notebook/Pipeline Spacy.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "executionInfo": {
     "elapsed": 10793,
     "status": "ok",
     "timestamp": 1686305120892,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "yDdTUV5JECAw"
   },
   "outputs": [],
   "source": [
    "import spacy\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 412,
     "status": "ok",
     "timestamp": 1686305121283,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "yLvcPYKkD5l1",
    "outputId": "149e670f-a56f-43c3-8036-f43a2c20f350"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aucun                32663\n",
       "geogFeat              2704\n",
       "geogName name         2117\n",
       "geogFeat geogName     1463\n",
       "geogName               910\n",
       "Name: Label, dtype: int64"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"./train_2.csv\")\n",
    "df.Label.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 3,
     "status": "ok",
     "timestamp": 1686305121627,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "sOC_kK2_iT4G",
    "outputId": "e79708ef-3748-4adf-9de0-48cf95dfd978"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "aucun                32663\n",
       "geogFeat              2704\n",
       "name geogName         2117\n",
       "geogName geogFeat     1463\n",
       "geogName               910\n",
       "Name: Label, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's clean the labels\n",
    "def clean_label(string):\n",
    "    string = string.split(\" \")\n",
    "    string = \" \".join(list(set(string)))  \n",
    "    return(string)\n",
    "df[\"Label\"] = df[\"Label\"].apply(clean_label)\n",
    "\n",
    "def isgeogFeat(string):\n",
    "    geogFeat = False\n",
    "    if \"geogFeat\" in string:\n",
    "        geogFeat = True\n",
    "    return(geogFeat)\n",
    "def isgeogName(string):\n",
    "    geogName = False\n",
    "    if \"geogName\" in string:\n",
    "        geogName = True\n",
    "    return(geogName)\n",
    "def isName(string):\n",
    "    Name = False\n",
    "    if (\"name\" in string):\n",
    "        Name = True\n",
    "    return(Name)\n",
    "\n",
    "df[\"Label\"] = df[\"Label\"].apply(clean_label)\n",
    "labels_count = df.Label.value_counts()\n",
    "labels_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "39857"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(labels_count)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Geographic labels :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7194"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(labels_count[labels_count.index!= 'aucun'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18.049527059236773"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(labels_count[labels_count.index!= 'aucun'])*100/sum(labels_count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# libraries\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    " \n",
    "# set width of bars\n",
    "barWidth = 0.5\n",
    " \n",
    "# set heights of bars\n",
    "bars1 = labels_count\n",
    " \n",
    "# Set position of bar on X axis\n",
    "r1 = np.arange(len(bars1))\n",
    "\n",
    "colors = [\"#aec7e8\", \"#c2e699\", \"#ffb366\", \"#f28d35\", \"#d9b3e6\"]\n",
    " \n",
    "# Make the plot\n",
    "plt.bar(r1, bars1, color=colors, width=barWidth)\n",
    "\n",
    "# Add xticks on the middle of the group bars\n",
    "plt.xticks([r  for r in range(len(bars1))], labels_count.index)\n",
    "plt.xticks(rotation=60, ha='right')\n",
    "#plt.xlabel('Classes')\n",
    "plt.ylabel('Nombre d\\'exemples')\n",
    "\n",
    "# Add values on top of the bars\n",
    "for i, value in enumerate(bars1):\n",
    "    plt.text(i, value + 0.05,f'n={value}', ha='center', va='bottom')\n",
    "    \n",
    "# Create legend & Show graphic\n",
    "#plt.legend()\n",
    "plt.savefig(\"class_distrib.png\", dpi = 300, bbox_inches = 'tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting fr-core-news-lg==3.6.0\n",
      "  Downloading https://github.com/explosion/spacy-models/releases/download/fr_core_news_lg-3.6.0/fr_core_news_lg-3.6.0-py3-none-any.whl (571.8 MB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m571.8/571.8 MB\u001b[0m \u001b[31m5.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hRequirement already satisfied: spacy<3.7.0,>=3.6.0 in /home/sarah/.local/lib/python3.9/site-packages (from fr-core-news-lg==3.6.0) (3.6.1)\n",
      "Requirement already satisfied: spacy-legacy<3.1.0,>=3.0.11 in /home/sarah/.local/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (3.0.12)\n",
      "Requirement already satisfied: spacy-loggers<2.0.0,>=1.0.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (1.0.3)\n",
      "Requirement already satisfied: murmurhash<1.1.0,>=0.28.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (1.0.8)\n",
      "Requirement already satisfied: cymem<2.1.0,>=2.0.2 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.0.6)\n",
      "Requirement already satisfied: preshed<3.1.0,>=3.0.2 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (3.0.7)\n",
      "Requirement already satisfied: thinc<8.2.0,>=8.1.8 in /home/sarah/.local/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (8.1.12)\n",
      "Requirement already satisfied: wasabi<1.2.0,>=0.9.1 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (0.10.1)\n",
      "Requirement already satisfied: srsly<3.0.0,>=2.4.3 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.4.4)\n",
      "Requirement already satisfied: catalogue<2.1.0,>=2.0.6 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.0.8)\n",
      "Requirement already satisfied: typer<0.10.0,>=0.3.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (0.4.2)\n",
      "Requirement already satisfied: pathy>=0.10.0 in /home/sarah/.local/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (0.10.2)\n",
      "Requirement already satisfied: smart-open<7.0.0,>=5.2.1 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (5.2.1)\n",
      "Requirement already satisfied: tqdm<5.0.0,>=4.38.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (4.65.0)\n",
      "Requirement already satisfied: numpy>=1.15.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (1.24.3)\n",
      "Requirement already satisfied: requests<3.0.0,>=2.13.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.31.0)\n",
      "Requirement already satisfied: pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4 in /home/sarah/.local/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (1.10.12)\n",
      "Requirement already satisfied: jinja2 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.11.3)\n",
      "Requirement already satisfied: setuptools in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (68.0.0)\n",
      "Requirement already satisfied: packaging>=20.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (23.0)\n",
      "Requirement already satisfied: langcodes<4.0.0,>=3.2.0 in /mnt/anaconda/lib/python3.9/site-packages (from spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (3.3.0)\n",
      "Requirement already satisfied: typing-extensions>=4.2.0 in /home/sarah/.local/lib/python3.9/site-packages (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (4.4.0)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in /mnt/anaconda/lib/python3.9/site-packages (from requests<3.0.0,>=2.13.0->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.0.4)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /mnt/anaconda/lib/python3.9/site-packages (from requests<3.0.0,>=2.13.0->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (3.4)\n",
      "Requirement already satisfied: urllib3<3,>=1.21.1 in /mnt/anaconda/lib/python3.9/site-packages (from requests<3.0.0,>=2.13.0->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (1.26.16)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /mnt/anaconda/lib/python3.9/site-packages (from requests<3.0.0,>=2.13.0->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2023.7.22)\n",
      "Requirement already satisfied: blis<0.8.0,>=0.7.8 in /mnt/anaconda/lib/python3.9/site-packages (from thinc<8.2.0,>=8.1.8->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (0.7.8)\n",
      "Requirement already satisfied: confection<1.0.0,>=0.0.1 in /mnt/anaconda/lib/python3.9/site-packages (from thinc<8.2.0,>=8.1.8->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (0.0.3)\n",
      "Requirement already satisfied: click<9.0.0,>=7.1.1 in /mnt/anaconda/lib/python3.9/site-packages (from typer<0.10.0,>=0.3.0->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (8.0.4)\n",
      "Requirement already satisfied: MarkupSafe>=0.23 in /mnt/anaconda/lib/python3.9/site-packages (from jinja2->spacy<3.7.0,>=3.6.0->fr-core-news-lg==3.6.0) (2.0.1)\n",
      "Installing collected packages: fr-core-news-lg\n",
      "Successfully installed fr-core-news-lg-3.6.0\n",
      "\u001b[38;5;2m✔ Download and installation successful\u001b[0m\n",
      "You can now load the package via spacy.load('fr_core_news_lg')\n"
     ]
    }
   ],
   "source": [
    "!python -m  spacy download fr_core_news_lg --user"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get the sentences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.replace('\"', '', regex=True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "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>Token</th>\n",
       "      <th>Label</th>\n",
       "      <th>Sentence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>01</td>\n",
       "      <td>aucun</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>4.5.3</td>\n",
       "      <td>aucun</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>.</td>\n",
       "      <td>aucun</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>Port</td>\n",
       "      <td>geogName geogFeat</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>du</td>\n",
       "      <td>geogName</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39852</th>\n",
       "      <td>39852</td>\n",
       "      <td>portée</td>\n",
       "      <td>aucun</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39853</th>\n",
       "      <td>39853</td>\n",
       "      <td>sur</td>\n",
       "      <td>aucun</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39854</th>\n",
       "      <td>39854</td>\n",
       "      <td>les</td>\n",
       "      <td>aucun</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39855</th>\n",
       "      <td>39855</td>\n",
       "      <td>cartes</td>\n",
       "      <td>aucun</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39856</th>\n",
       "      <td>39856</td>\n",
       "      <td>.</td>\n",
       "      <td>aucun</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>39857 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Id   Token              Label  Sentence\n",
       "0          0      01              aucun         0\n",
       "1          1   4.5.3              aucun         0\n",
       "2          2       .              aucun         0\n",
       "3          3    Port  geogName geogFeat         1\n",
       "4          4      du           geogName         1\n",
       "...      ...     ...                ...       ...\n",
       "39852  39852  portée              aucun      1701\n",
       "39853  39853     sur              aucun      1701\n",
       "39854  39854     les              aucun      1701\n",
       "39855  39855  cartes              aucun      1701\n",
       "39856  39856       .              aucun      1701\n",
       "\n",
       "[39857 rows x 4 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dot_index = [i for i,v in enumerate(df.Token.tolist()) if v==\".\"]\n",
    "sentence = list()\n",
    "s = 0\n",
    "for i in list(range(len(df.Token.tolist()))):\n",
    "    if df.Token.tolist()[i] != \".\":\n",
    "        sentence.append(s)\n",
    "    else:\n",
    "        sentence.append(s)\n",
    "        s = s+1\n",
    "df['Sentence'] = sentence\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create the training and test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "executionInfo": {
     "elapsed": 8979,
     "status": "ok",
     "timestamp": 1686305140070,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "hxaeXkUDrVzR"
   },
   "outputs": [],
   "source": [
    "from spacy.tokens import Doc\n",
    "from spacy.training import Example\n",
    "nlp = spacy.blank(\"fr\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_docs = []\n",
    "for s in list(set(df.Sentence)):\n",
    "    words = df[df.Sentence == s].Token.tolist()\n",
    "    labels = df[df.Sentence == s].Label.tolist()\n",
    "    doc = Doc(nlp.vocab, words=words)\n",
    "    entities_tmp = []\n",
    "    for i, w in enumerate(words):\n",
    "        if i == 0:\n",
    "            entities_tmp = [(0, len(w), labels[0])]\n",
    "            start = len(w)+1\n",
    "        else:\n",
    "            entities_tmp.append((start, start + len(w), labels[i]))\n",
    "            start = start + 1 + len(w)\n",
    "    train_docs.append((\" \".join(words), {'entities':entities_tmp}))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data, test_data = train_test_split(train_docs, test_size=0.20, random_state=42)\n",
    "id_sentences_train, id_sentences_test = train_test_split(range(len(train_docs)), test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1361"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(id_sentences_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "341"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(id_sentences_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 2143918,
     "status": "ok",
     "timestamp": 1686310056555,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "k_0oT4wnJMWu",
    "outputId": "82ab67ad-2a93-4407-c68d-601ff8713c28"
   },
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "epocs = 5 \n",
    "\n",
    "if \"ner\" not in nlp.pipe_names:\n",
    "    ner = nlp.add_pipe(\"ner\")\n",
    "else:\n",
    "    ner = nlp.get_pipe(\"ner\")\n",
    "\n",
    "# Add the labels to the NER pipeline\n",
    "if \"ner\" not in nlp.pipe_names:\n",
    "    ner = nlp.add_pipe(\"ner\")\n",
    "else:\n",
    "    ner = nlp.get_pipe(\"ner\")\n",
    "\n",
    "# Add the labels to the NER pipeline\n",
    "for _, annotations in train_data:\n",
    "    for ent in annotations.get('entities'):\n",
    "     \n",
    "        ner.add_label(ent[2])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 6805it [02:42, 41.86it/s, loss=1.17e+3]                                                   \n"
     ]
    }
   ],
   "source": [
    "epocs=5\n",
    "# Initialize the training process\n",
    "optimizer = nlp.begin_training()\n",
    "\n",
    "progress_bar = tqdm(total=len(train_data), desc=\"Training\")\n",
    "# Train the token classification model\n",
    "for _ in range(epocs):  # Adjust the number of training iterations as needed\n",
    "    losses = {}\n",
    "    for text, annotations in train_data:\n",
    "        doc = nlp.make_doc(text)\n",
    "        example = Example.from_dict(doc, annotations)\n",
    "        nlp.update([example], sgd=optimizer, losses=losses)\n",
    "        progress_bar.set_postfix(loss=losses[\"ner\"])\n",
    "        progress_bar.update(1)\n",
    "progress_bar.close()\n",
    "\n",
    "model_output_dir = \"./defi_finetuned_sentences_5epocs.tar\"\n",
    "nlp.to_disk(model_output_dir)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Test the trained model on the test set\n",
    "correct = 0\n",
    "total = 0\n",
    "predicted_labels = []\n",
    "true_labels = []\n",
    "for text, label in test_data:\n",
    "    doc = Doc(nlp.vocab, words=text.split(\" \"))\n",
    "    #nlp.tagger(doc)\n",
    "    #nlp.parser(doc)\n",
    "    nlp.get_pipe(\"ner\")(doc)\n",
    "    predicted_label = [word.ent_type_ for word in doc]\n",
    "    predicted_labels = predicted_labels + predicted_label\n",
    "    tmp_labels = [ent[2] for ent in label.get('entities')]\n",
    "    true_labels= true_labels + tmp_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i, value in enumerate(predicted_labels):\n",
    "    if value == '':\n",
    "        predicted_labels[i] = \"aucun\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8586"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predicted_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8586"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(true_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro F1-score: 0.9501078200824894\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "precision, recall, f1_score, _ = precision_recall_fscore_support(true_labels, predicted_labels, average='weighted')\n",
    "print(\"Micro F1-score:\", f1_score)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's remove the \"aucun\" class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "ids_without_aucun = [i for i,v in enumerate(true_labels) if v!='aucun']\n",
    "predicted_without_aucun = [predicted_labels[i] for i in ids_without_aucun]\n",
    "true_without_aucun = [true_labels[i] for i in ids_without_aucun]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 368,
     "status": "ok",
     "timestamp": 1686322310641,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "-awkW9bfpAEe",
    "outputId": "441e7a68-993c-45ee-c489-359c5c9064e2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro F1-score: 0.843483889789112\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/mnt/anaconda/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1469: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "precision, recall, f1_score, _ = precision_recall_fscore_support(true_without_aucun, predicted_without_aucun, average='weighted')\n",
    "\n",
    "print(\"Micro F1-score:\", f1_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 541
    },
    "executionInfo": {
     "elapsed": 1471,
     "status": "ok",
     "timestamp": 1686322231285,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "2lBlGTRglmOx",
    "outputId": "a57bb73e-f502-43b9-e165-fa382b5916f7"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import numpy as np\n",
    "# Assuming you have the true labels and predicted labels stored in the variables true_labels and predicted_labels\n",
    "\n",
    "# Compute the confusion matrix\n",
    "cm = confusion_matrix(true_labels, predicted_labels)\n",
    "\n",
    "# Get the unique class labels\n",
    "class_labels = np.unique(np.concatenate((true_labels, predicted_labels)))\n",
    "\n",
    "# Create a DataFrame from the confusion matrix\n",
    "df_cm = pd.DataFrame(cm, index=class_labels, columns=class_labels)\n",
    "df_cm = df_cm.loc[['aucun','geogFeat','name geogName','geogName geogFeat','geogName']] # reorder\n",
    "df_cm = df_cm[['aucun','geogFeat','name geogName','geogName geogFeat','geogName']]\n",
    "\n",
    "# Create a heatmap using seaborn\n",
    "plt.figure(figsize=(6, 4))\n",
    "sns.heatmap(df_cm, annot=True, cmap='Blues', fmt='d')\n",
    "\n",
    "plt.xlabel('Label prédit')\n",
    "plt.ylabel('Vrai label')\n",
    "plt.xticks(rotation=60, ha='right')\n",
    "plt.savefig(\"mc_spacy.png\", dpi = 300, bbox_inches = 'tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 6,
     "status": "ok",
     "timestamp": 1686310074629,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "B3kfWyOaWO0_",
    "outputId": "1f2205b3-8db2-4854-c79a-2dffb095ec15"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class: aucun\n",
      "Precision: 0.9674303063759315\n",
      "Recall: 0.9881590076120665\n",
      "F1 Score: 0.9776847977684797\n",
      "Support: 7094\n",
      "-------------------\n",
      "Class: geogFeat\n",
      "Precision: 0.7935368043087971\n",
      "Recall: 0.7647058823529411\n",
      "F1 Score: 0.7788546255506608\n",
      "Support: 578\n",
      "-------------------\n",
      "Class: geogName\n",
      "Precision: 0.9398496240601504\n",
      "Recall: 0.702247191011236\n",
      "F1 Score: 0.8038585209003215\n",
      "Support: 178\n",
      "-------------------\n",
      "Class: geogName geogFeat\n",
      "Precision: 0.9539748953974896\n",
      "Recall: 0.7238095238095238\n",
      "F1 Score: 0.8231046931407942\n",
      "Support: 315\n",
      "-------------------\n",
      "Class: name geogName\n",
      "Precision: 0.8880778588807786\n",
      "Recall: 0.8669833729216152\n",
      "F1 Score: 0.8774038461538463\n",
      "Support: 421\n",
      "-------------------\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "# true_labels = [label for _, label in test_data]\n",
    "\n",
    "precision, recall, f1_score, support = precision_recall_fscore_support(\n",
    "    true_labels, predicted_labels, average=None\n",
    ")\n",
    "\n",
    "# Print the results for each class\n",
    "for label, prec, rec, f1, supp in zip(nlp.get_pipe(\"ner\").labels, precision, recall, f1_score, support):\n",
    "    print(f\"Class: {label}\")\n",
    "    print(f\"Precision: {prec}\")\n",
    "    print(f\"Recall: {rec}\")\n",
    "    print(f\"F1 Score: {f1}\")\n",
    "    print(f\"Support: {supp}\")\n",
    "    print(\"-------------------\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prediction on test file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "executionInfo": {
     "elapsed": 254,
     "status": "ok",
     "timestamp": 1686315795306,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "DOkkcc4IL4GR"
   },
   "outputs": [],
   "source": [
    "df_test = pd.read_csv(\"./defi-textmine-2024/test.csv\")\n",
    "df_test.replace('\"', '', regex=True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "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>Token</th>\n",
       "      <th>Sentence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>01</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2.1.4.8</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>Conduites</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>sous-marines</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26168</th>\n",
       "      <td>26168</td>\n",
       "      <td>103</td>\n",
       "      <td>1162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26169</th>\n",
       "      <td>26169</td>\n",
       "      <td>/</td>\n",
       "      <td>1162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26170</th>\n",
       "      <td>26170</td>\n",
       "      <td>0</td>\n",
       "      <td>1162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26171</th>\n",
       "      <td>26171</td>\n",
       "      <td>103</td>\n",
       "      <td>1162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26172</th>\n",
       "      <td>26172</td>\n",
       "      <td>SAINT-PIERRE-ET-MIQUELON</td>\n",
       "      <td>1162</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>26173 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Id                     Token  Sentence\n",
       "0          0                        01         0\n",
       "1          1                   2.1.4.8         0\n",
       "2          2                         .         0\n",
       "3          3                 Conduites         1\n",
       "4          4              sous-marines         1\n",
       "...      ...                       ...       ...\n",
       "26168  26168                       103      1162\n",
       "26169  26169                         /      1162\n",
       "26170  26170                         0      1162\n",
       "26171  26171                       103      1162\n",
       "26172  26172  SAINT-PIERRE-ET-MIQUELON      1162\n",
       "\n",
       "[26173 rows x 3 columns]"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dot_index = [i for i,v in enumerate(df_test.Token.tolist()) if v==\".\"]\n",
    "sentence = list()\n",
    "s = 0\n",
    "for i in list(range(len(df_test.Token.tolist()))):\n",
    "    if df.Token.tolist()[i] != \".\":\n",
    "        sentence.append(s)\n",
    "    else:\n",
    "        sentence.append(s)\n",
    "        s = s+1\n",
    "df_test['Sentence'] = sentence\n",
    "df_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "predicted_labels = []\n",
    "for s in list(set(df_test.Sentence)):\n",
    "    words = df_test[df_test.Sentence == s].Token.tolist()\n",
    "    doc = Doc(nlp.vocab, words=words)\n",
    "    nlp.get_pipe(\"ner\")(doc)\n",
    "    predicted_label = [word.ent_type_ for word in doc]\n",
    "    predicted_labels = predicted_labels + predicted_label\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "for index, item in enumerate(predicted_labels):\n",
    "    if item == '':\n",
    "        predicted_labels[index] = 'aucun'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_test['Label'] = predicted_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "executionInfo": {
     "elapsed": 267,
     "status": "ok",
     "timestamp": 1686315895040,
     "user": {
      "displayName": "mehtab Alam",
      "userId": "10929531033452263395"
     },
     "user_tz": -120
    },
    "id": "ESM2ER5uOIC2"
   },
   "outputs": [],
   "source": [
    "df_test = df_test[['Id', 'Label']]\n",
    "df_test.to_csv('./defi-textmine-2024/spacy_sentences_fr_epoch5.csv', sep=',', encoding='utf-8', index=False)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyMI4zOoHOL9NpA8OltGNSn9",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}