train_classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /home/aivendil/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package punkt to /home/aivendil/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "from pymystem3 import Mystem\n",
    "from string import punctuation\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.cluster import KMeans\n",
    "import pickle\n",
    "nltk.download('stopwords')\n",
    "nltk.download('punkt')\n",
    "from bs4 import BeautifulSoup"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Препроцессинг текста\n",
    "Удаляем html теги, пунктуацию, приводим слова к инфнитиву и разбиваем строку по пробелам"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocess_text(text):\n",
    "    soup = BeautifulSoup(text)\n",
    "    text = soup.get_text()\n",
    "    mystem = Mystem()\n",
    "    russian_stopwords = stopwords.words(\"russian\")\n",
    "    tokens = mystem.lemmatize(text.lower())\n",
    "    tokens = [token for token in tokens if token not in russian_stopwords\\\n",
    "              and token != \" \" \\\n",
    "              and token.strip() not in punctuation]\n",
    "    \n",
    "    text = \" \".join(tokens)\n",
    "    \n",
    "    return tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def drop_tags(text):\n",
    "    soup = BeautifulSoup(text)\n",
    "    text = soup.get_text()\n",
    "    mystem = Mystem()\n",
    "    russian_stopwords = stopwords.words(\"russian\")\n",
    "    tokens = mystem.lemmatize(text.lower())\n",
    "    tokens = [token for token in tokens if token not in russian_stopwords\\\n",
    "              and token != \" \" \\\n",
    "              and token.strip() not in punctuation]\n",
    "    \n",
    "    text = \" \".join(tokens)\n",
    "    \n",
    "    return text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Загрузка датасета"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5525, 10)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.read_csv('extractvacancy_vacancy_t.csv')\n",
    "dataset.head()\n",
    "dataset.fillna('', inplace=True)\n",
    "dataset.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Формирование корпуса, на котором будет производится обучение."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset['full_sting'] = '{name} {duty} {qualification}'.format(name=dataset.name, duty=dataset.duty, qualification=dataset.qualification)\n",
    "# dataset.name + dataset.duty\n",
    "dataset.full_sting= dataset.full_sting.map(drop_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Обучение векторизатора"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = TfidfVectorizer(max_features=1000)\n",
    "X = vectorizer.fit_transform(dataset.full_sting)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle.dump(vectorizer, open('vectorizer_100', 'wb'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/aivendil/.local/lib/python3.7/site-packages/sklearn/cluster/k_means_.py:971: ConvergenceWarning: Number of distinct clusters (1) found smaller than n_clusters (10). Possibly due to duplicate points in X.\n",
      "  return_n_iter=True)\n"
     ]
    }
   ],
   "source": [
    "kmeans = KMeans(n_clusters=10)\n",
    "pred = kmeans.fit_predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = pd.Series(pred)\n",
    "pickle.dump(kmeans, open('kmeans_100','wb'))"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}