mapping_pipeline.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Calcul des mappings - Introduction\n",
    "Ce notebook sert à calculer un mapping entre deux graphes : il s'agit, pour chaque noeud du graphe A, d'associer un noeud du graphe B.\n",
    "\n",
    "Ici, les noeuds possèdent un libellé. On calcule l'embedding d'un noeud comme la concaténation de l'embedding de son parent et celui de son libellé. L'embedding des libellé est calculé comme la moyenne des embeddings des mots, eux-mêmes tirés de Fast Text.\n",
    "\n",
    "Au préalable, les données sont organisées dans `./parsing_pipeline.ipynb` selon le format suivant :\n",
    "```\n",
    "interface Node {\n",
    "    name: string,\n",
    "    children: Node[],\n",
    "}\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Commandes pour générer les fichiers `.csv` depuis le dump envoyé par la Direction du Budget :\n",
    "```\n",
    "ls | xargs -i basename {} .xlsx | xargs -i ssconvert {}.xlsx {}.csv\n",
    "find -name '*.csv' | xargs -i wc -l {}\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/alexis/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import csv\n",
    "import json\n",
    "import keras\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "from collections import defaultdict\n",
    "from sklearn.neighbors import NearestNeighbors\n",
    "from tqdm import tqdm_notebook, tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "EMBEDDING_FILEPATH = '/home/alexis/mva/word_embeddings/wiki.fr.vec'\n",
    "\n",
    "PLF_NOMENCLATURES_FOLDER = './raw_data/'\n",
    "PLF_GRAPH_OUTPUT_FOLDER = '../server/resources/plf_all_nodes/'\n",
    "VOTES_OUTPUT_FOLDER = '../server/resources/plf_votes/'\n",
    "MAPPINGS_OUTPUT_FOLDER = '../server/resources/plf_mappings/'\n",
    "\n",
    "WORD_EMBEDDING_SIZE = 300\n",
    "K = 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chargement des embeddings"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load FastText pretrained embeddings. These should be downloaded from `https://s3-us-west-1.amazonaws.com/fasttext-vectors/wiki.fr.vec`.\n",
    "\n",
    "`EMBEDDING_FILE_PATH` should point to where this file is stored."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gensim\n",
    "from gensim.models import Word2Vec\n",
    "pretrained_embeddings = gensim.models.KeyedVectors.load_word2vec_format(EMBEDDING_FILEPATH, binary=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Mapping de nomenclatures"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pipeline\n",
    "Dans un premier temps, on calcule les embeddings de chacun des noeuds.\n",
    "Pour cela, on parcourt l'arbre récursivement, en ajoutant chaque fois l'embedding\n",
    "du noeud actuel à une liste.\n",
    "L'embedding d'un noeud est pour moitié celui de son parent, pour moitié l'embedding\n",
    "de son propre libellé. L'embedding du libellé est lui-même la moyenne des embedding\n",
    "de chacun des mots du libellé."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_all_nodes_from_json(root_node):\n",
    "    nodes = []\n",
    "\n",
    "    def rec_build_embedding(node, parent_embedding, accPath):\n",
    "        ## Split current node's name\n",
    "        words = keras.preprocessing.text.text_to_word_sequence(\n",
    "            node['name'],\n",
    "            filters=\"!\\\"#$%&()*+,-./:;<=>?@[\\\\]^_`{|}~\\t\\n'´’™©®«»\",\n",
    "            lower=True,\n",
    "            split=' '\n",
    "        )\n",
    "\n",
    "        ## Initiate sentence embedding;\n",
    "        ## Sentence embedding will be the average\n",
    "        ## of all present word embeddings.\n",
    "        sentence_embedding = np.zeros((1, WORD_EMBEDDING_SIZE))\n",
    "\n",
    "        ## Create embedding for each word\n",
    "        for word in words:\n",
    "            try:\n",
    "                sentence_embedding += pretrained_embeddings[word.lower()]\n",
    "            except:\n",
    "                # If word does not exist in pretrained embedding,\n",
    "                # give it null embedding\n",
    "                pass\n",
    "\n",
    "        ## Normalize sentence embedding\n",
    "        sentence_embedding /= max(len(words), 1)\n",
    "\n",
    "        ## Concatenate embedding with parent node embedding\n",
    "        node_embedding = np.concatenate((\n",
    "            np.array(parent_embedding[0]).reshape((1, WORD_EMBEDDING_SIZE)) +\n",
    "            np.array(parent_embedding[1]).reshape((1, WORD_EMBEDDING_SIZE)),\n",
    "            sentence_embedding\n",
    "        ))\n",
    "        \n",
    "        r = dict({\n",
    "            'embedding': node_embedding\n",
    "        })\n",
    "        \n",
    "        for key in node:\n",
    "            if key != 'children':\n",
    "                r[key] = node[key]\n",
    "        \n",
    "        r['libelles'] = accPath + [node['name']]\n",
    "\n",
    "        nodes.append(r)\n",
    "        \n",
    "        if 'children' in node:\n",
    "            for child in node['children']:\n",
    "                rec_build_embedding(child, node_embedding, accPath + [node['name']])\n",
    "    \n",
    "    for node in root_node['children']:\n",
    "        rec_build_embedding(node, np.zeros((2, WORD_EMBEDDING_SIZE)), [])\n",
    "\n",
    "    return nodes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Une simple fonction pour effectuer un KNN avec scikit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_nearest_neighbours(source_nodes, target_nodes, k=2):\n",
    "    # Stack and flatten embeddings\n",
    "    target_embeddings = np.array([node['embedding'].flatten() for node in target_nodes])\n",
    "    source_embeddings = np.array([node['embedding'].flatten() for node in source_nodes])\n",
    "\n",
    "    # Initiate scikit learn NearestNeighbors\n",
    "    neighbours = NearestNeighbors(n_neighbors=k, algorithm='ball_tree').fit(target_embeddings)\n",
    "\n",
    "    distances, indices = neighbours.kneighbors(source_embeddings)\n",
    "\n",
    "    return distances, indices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Une simple fonction pour écrire les mappings générés dans un fichier json.\n",
    "Ils sont triés par `key` : on récupère le code de chaque noeud, puis on écrit la liste\n",
    "des noeuds suggérés (ici, on se tient à une liste de longueur 1)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "def neighbours_to_json_file(source_nodes, target_nodes, indices, distances, output_filepath):\n",
    "    mappings_dict = defaultdict(list)\n",
    "\n",
    "    for source_index, (row_indices, row_distances) in enumerate(zip(indices, distances)):\n",
    "        # Fill mappingsDict\n",
    "        mappings_dict[source_nodes[source_index]['code']].append(dict({\n",
    "            'distance': row_distances[0],\n",
    "            'libelles': target_nodes[row_indices[0]]['libelles'],\n",
    "            'code': target_nodes[row_indices[0]]['code']\n",
    "        }))\n",
    "\n",
    "    with open(output_filepath, 'w+') as f:\n",
    "        json.dump(mappings_dict, f, default=lambda o: o.__dict__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quelques tests pour vérifier que le mapping s'est correctement passé."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tests(distances):\n",
    "    print('Percentage of not directly matched: {:.2f} % ({} / {})'.format(\n",
    "        distances[distances[:,0] != 0].shape[0] / distances.shape[0] * 100,\n",
    "        distances[distances[:,0] != 0].shape[0],\n",
    "        distances.shape[0]\n",
    "    ))\n",
    "    \n",
    "    print('Percentage of not uniquely matched: {:.2f} % ({} / {})'.format(\n",
    "        len([1 for d in distances if (d[0] == 0 and d[1] == 0)]) / distances.shape[0] * 100,\n",
    "        len([1 for d in distances if (d[0] == 0 and d[1] == 0)]),\n",
    "        distances.shape[0]\n",
    "    ))\n",
    "    \n",
    "    f = plt.figure(figsize=(10, 4))\n",
    "    ax = plt.subplot(1, 1, 1)\n",
    "    ax.set_title('Distribution of distances of imperfectly matched nodes')\n",
    "    n, bins, patches = ax.hist(distances[distances[:,0] != 0][:,0], 50)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Enfin, une fonction qui exécute les blocks précédents."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_pipeline(source_tree_file, target_tree_file, output_filepath):\n",
    "    # Load json files\n",
    "    with open(source_tree_file, 'r') as source_file:\n",
    "        source_root_node = json.load(source_file)\n",
    "    source_nodes = build_all_nodes_from_json(source_root_node)\n",
    "\n",
    "    with open(target_tree_file, 'r') as source_file:\n",
    "        target_root_node = json.load(source_file)\n",
    "    target_nodes = build_all_nodes_from_json(target_root_node)\n",
    "\n",
    "    # Compute neighbours\n",
    "    print('Compute KNN...')\n",
    "    distances, indices = find_nearest_neighbours(source_nodes, target_nodes, k=K)\n",
    "\n",
    "    # Output results\n",
    "    neighbours_to_json_file(source_nodes, target_nodes, indices, distances, output_filepath)\n",
    "\n",
    "    # Tests\n",
    "    tests(distances)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "SOURCE_TREE_FILE = '../server/resources/recettes/2018.json'\n",
    "TARGET_TREE_FILE = '../server/resources/recettes/2019.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(SOURCE_TREE_FILE, 'r') as source_file:\n",
    "    source_root_node = json.load(source_file)\n",
    "\n",
    "source_nodes = build_all_nodes_from_json(source_root_node)\n",
    "\n",
    "with open(TARGET_TREE_FILE, 'r') as source_file:\n",
    "    target_root_node = json.load(source_file)\n",
    "\n",
    "target_nodes = build_all_nodes_from_json(target_root_node)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 70.9 ms, sys: 3.99 ms, total: 74.9 ms\n",
      "Wall time: 71.7 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "distances, indices = find_nearest_neighbours(source_nodes, target_nodes, k=K)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [],
   "source": [
    "neighbours_to_json_file(source_nodes, target_nodes, indices, distances, '../server/resources/plf_mappings/recettes_2018_to_recettes_2019')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of not directly matched: 1.17 % (2 / 171)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 171)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9673df9470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tests(distances)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Création des mappings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [],
   "source": [
    "INFO = {\n",
    "    'PLF': {\n",
    "        'tree_folder': './parsed_data',\n",
    "        'to_folder': '../server/resources/mappings/plf'\n",
    "    },\n",
    "    'recettes': {\n",
    "        'tree_folder': '../server/resources/recettes',\n",
    "        'to_folder': '../server/resources/mappings/recettes'\n",
    "    }\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_mapping(source_year, target_year, source_type):\n",
    "    tree_folder = INFO[source_type]['tree_folder']\n",
    "    to_folder = INFO[source_type]['to_folder']\n",
    "\n",
    "    print('Compute {} to {}'.format(source_year, target_year))\n",
    "    source_tree_file = os.path.join(tree_folder, '{}.json'.format(source_year))\n",
    "    target_tree_file = os.path.join(tree_folder, '{}.json'.format(target_year))\n",
    "    run_pipeline(source_tree_file, target_tree_file, os.path.join(to_folder, '{}_to_{}.json'.format(source_year, target_year)))\n",
    "    print()\n",
    "\n",
    "def compute_mappings(min_year, max_year, source_type, bothDirections=True):\n",
    "    for year in range(min_year, max_year):\n",
    "        compute_mapping(year, year+1, source_type)\n",
    "        if bothDirections:\n",
    "            compute_mapping(year+1, year, source_type)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### PLF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Compute 2012 to 2013\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 26.93 % (755 / 2804)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2804)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f96f2ad2fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2013 to 2012\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 23.73 % (637 / 2684)\n",
      "Percentage of not uniquely matched: 0.07 % (2 / 2684)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9673371e48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2013 to 2014\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 17.44 % (468 / 2684)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2684)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9673e05c18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2014 to 2013\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 18.14 % (491 / 2707)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2707)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f96781ff2b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2014 to 2015\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 23.46 % (635 / 2707)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2707)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f96783c6b38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2015 to 2014\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 16.05 % (396 / 2468)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2468)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967847c1d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2015 to 2016\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 6.69 % (165 / 2468)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2468)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967849ed68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2016 to 2015\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 7.25 % (180 / 2483)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2483)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967898f198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2016 to 2017\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 10.03 % (249 / 2483)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2483)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f96789742e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2017 to 2016\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 9.59 % (237 / 2471)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2471)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967371d5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2017 to 2018\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 18.49 % (457 / 2471)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2471)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967807f908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2018 to 2017\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 18.99 % (472 / 2486)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 2486)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f967851c048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2018 to 2019\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 4.42 % (110 / 2486)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2486)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9672cb4be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2019 to 2018\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 3.06 % (75 / 2451)\n",
      "Percentage of not uniquely matched: 0.16 % (4 / 2451)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9672bb0400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "compute_mappings(2012, 2019, 'PLF')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Recettes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Compute 2018 to 2019\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 1.17 % (2 / 171)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 171)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9673246208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Compute 2019 to 2018\n",
      "Compute KNN...\n",
      "Percentage of not directly matched: 1.17 % (2 / 171)\n",
      "Percentage of not uniquely matched: 0.00 % (0 / 171)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9672a793c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "compute_mappings(2018, 2019, 'recettes')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mon sacrifice est introduit daaaaans le volcan."
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}