",
     "iopub.status.busy": "2021-05-20T17:27:08.503089Z",
     "iopub.status.idle": "2021-05-20T17:27:08.515409Z",
     "shell.execute_reply": "2021-05-20T17:27:08.514678Z"
    },
    "papermill": {
     "duration": 34.056209,
     "end_time": "2021-05-20T17:27:08.515609",
     "exception": false,
     "start_time": "2021-05-20T17:26:34.459400",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "preds = xg_reg.predict(X_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "tough-malawi",
   "metadata": {
    "papermill": {
     "duration": 34.275726,
     "end_time": "2021-05-20T17:28:17.147375",
     "exception": false,
     "start_time": "2021-05-20T17:27:42.871649",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "~0.69 not bad, let's see on test :) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "continued-polyester",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:29:25.760694Z",
     "iopub.status.busy": "2021-05-20T17:29:25.760108Z",
     "iopub.status.idle": "2021-05-20T17:29:25.763883Z",
     "shell.execute_reply": "2021-05-20T17:29:25.763424Z"
    },
    "papermill": {
     "duration": 34.088093,
     "end_time": "2021-05-20T17:29:25.764019",
     "exception": false,
     "start_time": "2021-05-20T17:28:51.675926",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.656116743177399"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sqrt(mse(y_val[:,0], preds)) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "burning-olive",
   "metadata": {
    "papermill": {
     "duration": 34.155702,
     "end_time": "2021-05-20T17:30:34.141671",
     "exception": false,
     "start_time": "2021-05-20T17:29:59.985969",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# W2V Naive Augmentation \n",
    "\n",
    "In this section, we'll try to instantiate a naive augmentation pipeline for Word2Vec features. The rationale behind the following strategy is that W2V features are essentially vectors so we can augment our \"feature set\" by perturbing these vectors with a small gaussian noise, in order to facilitate the learning procedure for our regressor of choice. Moreover, we'd like to see if by augmenting our \"feature set\" in this way we can **break** the structure extracted by the supervised ```UMAP```.\n",
    "\n",
    "Of course, this naive augmentation doesn't make a lot of sense from a word/sentence perspective\n",
    "since it might be that the perturbed vectors don't correspond to any word/sentence themselves.\n",
    "\n",
    "With the following method, ```augment_train_w2v```, we augment the training set by adding a small gaussian noise on w2v's features while slightly modifying (by exploiting the **standard deviation** of each target) the new target variable corresponding to the augmented sample."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "inappropriate-apache",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:31:43.009446Z",
     "iopub.status.busy": "2021-05-20T17:31:43.008601Z",
     "iopub.status.idle": "2021-05-20T17:31:43.010621Z",
     "shell.execute_reply": "2021-05-20T17:31:43.011097Z"
    },
    "papermill": {
     "duration": 34.381037,
     "end_time": "2021-05-20T17:31:43.011285",
     "exception": false,
     "start_time": "2021-05-20T17:31:08.630248",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def augment_train_w2v(X_train, y_train, y_std, times = 5):\n",
    "    \n",
    "    augmented_w2v_X = X_train.copy()\n",
    "    augmented_w2v_y = y_train.copy()\n",
    "    \n",
    "    for j in range(0, times - 1):\n",
    "        for i in range(0, X_train.shape[0]):\n",
    "        \n",
    "            new_w2v = X_train[i,:] + np.random.uniform(1,8)*1e-4*np.random.randn(300)       # np.random.uniform(1,8)*1e-3*np.random.randn(300)\n",
    "            augmented_w2v_X = np.vstack((augmented_w2v_X, new_w2v))\n",
    "            \n",
    "            new_y = y_train[i] + np.random.choice([-1, 1])*y_std[i]*0.05\n",
    "            augmented_w2v_y = np.append(augmented_w2v_y, new_y)\n",
    "        \n",
    "    return augmented_w2v_X, augmented_w2v_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "figured-background",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:32:51.921955Z",
     "iopub.status.busy": "2021-05-20T17:32:51.921371Z",
     "iopub.status.idle": "2021-05-20T17:34:14.070271Z",
     "shell.execute_reply": "2021-05-20T17:34:14.069675Z"
    },
    "papermill": {
     "duration": 116.726804,
     "end_time": "2021-05-20T17:34:14.070437",
     "exception": false,
     "start_time": "2021-05-20T17:32:17.343633",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "aug_X_train, aug_Y_train = augment_train_w2v(X_train, y_train[:,0], y_train[:,1], times = 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "wooden-yahoo",
   "metadata": {
    "papermill": {
     "duration": 34.331457,
     "end_time": "2021-05-20T17:35:22.686733",
     "exception": false,
     "start_time": "2021-05-20T17:34:48.355276",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We decided to shuffle again the new augmented training set by keeping the information on the sample-target couple. That's what the following method ```shuffle_couples``` does. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "mexican-smith",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:36:31.271455Z",
     "iopub.status.busy": "2021-05-20T17:36:31.270844Z",
     "iopub.status.idle": "2021-05-20T17:36:31.273700Z",
     "shell.execute_reply": "2021-05-20T17:36:31.273241Z"
    },
    "papermill": {
     "duration": 34.111542,
     "end_time": "2021-05-20T17:36:31.273905",
     "exception": false,
     "start_time": "2021-05-20T17:35:57.162363",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def shuffle_couples(a, b):\n",
    "    assert len(a) == len(b)\n",
    "    p = np.random.permutation(len(a))\n",
    "    return a[p], b[p]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "initial-montgomery",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:37:39.943905Z",
     "iopub.status.busy": "2021-05-20T17:37:39.943307Z",
     "iopub.status.idle": "2021-05-20T17:37:39.962360Z",
     "shell.execute_reply": "2021-05-20T17:37:39.961878Z"
    },
    "papermill": {
     "duration": 34.448727,
     "end_time": "2021-05-20T17:37:39.962501",
     "exception": false,
     "start_time": "2021-05-20T17:37:05.513774",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "aug_X_train, aug_Y_train = shuffle_couples(aug_X_train, aug_Y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "conventional-robert",
   "metadata": {
    "papermill": {
     "duration": 34.489595,
     "end_time": "2021-05-20T17:38:48.528078",
     "exception": false,
     "start_time": "2021-05-20T17:38:14.038483",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Now let's see if the previously seen structure still exist ( you can try to change the parameter ```times``` of ```augment_train_w2v``` to see weather it affects something or not) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "appreciated-italic",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:39:57.028568Z",
     "iopub.status.busy": "2021-05-20T17:39:57.027871Z",
     "iopub.status.idle": "2021-05-20T17:40:44.904166Z",
     "shell.execute_reply": "2021-05-20T17:40:44.904639Z"
    },
    "papermill": {
     "duration": 82.077662,
     "end_time": "2021-05-20T17:40:44.904829",
     "exception": false,
     "start_time": "2021-05-20T17:39:22.827167",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "mapper_2 = umap.UMAP(n_neighbors= 15, n_components = 2, min_dist = 0.1, target_metric = 'l1' , n_epochs = 1000).fit(aug_X_train)#, y=aug_Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "ideal-crack",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:41:54.145861Z",
     "iopub.status.busy": "2021-05-20T17:41:54.145302Z",
     "iopub.status.idle": "2021-05-20T17:42:12.986014Z",
     "shell.execute_reply": "2021-05-20T17:42:12.985424Z"
    },
    "papermill": {
     "duration": 53.340348,
     "end_time": "2021-05-20T17:42:12.986183",
     "exception": false,
     "start_time": "2021-05-20T17:41:19.645835",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "val_embedding_2 = mapper_2.transform(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "large-eating",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:43:21.231131Z",
     "iopub.status.busy": "2021-05-20T17:43:21.230512Z",
     "iopub.status.idle": "2021-05-20T17:43:21.514085Z",
     "shell.execute_reply": "2021-05-20T17:43:21.501597Z"
    },
    "papermill": {
     "duration": 34.484478,
     "end_time": "2021-05-20T17:43:21.514244",
     "exception": false,
     "start_time": "2021-05-20T17:42:47.029766",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, figsize=(14, 10))\n",
    "plt.scatter(*mapper_2.embedding_.T, s=3, c=aug_Y_train, cmap='Spectral', alpha=1.0)\n",
    "plt.setp(ax, xticks=[], yticks=[])\n",
    "ax.patch.set_facecolor('black')\n",
    "fg_color = 'black'\n",
    "cbar = plt.colorbar()\n",
    "plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color=fg_color)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "former-buffer",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:44:29.855607Z",
     "iopub.status.busy": "2021-05-20T17:44:29.852836Z",
     "iopub.status.idle": "2021-05-20T17:44:29.982013Z",
     "shell.execute_reply": "2021-05-20T17:44:29.981567Z"
    },
    "papermill": {
     "duration": 34.544951,
     "end_time": "2021-05-20T17:44:29.982178",
     "exception": false,
     "start_time": "2021-05-20T17:43:55.437227",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, figsize=(14, 10))\n",
    "plt.scatter(*val_embedding_2.T, s=3, c=y_val[:,0], cmap='Spectral', alpha=1.0)\n",
    "plt.setp(ax, xticks=[], yticks=[])\n",
    "ax.patch.set_facecolor('black')\n",
    "fg_color = 'black'\n",
    "cbar = plt.colorbar()\n",
    "plt.setp(plt.getp(cbar.ax.axes, 'yticklabels'), color=fg_color)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "occasional-annex",
   "metadata": {
    "papermill": {
     "duration": 34.427535,
     "end_time": "2021-05-20T17:45:38.911771",
     "exception": false,
     "start_time": "2021-05-20T17:45:04.484236",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Apparently the structure is affected by this augmentation strategy. After that we can take a look at the ```rmse``` score, maybe we can get a better result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "religious-stranger",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:46:47.311012Z",
     "iopub.status.busy": "2021-05-20T17:46:47.310454Z",
     "iopub.status.idle": "2021-05-20T17:48:29.251994Z",
     "shell.execute_reply": "2021-05-20T17:48:29.252430Z"
    },
    "papermill": {
     "duration": 136.244362,
     "end_time": "2021-05-20T17:48:29.252619",
     "exception": false,
     "start_time": "2021-05-20T17:46:13.008257",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBRegressor(alpha=10, base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "             colsample_bynode=1, colsample_bytree=0.5, gamma=0, gpu_id=-1,\n",
       "             importance_type='gain', interaction_constraints='',\n",
       "             learning_rate=0.1, max_delta_step=0, max_depth=5,\n",
       "             min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "             n_estimators=1000, n_jobs=4, num_parallel_tree=1, random_state=0,\n",
       "             reg_alpha=10, reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
       "             tree_method='exact', validate_parameters=1, verbosity=1)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xg_reg = xgb.XGBRegressor(objective ='reg:squarederror', colsample_bytree = 0.5, learning_rate = 0.1,\n",
    "                max_depth = 5, alpha = 10, n_estimators = 1000, verbosity = 1)\n",
    "\n",
    "xg_reg.fit(aug_X_train, aug_Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "later-seafood",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:49:37.668208Z",
     "iopub.status.busy": "2021-05-20T17:49:37.667474Z",
     "iopub.status.idle": "2021-05-20T17:49:37.674673Z",
     "shell.execute_reply": "2021-05-20T17:49:37.675115Z"
    },
    "papermill": {
     "duration": 34.349513,
     "end_time": "2021-05-20T17:49:37.675281",
     "exception": false,
     "start_time": "2021-05-20T17:49:03.325768",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "preds = xg_reg.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "productive-assurance",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:50:51.125583Z",
     "iopub.status.busy": "2021-05-20T17:50:51.124897Z",
     "iopub.status.idle": "2021-05-20T17:50:51.130193Z",
     "shell.execute_reply": "2021-05-20T17:50:51.130596Z"
    },
    "papermill": {
     "duration": 39.259835,
     "end_time": "2021-05-20T17:50:51.130762",
     "exception": false,
     "start_time": "2021-05-20T17:50:11.870927",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6792162014766102"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sqrt(mse(y_val[:,0], preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "moral-syntax",
   "metadata": {
    "papermill": {
     "duration": 34.040762,
     "end_time": "2021-05-20T17:51:59.693291",
     "exception": false,
     "start_time": "2021-05-20T17:51:25.652529",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "The result is pretty much the same, we need to devise a better augmentation strategy, maybe by exploiting the information gained with ```UMAP```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "arbitrary-edmonton",
   "metadata": {
    "papermill": {
     "duration": 34.535069,
     "end_time": "2021-05-20T17:53:08.416180",
     "exception": false,
     "start_time": "2021-05-20T17:52:33.881111",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Test Data\n",
    "\n",
    "Preparing submission with test data by basically re-doing the previously seen operations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "arctic-mortgage",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:54:17.069826Z",
     "iopub.status.busy": "2021-05-20T17:54:17.069266Z",
     "iopub.status.idle": "2021-05-20T17:54:17.099664Z",
     "shell.execute_reply": "2021-05-20T17:54:17.100078Z"
    },
    "papermill": {
     "duration": 34.522018,
     "end_time": "2021-05-20T17:54:17.100251",
     "exception": false,
     "start_time": "2021-05-20T17:53:42.578233",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>url_legal</th>\n",
       "      <th>license</th>\n",
       "      <th>excerpt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>c0f722661</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>My hope lay in Jack's promise that he would ke...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>f0953f0a5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Dotty continued to go to Mrs. Gray's every nig...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0df072751</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>It was a bright and cheerful scene that greete...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>04caf4e0c</td>\n",
       "      <td>https://en.wikipedia.org/wiki/Cell_division</td>\n",
       "      <td>CC BY-SA 3.0</td>\n",
       "      <td>Cell division is the process by which a parent...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0e63f8bea</td>\n",
       "      <td>https://en.wikipedia.org/wiki/Debugging</td>\n",
       "      <td>CC BY-SA 3.0</td>\n",
       "      <td>Debugging is the process of finding and resolv...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          id                                    url_legal       license  \\\n",
       "0  c0f722661                                          NaN           NaN   \n",
       "1  f0953f0a5                                          NaN           NaN   \n",
       "2  0df072751                                          NaN           NaN   \n",
       "3  04caf4e0c  https://en.wikipedia.org/wiki/Cell_division  CC BY-SA 3.0   \n",
       "4  0e63f8bea      https://en.wikipedia.org/wiki/Debugging  CC BY-SA 3.0   \n",
       "\n",
       "                                             excerpt  \n",
       "0  My hope lay in Jack's promise that he would ke...  \n",
       "1  Dotty continued to go to Mrs. Gray's every nig...  \n",
       "2  It was a bright and cheerful scene that greete...  \n",
       "3  Cell division is the process by which a parent...  \n",
       "4  Debugging is the process of finding and resolv...  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data =  pd.read_csv(\"/kaggle/input/commonlitreadabilityprize/test.csv\")\n",
    "test_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "continuing-benefit",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:55:25.695880Z",
     "iopub.status.busy": "2021-05-20T17:55:25.695318Z",
     "iopub.status.idle": "2021-05-20T17:55:25.703324Z",
     "shell.execute_reply": "2021-05-20T17:55:25.702796Z"
    },
    "papermill": {
     "duration": 34.341143,
     "end_time": "2021-05-20T17:55:25.703463",
     "exception": false,
     "start_time": "2021-05-20T17:54:51.362320",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Replaced 0 words with characters with an ordinal >= 128 in the test data.\n"
     ]
    }
   ],
   "source": [
    "test_data.excerpt = test_data['excerpt'].apply(str)\n",
    "non_ascii_words = remove_ascii_words(test_data)\n",
    "\n",
    "print(\"Replaced {} words with characters with an ordinal >= 128 in the test data.\".format(\n",
    "    len(non_ascii_words)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "appropriate-success",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:56:34.212156Z",
     "iopub.status.busy": "2021-05-20T17:56:34.211265Z",
     "iopub.status.idle": "2021-05-20T17:56:34.213954Z",
     "shell.execute_reply": "2021-05-20T17:56:34.213515Z"
    },
    "papermill": {
     "duration": 34.302724,
     "end_time": "2021-05-20T17:56:34.214107",
     "exception": false,
     "start_time": "2021-05-20T17:55:59.911383",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "w2v_preprocessing(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "owned-suggestion",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:57:42.672037Z",
     "iopub.status.busy": "2021-05-20T17:57:42.671214Z",
     "iopub.status.idle": "2021-05-20T17:57:42.673718Z",
     "shell.execute_reply": "2021-05-20T17:57:42.673194Z"
    },
    "papermill": {
     "duration": 34.105056,
     "end_time": "2021-05-20T17:57:42.673845",
     "exception": false,
     "start_time": "2021-05-20T17:57:08.568789",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "test_data.drop(test_data[test_data.tokenized_sentences.str.len() == 0].index, inplace= True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "genuine-chuck",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T17:58:51.315346Z",
     "iopub.status.busy": "2021-05-20T17:58:51.314636Z",
     "iopub.status.idle": "2021-05-20T17:58:51.318109Z",
     "shell.execute_reply": "2021-05-20T17:58:51.318479Z"
    },
    "papermill": {
     "duration": 34.297236,
     "end_time": "2021-05-20T17:58:51.318652",
     "exception": false,
     "start_time": "2021-05-20T17:58:17.021416",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of sentences: 25633.\n",
      "Number of texts: 7.\n"
     ]
    }
   ],
   "source": [
    "#create dictionary with all sentences\n",
    "sentences_test = []\n",
    "for sentence_group in test_data.tokenized_sentences:\n",
    "    sentences_test.extend(sentence_group)\n",
    "\n",
    "print(\"Number of sentences: {}.\".format(len(sentences)))\n",
    "print(\"Number of texts: {}.\".format(len(test_data)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "informed-divorce",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T18:00:00.202540Z",
     "iopub.status.busy": "2021-05-20T18:00:00.201920Z",
     "iopub.status.idle": "2021-05-20T18:00:04.127631Z",
     "shell.execute_reply": "2021-05-20T18:00:04.128228Z"
    },
    "papermill": {
     "duration": 38.727493,
     "end_time": "2021-05-20T18:00:04.128509",
     "exception": false,
     "start_time": "2021-05-20T17:59:25.401016",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "test_data['w2v_features'] = list(map(lambda sen_group:\n",
    "                                     get_w2v_features(W2Vmodel, sen_group),\n",
    "                                     test_data.tokenized_sentences))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "gorgeous-backing",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T18:01:12.788697Z",
     "iopub.status.busy": "2021-05-20T18:01:12.787859Z",
     "iopub.status.idle": "2021-05-20T18:01:12.790873Z",
     "shell.execute_reply": "2021-05-20T18:01:12.790463Z"
    },
    "papermill": {
     "duration": 34.411229,
     "end_time": "2021-05-20T18:01:12.791002",
     "exception": false,
     "start_time": "2021-05-20T18:00:38.379773",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "test_data[\"w2v_resh_features\"] = test_data[\"w2v_features\"].apply(lambda x : x.reshape(1,-1) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "equivalent-humor",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-20T18:02:21.120595Z",
     "iopub.status.busy": "2021-05-20T18:02:21.119974Z",
     "iopub.status.idle": "2021-05-20T18:02:21.122107Z",
     "shell.execute_reply": "2021-05-20T18:02:21.122534Z"
    },
    "papermill": {
     "duration": 34.21501,
     "end_time": "2021-05-20T18:02:21.122687",
     "exception": false,
     "start_time": "2021-05-20T18:01:46.907677",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "arr_w2v_test = test_data.w2v_resh_features[0]\n",
    "for i in range(1, len(test_data)):\n",
    "    arr_w2v_test = np.vstack((arr_w2v_test, test_data.w2v_resh_features[i]))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 6272.836049,
   "end_time": "2021-05-20T18:02:58.466764",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-05-20T16:18:25.630715",
   "version": "2.3.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}