model_execution.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gensim\n",
    "import pymystem3\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "code = pd.read_csv('tmp/criminal_code.csv')\n",
    "df = pd.read_csv('tmp/judicial_orders.csv').rename(columns={'title':'name'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/gzhigalov/envs/diploma/lib/python3.6/site-packages/ipykernel_launcher.py:6: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version\n",
      "of pandas will change to not sort by default.\n",
      "\n",
      "To accept the future behavior, pass 'sort=False'.\n",
      "\n",
      "To retain the current behavior and silence the warning, pass 'sort=True'.\n",
      "\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "code['lemmas'] = code['lemmas'].map(eval)\n",
    "df['lemmas'] = df['lemmas'].map(eval)\n",
    "code['source'] = 'criminal_code'\n",
    "df['source'] = 'court_orders'\n",
    "df['body'] = df['intro'] + df['case'] + df['result']\n",
    "df = pd.concat([code[['name', 'number', 'lemmas','body', 'source']], \n",
    "                df[['name', 'lemmas', 'body', 'source']]], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "td = lambda row: gensim.models.doc2vec.TaggedDocument(words=row['lemmas'], tags=[row.name])\n",
    "documents = df.apply(td, axis=1).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "model = gensim.models.doc2vec.Doc2Vec()\n",
    "model.build_vocab(documents)\n",
    "model.train(documents, total_examples=model.corpus_count, epochs=3000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('tmp/doc2vec/code_and_court_orders.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = gensim.models.doc2vec.Doc2Vec.load('tmp/doc2vec/code_and_court_orders.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "stem = pymystem3.Mystem(entire_input=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "text = \"\"\"Группой из 10 человек было совершено нападение на сберегательную кассу в момент её закрытия\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/gzhigalov/envs/diploma/lib/python3.6/site-packages/gensim/matutils.py:737: FutureWarning: Conversion of the second argument of issubdtype from `int` to `np.signedinteger` is deprecated. In future, it will be treated as `np.int64 == np.dtype(int).type`.\n",
      "  if np.issubdtype(vec.dtype, np.int):\n"
     ]
    }
   ],
   "source": [
    "new_vector = model.infer_vector(stem.lemmatize(text))\n",
    "sims = model.docvecs.most_similar([new_vector], topn=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sqlite3\n",
    "conn = sqlite3.connect('data.db')\n",
    "df[df['source'] == 'criminal_code'].reset_index()[['index', 'name', 'number', 'body']].to_sql('data', con=conn, index=False, if_exists='replace')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_code_info(sims, n=5):\n",
    "    cursor = conn.execute(f\"SELECT * FROM data WHERE `index` in ({','.join([str(x[0]) for x in sims])})\")\n",
    "    data = cursor.fetchmany(n)\n",
    "    cursor.close()\n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.4224820137023926 Общие условия уголовной ответственности\n"
     ]
    }
   ],
   "source": [
    "for i, p in sims:\n",
    "    if df.loc[i, 'source'] == 'criminal_code':\n",
    "        print(p, df.loc[i, 'name'])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "diploma",
   "language": "python",
   "name": "diploma"
  },
  "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
}