nlpWorkTools.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6eeb8cd6",
   "metadata": {},
   "source": [
    "中文nlpTools推荐：ltp(性能好)和jiagu(功能多),Synonyms(相似度常用)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "departmental-boutique",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# 字符串数量统计"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "employed-longitude",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-25T12:25:32.048868Z",
     "start_time": "2021-04-25T12:25:32.036868Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "统计结果为:\n",
      "{'p': 12, 'n': 13, '<': 8, '*': 10, '?': 12, '@': 10, ';': 10, 'f': 10, 'V': 11, '%': 13, '$': 13, 'm': 12, 'A': 9, 'T': 6, 'x': 7, '!': 7, 'O': 4, 'y': 14, '4': 12, '>': 12, 'l': 7, 'P': 15, 'E': 9, '-': 4, 'j': 9, '\\\\': 12, 'S': 10, '3': 9, '&': 11, 'z': 11, 's': 15, 'U': 6, '5': 17, '6': 10, '\"': 15, 'u': 15, ')': 7, ']': 12, 'b': 5, '|': 8, '9': 13, 'J': 14, 'F': 16, 'q': 12, '{': 8, 'Y': 10, 'C': 13, 'X': 7, ':': 15, '.': 12, 'k': 16, 'o': 5, 'c': 15, '_': 10, 'B': 16, 'M': 6, 'w': 23, 'R': 16, '1': 17, 'd': 15, '0': 12, 'D': 6, 'W': 13, '=': 13, 'a': 9, '~': 14, '7': 8, 'K': 10, 'e': 8, '8': 9, 'L': 5, '[': 9, 'i': 13, 'N': 10, 'r': 12, ',': 7, \"'\": 7, '}': 9, '(': 9, 'h': 13, '`': 13, 'v': 12, 'Q': 7, 'g': 12, 't': 15, 'G': 9, '/': 11, '^': 9, '2': 9, '+': 6, 'Z': 3, '#': 6, 'I': 9, 'H': 12}\n"
     ]
    }
   ],
   "source": [
    "# 下面的代码首先生成包含1000个随机字符的字符串\n",
    "# 然后统计每个字符的出现次数\n",
    "import string\n",
    "import random\n",
    "x = string.ascii_letters + string.digits + string.punctuation\n",
    "# print('\\n可使用的字符集:\\n', len(x))   \n",
    "y = [random.choice(x) for i in range(1000)]\n",
    "# print(y)\n",
    "z = ''.join(y)\n",
    "d = dict()\n",
    "# 遍历生成并且统计数量\n",
    "for ch in z:\n",
    "\td[ch] = d.get(ch, 0) + 1  #千万不要使用d[ch] =d[ch] + 1，否则字典没有ch的key，出错\n",
    "# print('\\n产生字符串为:\\n',z,sep='')\n",
    "print('\\n统计结果为:\\n',d,sep='')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "stock-capital",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# 句子相似度计算"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "considerable-discipline",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-25T16:28:03.951035Z",
     "start_time": "2021-04-25T16:28:03.941033Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8571428571428571\n",
      "0.3157894736842105\n"
     ]
    }
   ],
   "source": [
    "# 计算句子相似度\n",
    "import jieba\n",
    "\n",
    "from difflib import SequenceMatcher#导入库\n",
    "def similarity(a, b):\n",
    "    return SequenceMatcher(None, a, b).ratio()#引用ratio方法，返回序列相似性的度量\n",
    "print(similarity((6,6,6,6,6,6,6), (6,6,6,6,6,6,7)))\n",
    "\n",
    "def Jaccrad(model, reference):  # terms_reference为源句子，terms_model为候选句子\n",
    "    terms_reference = jieba.cut(reference)  # 默认精准模式\n",
    "    terms_model = jieba.cut(model)\n",
    "    grams_reference = set(terms_reference)  # 去重；如果不需要就改为list\n",
    "    grams_model = set(terms_model)\n",
    "    temp = 0\n",
    "    for i in grams_reference:\n",
    "        if i in grams_model:\n",
    "            temp = temp + 1\n",
    "    fenmu = len(grams_model) + len(grams_reference) - temp  # 并集\n",
    "    jaccard_coefficient = float(temp / fenmu)  # 交集\n",
    "    return jaccard_coefficient\n",
    "\n",
    "\n",
    "a = \"飞控系统主要组成以及交联设备:飞行传感器系统(子系统)的说明是什么？\"\n",
    "b = \"飞行传感器系统(子系统)\"\n",
    "jaccard_coefficient = Jaccrad(a, b)\n",
    "print(jaccard_coefficient)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "jewish-rendering",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "# 在句子中找到p然后分割后计算相似度来找s\n",
    "import pymongo\n",
    "client=pymongo.MongoClient('mongodb://localhost:27017')\n",
    "db = client.cndbpedia03\n",
    "collection = db.triples\n",
    "import bson\n",
    "import jieba\n",
    "\n",
    "def Jaccrad(model, reference):  # terms_reference为源句子，terms_model为候选句子,计算句子相似度\n",
    "    terms_reference = jieba.cut(reference)  # 默认精准模式\n",
    "    terms_model = jieba.cut(model)\n",
    "    grams_reference = set(terms_reference)  # 去重；如果不需要就改为list\n",
    "    grams_model = set(terms_model)\n",
    "    temp = 0\n",
    "    for i in grams_reference:\n",
    "        if i in grams_model:\n",
    "            temp = temp + 1\n",
    "    fenmu = len(grams_model) + len(grams_reference) - temp  # 并集\n",
    "    jaccard_coefficient = float(temp / fenmu)  # 交集\n",
    "    return jaccard_coefficient\n",
    "\n",
    "\n",
    "a = \"飞控系统主要组成以及交联设备:飞行传感器系统(子系统)的说明是什么？\"\n",
    "\n",
    "str_input = \"飞控系统的说明是什么\"\n",
    "# jaccard_coefficient = Jaccrad(str_input, b)\n",
    "\n",
    "results = collection.find({})\n",
    "for result in results:\n",
    "    if result['p'] in str_input:\n",
    "        str_split = str_input.split(result['p'])\n",
    "        str_p = result['p']\n",
    "        break\n",
    "str_s = str_split[0]\n",
    "print(str_s)\n",
    "print(str_p)\n",
    "\n",
    "# jaccard_coefficient = Jaccrad(str_s, b)\n",
    "results = collection.find({\"p\":str_p})\n",
    "for result in results:\n",
    "    jacc_score = Jaccrad(str_s,result['s'])\n",
    "    # print(jacc_score)\n",
    "    if jacc_score > 0.1:\n",
    "        print(result['s'],result['p'],result['o'])\n",
    "print(\"--------------------------------------------\")\n",
    "# print(collection.find({\"p\": \"用途\"}).count()) \n",
    "\n",
    "# jaccard_coefficient = Jaccrad(a, b)\n",
    "# print(jaccard_coefficient) //一个系数"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "unsigned-circular",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# ⭐哈工大LTP-亲测中文处理简单好用"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "441ce137",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T14:21:48.840929Z",
     "start_time": "2021-05-04T14:21:33.361643Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "#pip install ltp\n",
    "from ltp import LTP\n",
    "ltp = LTP()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "38b4e773",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T14:32:43.747019Z",
     "start_time": "2021-05-04T14:32:42.216911Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "seg:  [['我', '在', '复旦', '大学', '马路边', '看见', '了', '计算机', '学院', '肖仰华', '老师', '。']]\n",
      "pos:  [['r', 'p', 'nz', 'n', 'n', 'v', 'u', 'n', 'n', 'nh', 'n', 'wp']]\n",
      "ner:  [[('Ni', 2, 3), ('Ni', 7, 8), ('Nh', 9, 9)]]\n",
      "srl:  [[[], [], [], [], [], [('A0', 0, 0), ('ARGM-LOC', 1, 4), ('A1', 7, 10)], [], [], [], [], [], []]]\n",
      "dep [[(1, 6, 'SBV'), (2, 6, 'ADV'), (3, 4, 'ATT'), (4, 5, 'ATT'), (5, 2, 'POB'), (6, 0, 'HED'), (7, 6, 'RAD'), (8, 9, 'ATT'), (9, 11, 'ATT'), (10, 11, 'ATT'), (11, 6, 'VOB'), (12, 6, 'WP')]]\n",
      "sdp [[(1, 6, 'AGT'), (2, 5, 'mRELA'), (3, 4, 'FEAT'), (4, 5, 'FEAT'), (5, 6, 'LOC'), (6, 0, 'Root'), (7, 6, 'mDEPD'), (8, 9, 'FEAT'), (9, 11, 'FEAT'), (10, 11, 'FEAT'), (11, 6, 'DATV'), (12, 6, 'mPUNC')]]\n"
     ]
    }
   ],
   "source": [
    "from ltp import LTP\n",
    "\n",
    "ltp = LTP()  # 默认加载 Small 模型\n",
    "seg, hidden = ltp.seg([\"我在复旦大学马路边看见了计算机学院肖仰华老师。\"])\n",
    "print('seg: ',seg)\n",
    "# print(hidden)\n",
    "pos = ltp.pos(hidden)\n",
    "print('pos: ',pos)\n",
    "ner = ltp.ner(hidden)\n",
    "print('ner: ',ner)\n",
    "srl = ltp.srl(hidden)\n",
    "print('srl: ',srl)\n",
    "dep = ltp.dep(hidden)\n",
    "print('dep',dep)\n",
    "sdp = ltp.sdp(hidden)\n",
    "print('sdp',sdp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "practical-reputation",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T14:23:15.323728Z",
     "start_time": "2021-05-04T14:23:14.745648Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['他叫汤姆去拿外衣。', '汤姆生病了。', '他去了医院。']\n",
      "[['r', 'v', 'nh', 'v', 'v', 'n', 'wp']]\n",
      "[['我', '在', '复旦', '大学', '马路边', '看见', '了', '计算机', '学院', '肖仰华', '老师', '。']]\n",
      "[[('Ni', 2, 3), ('Ni', 7, 8), ('Nh', 9, 9)]]\n",
      "Ni : 复旦大学\n",
      "[[[], [('A0', 0, 0), ('A1', 2, 2), ('A2', 3, 5)], [], [], [('A0', 2, 2), ('A1', 5, 5)], [], []]]\n",
      "[[(1, [('A0', 0, 0), ('A1', 2, 2), ('A2', 3, 5)]), (4, [('A0', 2, 2), ('A1', 5, 5)])]]\n",
      "[[(1, 2, 'SBV'), (2, 0, 'HED'), (3, 2, 'DBL'), (4, 5, 'ADV'), (5, 2, 'VOB'), (6, 5, 'VOB'), (7, 2, 'WP')]]\n",
      "[[(1, 2, 'AGT'), (2, 0, 'Root'), (3, 2, 'DATV'), (3, 4, 'AGT'), (3, 5, 'AGT'), (4, 2, 'eSUCC'), (5, 2, 'eSUCC'), (5, 4, 'eSUCC'), (6, 5, 'PAT'), (7, 2, 'mPUNC')]]\n"
     ]
    }
   ],
   "source": [
    "#分句\n",
    "sents = ltp.sent_split([\"他叫汤姆去拿外衣。汤姆生病了。他去了医院。\"])\n",
    "print(sents)\n",
    "# user_dict.txt 是词典文件， max_window是最大前向分词窗口,用户自定义词典\n",
    "ltp.init_dict(path=\"user_dict.txt\", max_window=4)\n",
    "# 也可以在代码中添加自定义的词语\n",
    "ltp.add_words(words=[\"负重前行\", \"长江大桥\"], max_window=4)\n",
    "#分词\n",
    "segment,_= ltp.seg([\"他叫汤姆去拿外衣。\"])\n",
    "segment, hidden = ltp.seg([\"他/叫/汤姆/去/拿/外衣/。\".split('/')], is_preseged=True)\n",
    "# print(hidden)\n",
    "#词性标注\n",
    "pos = ltp.pos(hidden)\n",
    "print(pos)\n",
    "# NER\n",
    "seg, hidden = ltp.seg([\"我在复旦大学马路边看见了计算机学院肖仰华老师。\"])\n",
    "print(seg)\n",
    "ner = ltp.ner(hidden)\n",
    "print(ner)\n",
    "# [['他', '叫', '汤姆', '去', '拿', '外衣', '。']]\n",
    "# [[('Nh', 2, 2)]]\n",
    "tag, start, end = ner[0][0]\n",
    "print(tag,\":\", \"\".join(seg[0][start:end + 1]))\n",
    "# Nh : 汤姆\n",
    "# 语义角色标注\n",
    "seg, hidden = ltp.seg([\"他叫汤姆去拿外衣。\"])\n",
    "srl1 = ltp.srl(hidden)\n",
    "print(srl1)\n",
    "srl2 = ltp.srl(hidden, keep_empty=False)\n",
    "print(srl2)\n",
    "# 依存句法分析:在依存句法当中，虚节点ROOT占据了0位置，因此节点的下标从1开始。\n",
    "seg, hidden = ltp.seg([\"他叫汤姆去拿外衣。\"])\n",
    "dep = ltp.dep(hidden)\n",
    "print(dep)\n",
    "# 语义依存分析(树)\n",
    "seg, hidden = ltp.seg([\"他叫汤姆去拿外衣。\"])\n",
    "sdp = ltp.sdp(hidden)\n",
    "print(sdp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "motivated-worry",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T14:34:48.400380Z",
     "start_time": "2021-05-04T14:34:47.055138Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['飞控', '系统', '主要', '组成', '以及', '交联', '设备', ':', '飞行', '传感', '系统', '（', '子系统', '）']]\n",
      "[[]]\n"
     ]
    }
   ],
   "source": [
    "# NER\n",
    "from ltp import LTP\n",
    "ltp = LTP()\n",
    "seg, hidden = ltp.seg([\"飞控系统主要组成以及交联设备:飞行传感系统（子系统）\"])\n",
    "print(seg)\n",
    "ner = ltp.ner(hidden)\n",
    "print(ner)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "spiritual-houston",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# 甲言-古汉语信息处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "occasional-knock",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-01T15:42:03.998819Z",
     "start_time": "2021-04-01T15:42:02.876618Z"
    },
    "hidden": true,
    "scrolled": true
   },
   "outputs": [
    {
     "ename": "FileNotFoundError",
     "evalue": "[Errno 2] No such file or directory: '庄子.txt'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mFileNotFoundError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-38-f71b56be7541>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mconstructor\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPMIEntropyLexiconConstructor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mlexicon\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconstructor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstruct_lexicon\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'庄子.txt'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0mconstructor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlexicon\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'庄子词库.csv'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/Deep/lib/python3.6/site-packages/jiayan/lexicon/pmi_entropy_constructor.py\u001b[0m in \u001b[0;36mconstruct_lexicon\u001b[0;34m(self, data_file)\u001b[0m\n\u001b[1;32m     58\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mconstruct_lexicon\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata_file\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 60\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbuild_trie_trees\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_file\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     61\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcompute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m         \u001b[0mlexicon\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/Deep/lib/python3.6/site-packages/jiayan/lexicon/pmi_entropy_constructor.py\u001b[0m in \u001b[0;36mbuild_trie_trees\u001b[0;34m(self, data_file)\u001b[0m\n\u001b[1;32m     69\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     70\u001b[0m         \u001b[0mstart\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 71\u001b[0;31m         \u001b[0;32mfor\u001b[0m \u001b[0mtext\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtext_iterator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_file\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     72\u001b[0m             \u001b[0mlength\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     73\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlength\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/Deep/lib/python3.6/site-packages/jiayan/utils.py\u001b[0m in \u001b[0;36mtext_iterator\u001b[0;34m(data_file, keep_punc)\u001b[0m\n\u001b[1;32m     15\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mtext_iterator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkeep_punc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m     \u001b[0;34m\"\"\" A help function to provide clean zh char lines of a given file. \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 17\u001b[0;31m     \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'r'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mencoding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'utf-8'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     18\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mline\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     19\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mseg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mline\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstrip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msplit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mFileNotFoundError\u001b[0m: [Errno 2] No such file or directory: '庄子.txt'"
     ]
    }
   ],
   "source": [
    "from jiayan import PMIEntropyLexiconConstructor\n",
    "\n",
    "constructor = PMIEntropyLexiconConstructor()\n",
    "lexicon = constructor.construct_lexicon('庄子.txt')\n",
    "constructor.save(lexicon, '庄子词库.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "103744a6",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# ⭐jiagu自然语言处理工具 - 方便快捷 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9558abb",
   "metadata": {
    "hidden": true
   },
   "source": [
    "Jiagu使用大规模语料训练而成。将提供中文分词、词性标注、命名实体识别、情感分析、知识图谱关系抽取、关键词抽取、文本摘要、新词发现、情感分析、文本聚类等常用自然语言处理功能。参考了各大工具优缺点制作，将Jiagu回馈给大家。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "087d2eac",
   "metadata": {
    "hidden": true
   },
   "source": [
    "1. 快速上手：分词、词性标注、命名实体识别"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "introductory-sender",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:04:39.860326Z",
     "start_time": "2021-05-04T04:04:39.838324Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['飞控', '系统', '主要', '组成', '以及', '交', '联', '设备', ':', '飞行', '传感', '系统', '（', '子系统', '）']\n",
      "['n', 'n', 'a', 'v', 'c', 'v', 'v', 'n', 'w', 'v', 'n', 'n', 'w', 'n', 'w']\n",
      "['O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O']\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "#jiagu.init() # 可手动初始化，也可以动态初始化\n",
    "\n",
    "text = '飞控系统主要组成以及交联设备:飞行传感系统（子系统）'\n",
    "\n",
    "words = jiagu.seg(text) # 分词\n",
    "print(words)\n",
    "\n",
    "pos = jiagu.pos(words) # 词性标注\n",
    "print(pos)\n",
    "\n",
    "ner = jiagu.ner(words) # 命名实体识别\n",
    "print(ner)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a4d7610",
   "metadata": {
    "hidden": true
   },
   "source": [
    "2. 中文分词"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "72a674b2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:07:18.114846Z",
     "start_time": "2021-05-04T04:07:18.098847Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['汉服', '和', '服装', '、', '维基', '图谱']\n",
      "['汉服和服装', '、', '维基', '图谱']\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "text = '汉服和服装、维基图谱'\n",
    "\n",
    "words = jiagu.seg(text)\n",
    "print(words)\n",
    "\n",
    "# jiagu.load_userdict('dict/user.dict') # 加载自定义字典，支持字典路径、字典列表形式。\n",
    "jiagu.load_userdict(['汉服和服装'])\n",
    "\n",
    "words = jiagu.seg(text) # 自定义分词，字典分词模式有效\n",
    "print(words)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7eee1d8a",
   "metadata": {
    "hidden": true
   },
   "source": [
    "3. 知识图谱关系抽取\n",
    "\n",
    "仅用于测试用，可以pip3 install jiagu==0.1.8，只能使用百科的描述进行测试。效果更佳的后期将会开放api。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "367edd5c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:27:31.797505Z",
     "start_time": "2021-05-04T04:27:31.760059Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['姚明', '出生日期', '1980年9月12日'], ['姚明', '出生地', '上海市徐汇区'], ['姚明', '祖籍', '江苏省苏州市吴江区震泽镇']]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "list"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "# 吻别是由张学友演唱的一首歌曲。\n",
    "# 《盗墓笔记》是2014年欢瑞世纪影视传媒股份有限公司出品的一部网络季播剧，改编自南派三叔所著的同名小说，由郑保瑞和罗永昌联合导演，李易峰、杨洋、唐嫣、刘天佐、张智尧、魏巍等主演。\n",
    "\n",
    "text = '姚明1980年9月12日出生于上海市徐汇区，祖籍江苏省苏州市吴江区震泽镇，前中国职业篮球运动员，司职中锋，现任中职联公司董事长兼总经理。'\n",
    "knowledge = jiagu.knowledge(text)\n",
    "print(knowledge)\n",
    "type(knowledge)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "caec18c7",
   "metadata": {
    "hidden": true
   },
   "source": [
    "4. 关键词提取"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "bdc003b0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:18:04.759244Z",
     "start_time": "2021-05-04T04:18:04.724230Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['.', '工程', '万', '造林', '绿化']\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "text = '''\n",
    "该研究主持者之一、波士顿大学地球与环境科学系博士陈池（音）表示，“尽管中国和印度国土面积仅占全球陆地的9%，但两国为这一绿化过程贡献超过三分之一。考虑到人口过多的国家一般存在对土地过度利用的问题，这个发现令人吃惊。”\n",
    "NASA埃姆斯研究中心的科学家拉玛·内曼尼（Rama Nemani）说，“这一长期数据能让我们深入分析地表绿化背后的影响因素。我们一开始以为，植被增加是由于更多二氧化碳排放，导致气候更加温暖、潮湿，适宜生长。”\n",
    "“MODIS的数据让我们能在非常小的尺度上理解这一现象，我们发现人类活动也作出了贡献。”\n",
    "NASA文章介绍，在中国为全球绿化进程做出的贡献中，有42%来源于植树造林工程，对于减少土壤侵蚀、空气污染与气候变化发挥了作用。\n",
    "据观察者网过往报道，2017年我国全国共完成造林736.2万公顷、森林抚育830.2万公顷。其中，天然林资源保护工程完成造林26万公顷，退耕还林工程完成造林91.2万公顷。京津风沙源治理工程完成造林18.5万公顷。三北及长江流域等重点防护林体系工程完成造林99.1万公顷。完成国家储备林建设任务68万公顷。\n",
    "'''\t\t\t\t\n",
    "\n",
    "keywords = jiagu.keywords(text, 5) # 关键词\n",
    "print(keywords)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a11d49c4",
   "metadata": {
    "hidden": true
   },
   "source": [
    "5. 文本摘要"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "48f3a9ea",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:19:38.879471Z",
     "start_time": "2021-05-04T04:19:38.866472Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['”NASA文章介绍，在中国为全球绿化进程做出的贡献中，有42%来源于植树造林工程，对于减少土壤侵蚀、空气污染与气候变化发挥了作用。']\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "# fin = open('input.txt', 'r')\n",
    "# text = fin.read()\n",
    "# fin.close()\n",
    "\n",
    "text = '''该研究主持者之一、波士顿大学地球与环境科学系博士陈池（音）表示，“尽管中国和印度国土面积仅占全球陆地的9%，但两国为这一绿化过程贡献超过三分之一。考虑到人口过多的国家一般存在对土地过度利用的问题，这个发现令人吃惊。”\n",
    "NASA埃姆斯研究中心的科学家拉玛·内曼尼（Rama Nemani）说，“这一长期数据能让我们深入分析地表绿化背后的影响因素。我们一开始以为，植被增加是由于更多二氧化碳排放，导致气候更加温暖、潮湿，适宜生长。”\n",
    "“MODIS的数据让我们能在非常小的尺度上理解这一现象，我们发现人类活动也作出了贡献。”\n",
    "NASA文章介绍，在中国为全球绿化进程做出的贡献中，有42%来源于植树造林工程，对于减少土壤侵蚀、空气污染与气候变化发挥了作用。\n",
    "据观察者网过往报道，2017年我国全国共完成造林736.2万公顷、森林抚育830.2万公顷。其中，天然林资源保护工程完成造林26万公顷，退耕还林工程完成造林91.2万公顷。京津风沙源治理工程完成造林18.5万公顷。三北及长江流域等重点防护林体系工程完成造林99.1万公顷。完成国家储备林建设任务68万公顷。'''\n",
    "\n",
    "summarize = jiagu.summarize(text, 1) # 摘要\n",
    "print(summarize)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3405ce65",
   "metadata": {
    "hidden": true
   },
   "source": [
    "6. 新词发现"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "3c6c764c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:20:18.872495Z",
     "start_time": "2021-05-04T04:20:18.857484Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "import jiagu\n",
    "\n",
    "jiagu.findword('input.txt', 'output.txt')\n",
    "# 根据文本，利用信息熵做新词发现。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "318e1d0a",
   "metadata": {
    "hidden": true
   },
   "source": [
    "7. 情感分析"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b42d165e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:20:49.163746Z",
     "start_time": "2021-05-04T04:20:49.140765Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('negative', 0.9957030885091285)\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "text = '很讨厌还是个懒鬼'\n",
    "sentiment = jiagu.sentiment(text)\n",
    "print(sentiment)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59f49784",
   "metadata": {
    "hidden": true
   },
   "source": [
    "8. 文本聚类"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "f9d32579",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-04T04:21:21.475946Z",
     "start_time": "2021-05-04T04:21:21.458944Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: ['百度深度学习中文情感分析工具Senta试用及在线测试', '深度学习实践：从零开始做电影评论文本情感分析', '将不同长度的句子用BERT预训练模型编码，映射到一个固定长度的向量上', '现在可以快速测试一下spaCy的相关功能，我们以英文数据为例，spaCy目前主要支持英文和德文'], 1: ['自然语言处理工具包spaCy介绍', '情感分析是自然语言处理里面一个热门话题'], 2: ['AI Challenger 2018 文本挖掘类竞赛相关解决方案及代码汇总', 'BERT相关论文、文章和代码资源汇总']}\n"
     ]
    }
   ],
   "source": [
    "import jiagu\n",
    "\n",
    "docs = [\n",
    "        \"百度深度学习中文情感分析工具Senta试用及在线测试\",\n",
    "        \"情感分析是自然语言处理里面一个热门话题\",\n",
    "        \"AI Challenger 2018 文本挖掘类竞赛相关解决方案及代码汇总\",\n",
    "        \"深度学习实践：从零开始做电影评论文本情感分析\",\n",
    "        \"BERT相关论文、文章和代码资源汇总\",\n",
    "        \"将不同长度的句子用BERT预训练模型编码，映射到一个固定长度的向量上\",\n",
    "        \"自然语言处理工具包spaCy介绍\",\n",
    "        \"现在可以快速测试一下spaCy的相关功能，我们以英文数据为例，spaCy目前主要支持英文和德文\"\n",
    "    ]\n",
    "cluster = jiagu.text_cluster(docs)\n",
    "print(cluster)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e045473b",
   "metadata": {
    "hidden": true
   },
   "source": [
    "词性标注说明：\n",
    "\n",
    "n　　　普通名词\n",
    "nt　 　时间名词\n",
    "nd　 　方位名词\n",
    "nl　 　处所名词\n",
    "nh　 　人名\n",
    "nhf　　姓\n",
    "nhs　　名\n",
    "ns　 　地名\n",
    "nn 　　族名\n",
    "ni 　　机构名\n",
    "nz 　　其他专名\n",
    "v　　 动词\n",
    "vd　　趋向动词\n",
    "vl　　联系动词\n",
    "vu　　能愿动词\n",
    "a　 　形容词\n",
    "f　 　区别词\n",
    "m　 　数词　　\n",
    "q　 　量词\n",
    "d　 　副词\n",
    "r　 　代词\n",
    "p　　 介词\n",
    "c　 　连词\n",
    "u　　 助词\n",
    "e　 　叹词\n",
    "o　 　拟声词\n",
    "i　 　习用语\n",
    "j　　 缩略语\n",
    "h　　 前接成分\n",
    "k　　 后接成分\n",
    "g　 　语素字\n",
    "x　 　非语素字\n",
    "w　 　标点符号\n",
    "ws　　非汉字字符串\n",
    "wu　　其他未知的符号"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "635f8803",
   "metadata": {
    "hidden": true
   },
   "source": [
    "命名实体说明（采用BIO标记方式）:\n",
    "\n",
    "B-PER、I-PER   人名\n",
    "\n",
    "B-LOC、I-LOC   地名\n",
    "\n",
    "B-ORG、I-ORG   机构名"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "expanded-senior",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# ⭐Synonyms近义词和词向量获取"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "linear-provider",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-17T12:54:47.573356Z",
     "start_time": "2021-04-17T12:52:57.469993Z"
    },
    "hidden": true,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "smart_open library not found; falling back to local-filesystem-only\n",
      "[jieba] default dict file path ..\\data\\vocab.txt\n",
      "[jieba] default dict file path ..\\data\\vocab.txt\n",
      "[jieba] load default dict ..\\data\\vocab.txt ...\n",
      "[jieba] load default dict ..\\data\\vocab.txt ...\n",
      ">> Synonyms load wordseg dict [D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\synonyms\\data\\vocab.txt] ... \n",
      ">> Synonyms on loading stopwords [D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\synonyms\\data\\stopwords.txt] ...\n",
      "[Synonyms] on loading vectors [D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\synonyms\\data\\words.vector.gz] ...\n"
     ]
    }
   ],
   "source": [
    "import synonyms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "complicated-necessity",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-17T12:55:14.920589Z",
     "start_time": "2021-04-17T12:55:14.913585Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "飞控系统:  ([], [])\n"
     ]
    }
   ],
   "source": [
    "print(synonyms.nearby(\"飞控系统\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "manufactured-crisis",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-11T09:02:16.769138Z",
     "start_time": "2021-04-11T09:02:16.753138Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid token (<ipython-input-2-62fd3ea4ada2>, line 9)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-2-62fd3ea4ada2>\"\u001b[1;36m, line \u001b[1;32m9\u001b[0m\n\u001b[1;33m    095, 0.525344, 0.524009, 0.523101, 0.516046])\u001b[0m\n\u001b[1;37m      ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid token\n"
     ]
    }
   ],
   "source": [
    "import synonyms\n",
    "print(\"人脸: \", synonyms.nearby(\"人脸\"))\n",
    "print(\"识别: \", synonyms.nearby(\"识别\"))\n",
    "print(\"NOT_EXIST: \", synonyms.nearby(\"NOT_EXIST\"))\n",
    "\n",
    "synonyms.nearby(人脸, 10) = (\n",
    "    [\"图片\", \"图像\", \"通过观察\", \"数字图像\", \"几何图形\", \"脸部\", \"图象\", \"放大镜\", \"面孔\", \"Mii\"],\n",
    "    [0.597284, 0.580373, 0.568486, 0.535674, 0.531835, 0.530\n",
    "095, 0.525344, 0.524009, 0.523101, 0.516046])\n",
    "#\n",
    "两个句子的相似度比较\n",
    "\n",
    "    sen1 = \"发生历史性变革\"\n",
    "    sen2 = \"发生历史性变革\"\n",
    "    r = synonyms.compare(sen1, sen2, seg=True)\n",
    "    \n",
    "# 以友好的方式打印近义词，方便调试，display(WORD [, SIZE])调用了 synonyms#nearby 方法。\n",
    "\n",
    "synonyms.display(\"飞机\")\n",
    "# '飞机'近义词：\n",
    "#   1. 飞机:1.0\n",
    "#   2. 直升机:0.8423391\n",
    "#   3. 客机:0.8393003\n",
    "#   4. 滑翔机:0.7872388\n",
    "#   5. 军用飞机:0.7832081\n",
    "#   6. 水上飞机:0.77857226\n",
    "#   7. 运输机:0.7724742\n",
    "#   8. 航机:0.7664748\n",
    "#   9. 航空器:0.76592904\n",
    "#   10. 民航机:0.74209654\n",
    "\n",
    "\n",
    "#获得一个词语的向量，该向量为 numpy 的 array，当该词语是未登录词时，抛出 KeyError 异常。\n",
    "synonyms.v(\"飞机\")\n",
    "array([-2.412167  ,  2.2628384 , -7.0214124 ,  3.9381874 ,  0.8219283 ,\n",
    "       -3.2809453 ,  3.8747153 , -5.217062  , -2.2786229 , -1.2572327 ],\n",
    "      dtype=float32)\n",
    "# 中文分词\n",
    "\n",
    "synonyms.seg(\"中文近义词工具包\")\n",
    "# 分词结果，由两个 list 组成的元组，分别是单词和对应的词性。\n",
    "\n",
    "(['中文', '近义词', '工具包'], ['nz', 'n', 'n'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "falling-respect",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# fasttext"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "happy-section",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-12T03:38:56.722740Z",
     "start_time": "2021-04-12T03:38:53.148817Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning : `load_model` does not return WordVectorModel or SupervisedModel any more, but a `FastText` object which is very similar.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "准确率: (0, nan, nan)\n"
     ]
    }
   ],
   "source": [
    "# coidng:utf-8\n",
    "import fasttext\n",
    " \n",
    " \n",
    "# 模型的训练\n",
    "def train():\n",
    "    model = fasttext.train_supervised(\"train.txt\", lr=0.1, dim=100, epoch=500, word_ngrams=4, loss='softmax')\n",
    "    model.save_model(\"model_file.bin\")\n",
    " \n",
    " \n",
    "# 模型的测试\n",
    "def test_1():\n",
    "    classifier = fasttext.load_model(\"model_file.bin\")\n",
    "    # 测试模型\n",
    "    result = classifier.test(\"test.txt\")\n",
    "    print(\"准确率:\", result)\n",
    "    f1 = open('prediction.txt', 'w', encoding='utf-8')\n",
    "    with open('test.txt', encoding='utf-8') as fp:\n",
    "        for line in fp.readlines():\n",
    "            line = line.strip()\n",
    "            if line == '':\n",
    "                continue\n",
    "            f1.write(line + '\\t#####\\t' + classifier.predict([line])[0][0][0] + '\\n')\n",
    "    f1.close()\n",
    " \n",
    " \n",
    "if __name__ == '__main__':\n",
    "    train()\n",
    "    test_1()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "powered-criterion",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-12T03:39:22.459759Z",
     "start_time": "2021-04-12T03:39:19.354536Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "import fasttext\n",
    "trainDataFile = 'train.txt'\n",
    " \n",
    "classifier = fasttext.train_supervised(\n",
    "    input = trainDataFile,\n",
    "    label_prefix = '__label__',\n",
    "    dim = 256,\n",
    "    epoch = 50,\n",
    "    lr = 1,\n",
    "    lr_update_rate = 50,\n",
    "    min_count = 3,\n",
    "    loss = 'softmax',\n",
    "    word_ngrams = 2,\n",
    "    bucket = 1000000)\n",
    "classifier.save_model(\"Model.bin\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "opponent-speaking",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-12T03:40:35.086534Z",
     "start_time": "2021-04-12T03:40:33.547620Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "测试集上数据量 0\n",
      "测试集上准确率 nan\n",
      "测试集上召回率 nan\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning : `load_model` does not return WordVectorModel or SupervisedModel any more, but a `FastText` object which is very similar.\n"
     ]
    }
   ],
   "source": [
    "testDataFile = 'test.txt'\n",
    " \n",
    "classifier = fasttext.load_model('Model.bin') \n",
    " \n",
    "result = classifier.test(testDataFile)\n",
    "print('测试集上数据量', result[0])\n",
    "print('测试集上准确率', result[1])\n",
    "print('测试集上召回率', result[2])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7da68e6",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# lightNLP"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba29d558",
   "metadata": {
    "hidden": true
   },
   "source": [
    "是要train的，不是纯工具包\n",
    "使用说明在： https://lightnlp-cookbook.readthedocs.io/zh_CN/latest/preface.html"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14fb5b13",
   "metadata": {
    "hidden": true
   },
   "source": [
    "## 功能"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8e8b2563",
   "metadata": {
    "hidden": true
   },
   "source": [
    "    序列标注，sl\n",
    "        中文分词，cws\n",
    "        命名实体识别，ner\n",
    "        词性标注，pos\n",
    "        语义角色标注， srl\n",
    "    结构分析，sp\n",
    "        基于图的依存句法分析，gdp\n",
    "        基于转移的依存句法分析， tdp\n",
    "    句子关系，sr\n",
    "        语句相似度，ss\n",
    "        文本蕴含，te\n",
    "    文本分类，tc\n",
    "        关系抽取，re\n",
    "        情感极性分析，sa\n",
    "    文本生成，tg\n",
    "        语言模型，lm\n",
    "        聊天机器人，cb\n",
    "        机器翻译，mt\n",
    "        文本摘要，ts\n",
    "    词向量，we\n",
    "        词袋模型，cbow\n",
    "            base\n",
    "            hierarchical_softmax\n",
    "            negative_sampling\n",
    "        跳字模型，skip_gram\n",
    "        base\n",
    "        hierarchical_softmax\n",
    "        negative_sampling\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3851479",
   "metadata": {
    "hidden": true
   },
   "source": [
    "## 所用模型"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e74a9fa",
   "metadata": {
    "hidden": true
   },
   "source": [
    "所用模型：\n",
    "\n",
    "    ner: BiLstm-Crf\n",
    "    cws: BiLstm-Crf\n",
    "    pos: BiLstm-Crf\n",
    "    srl:BiLstm-Crf\n",
    "    sa: TextCnn\n",
    "    re: TextCnn,当前这里只是有监督关系抽取\n",
    "    lm: Lstm,基础的LSTM，没有使用Seq2Seq模型\n",
    "    ss: 共享LSTM + 曼哈顿距离\n",
    "    te:共享LSTM + 全连接\n",
    "    tdp: lstm + mlp + shift-reduce(移入规约)\n",
    "    gdp: lstm + mlp + biaffine（双仿射）\n",
    "    cbow: base、hierarchical_softmax、negative_sampling\n",
    "    skip_gram: base、hierarchical_softmax、negative_sampling\n",
    "    cb: Seq2Seq+Attention\n",
    "    mt: Seq2Seq+Attention\n",
    "    ts: Seq2Seq+Attention"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c8f028c",
   "metadata": {
    "hidden": true
   },
   "source": [
    "## 使用"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "212c40a5",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 关系抽取"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2940e5e8",
   "metadata": {
    "hidden": true
   },
   "source": [
    "训练数据示例如下，其中各列分别为实体1、实体2、关系、句子\n",
    "\n",
    "    钱钟书\t辛笛\t同门\t与辛笛京沪唱和聽钱钟书与钱钟书是清华校友，钱钟书高辛笛两班。\n",
    "    元武\t元华\tunknown\t于师傅在一次京剧表演中，选了元龙（洪金宝）、元楼（元奎）、元彪、成龙、元华、元武、元泰7人担任七小福的主角。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "organizational-serum",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "'''训练'''\n",
    "from lightnlp.tc import RE\n",
    "\n",
    "re = RE()\n",
    "\n",
    "train_path = '/home/lightsmile/Projects/NLP/ChineseNRE/data/people-relation/train.sample.txt'\n",
    "dev_path = '/home/lightsmile/Projects/NLP/ChineseNRE/data/people-relation/test.sample.txt'\n",
    "vec_path = '/home/lightsmile/NLP/embedding/word/sgns.zhihu.bigram-char'\n",
    "\n",
    "re.train(train_path, dev_path=dev_path, vectors_path=vec_path, save_path='./re_saves')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "63952c2b",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "'''测试'''\n",
    "re.load('./re_saves')\n",
    "re.test(dev_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "77ef4bcc",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "'''预测'''\n",
    "print(re.predict('钱钟书', '辛笛', '与辛笛京沪唱和聽钱钟书与钱钟书是清华校友，钱钟书高辛笛两班。'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f10db787",
   "metadata": {
    "hidden": true
   },
   "source": [
    "预测结果：\n",
    "\n",
    "(0.7306928038597107, '同门') # return格式为（预测概率，预测标签）"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f70f602",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# fastNLP"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12e77909",
   "metadata": {
    "hidden": true
   },
   "source": [
    "## 文本分类例程"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b274fc84",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 读取数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "40a79f6f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:19:26.230230Z",
     "start_time": "2021-05-05T05:02:34.760175Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  6%|▌         | 98.3k/1.76M [00:00<00:01, 870kB/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "http://212.129.155.247/dataset/chn_senti_corp.zip not found in cache, downloading to C:\\Users\\小钒\\AppData\\Local\\Temp\\tmpr02odrxb\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1.76M/1.76M [00:00<00:00, 2.34MB/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finish download from http://212.129.155.247/dataset/chn_senti_corp.zip\n",
      "Copy file to C:\\Users\\小钒\\.fastNLP\\dataset\\chn_senti_corp\n"
     ]
    }
   ],
   "source": [
    "'''fastNLP提供多种数据的自动下载与自动加载功能，对于这里我们要用到的数据，\n",
    "我们可以用 Loader 自动下载并加载该数据。 更多有关Loader的使用可以参考 loader'''\n",
    "from fastNLP.io import ChnSentiCorpLoader\n",
    "\n",
    "loader = ChnSentiCorpLoader()        # 初始化一个中文情感分类的loader\n",
    "data_dir = loader.download()         # 这一行代码将自动下载数据到默认的缓存地址, 并将该地址返回\n",
    "data_bundle = loader.load(data_dir)  # 这一行代码将从{data_dir}处读取数据至DataBundle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6638cb9c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:28:29.519281Z",
     "start_time": "2021-05-05T05:28:29.503280Z"
    },
    "hidden": true,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In total 3 datasets:\n",
      "\tdev has 1200 instances.\n",
      "\ttest has 1200 instances.\n",
      "\ttrain has 9600 instances.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''DataBundle的相关介绍，可以参考 DataBundle 。我们可以打印该data_bundle的基本信息。'''\n",
    "print(data_bundle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "61ce0cb5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:33:00.850623Z",
     "start_time": "2021-05-05T05:33:00.842627Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-------------------------------------------+--------+\n",
      "| raw_chars                                 | target |\n",
      "+-------------------------------------------+--------+\n",
      "| 选择珠江花园的原因就是方便，有电动扶梯... | 1      |\n",
      "| 15.4寸笔记本的键盘确实爽，基本跟台式机... | 1      |\n",
      "+-------------------------------------------+--------+\n"
     ]
    }
   ],
   "source": [
    "'''可以看出，该data_bundle中一个含有三个 DataSet 。\n",
    "通过下面的代码，我们可以查看DataSet的基本情况'''\n",
    "print(data_bundle.get_dataset('train')[:2])  # 查看Train集前两个sample"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "748814e1",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 预处理数据"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d819bbd7",
   "metadata": {
    "hidden": true
   },
   "source": [
    "在NLP任务中，预处理一般包括:\n",
    "\n",
    "    将一整句话切分成汉字或者词;\n",
    "    将文本转换为index\n",
    "fastNLP中也提供了多种数据集的处理类，这里我们直接使用fastNLP的ChnSentiCorpPipe。更多关于Pipe的说明可以参考 pipe 。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "52ffb556",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:39:32.267599Z",
     "start_time": "2021-05-05T05:39:28.341284Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In total 3 datasets:\n",
      "\tdev has 1200 instances.\n",
      "\ttest has 1200 instances.\n",
      "\ttrain has 9600 instances.\n",
      "In total 2 vocabs:\n",
      "\tchars has 4409 entries.\n",
      "\ttarget has 2 entries.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from fastNLP.io import ChnSentiCorpPipe\n",
    "\n",
    "pipe = ChnSentiCorpPipe()\n",
    "data_bundle = pipe.process(data_bundle)  # 所有的Pipe都实现了process()方法，且输入输出都为DataBundle类型\n",
    "\n",
    "print(data_bundle)  # 打印data_bundle，查看其变化"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13b37bfe",
   "metadata": {
    "hidden": true
   },
   "source": [
    "可以看到除了之前已经包含的3个 DataSet ,还新增了两个 Vocabulary 。我们可以打印DataSet中的内容"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "27442924",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:40:47.436527Z",
     "start_time": "2021-05-05T05:40:47.414525Z"
    },
    "hidden": true,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-------------------------+--------+------------------------+---------+\n",
      "| raw_chars               | target | chars                  | seq_len |\n",
      "+-------------------------+--------+------------------------+---------+\n",
      "| 选择珠江花园的原因就... | 0      | [338, 464, 1400, 78... | 106     |\n",
      "| 15.4寸笔记本的键盘确... | 0      | [50, 133, 20, 135, ... | 56      |\n",
      "+-------------------------+--------+------------------------+---------+\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "fastNLP.io.data_bundle.DataBundle"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(data_bundle.get_dataset('train')[:2])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "35adfa81",
   "metadata": {
    "hidden": true
   },
   "source": [
    "新增了一列为数字列表的chars，以及变为数字的target列。可以看出这两列的名称和刚好与data_bundle中两个Vocabulary的名称是一致的，我们可以打印一下Vocabulary看一下里面的内容"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f4ba6c41",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:43:59.163457Z",
     "start_time": "2021-05-05T05:43:59.153456Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vocabulary(['选', '择', '珠', '江', '花']...)\n"
     ]
    }
   ],
   "source": [
    "char_vocab = data_bundle.get_vocab('chars')\n",
    "print(char_vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a60e6f77",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T05:44:54.709480Z",
     "start_time": "2021-05-05T05:44:54.700480Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "'选'的index是338\n",
      "index:338对应的汉字是选\n"
     ]
    }
   ],
   "source": [
    "'''Vocabulary是一个记录着词语与index之间映射关系的类，比如'''\n",
    "index = char_vocab.to_index('选')\n",
    "print(\"'选'的index是{}\".format(index))  # 这个值与上面打印出来的第一个instance的chars的第一个index是一致的\n",
    "print(\"index:{}对应的汉字是{}\".format(index, char_vocab.to_word(index)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b0040fc",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 选择预训练词向量"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f353777",
   "metadata": {
    "hidden": true
   },
   "source": [
    "由于Word2vec, Glove, Elmo, Bert等预训练模型可以增强模型的性能，所以在训练具体任务前，选择合适的预训练词向量非常重要。 \n",
    "\n",
    "在fastNLP中我们提供了多种Embedding使得加载这些预训练模型的过程变得更加便捷。 \n",
    "\n",
    "这里我们先给出一个使用word2vec的中文汉字预训练的示例，之后再给出一个使用Bert的文本分类。 这里使用的预训练词向量为'cn-fastnlp-100d'\n",
    "\n",
    "fastNLP将自动下载该embedding至本地缓存， fastNLP支持使用名字指定的Embedding以及相关说明可以参见 fastNLP.embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7b98cec4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T06:08:09.901593Z",
     "start_time": "2021-05-05T06:08:07.766451Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  2%|▏         | 81.9k/3.70M [00:00<00:04, 763kB/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "http://212.129.155.247/embedding/cn_char_fastnlp_100d.zip not found in cache, downloading to C:\\Users\\小钒\\AppData\\Local\\Temp\\tmphek4mbvh\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3.70M/3.70M [00:01<00:00, 2.78MB/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finish download from http://212.129.155.247/embedding/cn_char_fastnlp_100d.zip\n",
      "Copy file to C:\\Users\\小钒\\.fastNLP\\embedding\\cn_char_fastnlp_100d\n",
      "Found 4321 out of 4409 words in the pre-training embedding.\n"
     ]
    }
   ],
   "source": [
    "from fastNLP.embeddings import StaticEmbedding\n",
    "word2vec_embed = StaticEmbedding(char_vocab, model_dir_or_name='cn-char-fastnlp-100d')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "31ec933d",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 创建模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "431f3625",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T06:17:19.555495Z",
     "start_time": "2021-05-05T06:17:19.523883Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "from torch import nn\n",
    "from fastNLP.modules import LSTM\n",
    "import torch\n",
    "\n",
    "# 定义模型\n",
    "class BiLSTMMaxPoolCls(nn.Module):\n",
    "    def __init__(self, embed, num_classes, hidden_size=400, num_layers=1, dropout=0.3):\n",
    "        super().__init__()\n",
    "        self.embed = embed\n",
    "\n",
    "        self.lstm = LSTM(self.embed.embedding_dim, hidden_size=hidden_size//2, num_layers=num_layers,\n",
    "                         batch_first=True, bidirectional=True)\n",
    "        self.dropout_layer = nn.Dropout(dropout)\n",
    "        self.fc = nn.Linear(hidden_size, num_classes)\n",
    "\n",
    "    def forward(self, chars, seq_len):  # 这里的名称必须和DataSet中相应的field对应，比如之前我们DataSet中有chars，这里就必须为chars\n",
    "        # chars:[batch_size, max_len]\n",
    "        # seq_len: [batch_size, ]\n",
    "        chars = self.embed(chars)\n",
    "        outputs, _ = self.lstm(chars, seq_len)\n",
    "        outputs = self.dropout_layer(outputs)\n",
    "        outputs, _ = torch.max(outputs, dim=1)\n",
    "        outputs = self.fc(outputs)\n",
    "\n",
    "        return {'pred':outputs}  # [batch_size,], 返回值必须是dict类型，且预测值的key建议设为pred\n",
    "\n",
    "# 初始化模型\n",
    "model = BiLSTMMaxPoolCls(word2vec_embed, len(data_bundle.get_vocab('target')))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "885f3333",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### 训练模型"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d3c41cbd",
   "metadata": {
    "hidden": true
   },
   "source": [
    "fastNLP提供了Trainer对象来组织训练过程，包括完成loss计算(所以在初始化Trainer的时候需要指定loss类型)，梯度更新(所以在初始化Trainer的时候需要提供优化器optimizer)以及在验证集上的性能验证(所以在初始化时需要提供一个Metric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "0da79395",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-05T06:22:15.180842Z",
     "start_time": "2021-05-05T06:19:09.866317Z"
    },
    "hidden": true,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input fields after batch(if batch size is 2):\n",
      "\ttarget: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "\tchars: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2, 106]) \n",
      "\tseq_len: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "target fields after batch(if batch size is 2):\n",
      "\ttarget: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "\n",
      "training epochs started 2021-05-05-14-19-10-237355\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "31997f639f4f4d438cdf0907e7e123b5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, layout=Layout(flex='2'), max=3000.0), HTML(value='')), layout=Layout(d…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "Performance on test is:\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, layout=Layout(flex='2'), max=19.0), HTML(value='')), layout=Layout(dis…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "Evaluate data in 10.34 seconds!\n",
      "[tester] \n",
      "AccuracyMetric: acc=0.603333\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'AccuracyMetric': {'acc': 0.603333}}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from fastNLP import Trainer\n",
    "from fastNLP import CrossEntropyLoss\n",
    "from torch.optim import Adam\n",
    "from fastNLP import AccuracyMetric\n",
    "\n",
    "loss = CrossEntropyLoss()\n",
    "optimizer = Adam(model.parameters(), lr=0.001)\n",
    "metric = AccuracyMetric()\n",
    "device = 0 if torch.cuda.is_available() else 'cpu'  # 如果有gpu的话在gpu上运行，训练速度会更快\n",
    "\n",
    "trainer = Trainer(train_data=data_bundle.get_dataset('train'), model=model, loss=loss,\n",
    "                  optimizer=optimizer, batch_size=32, dev_data=data_bundle.get_dataset('dev'),\n",
    "                  metrics=metric, device=device)\n",
    "trainer.train()  # 开始训练，训练完成之后默认会加载在dev上表现最好的模型\n",
    "\n",
    "# 在测试集上测试一下模型的性能\n",
    "from fastNLP import Tester\n",
    "print(\"Performance on test is:\")\n",
    "tester = Tester(data=data_bundle.get_dataset('test'), model=model, metrics=metric, batch_size=64, device=device)\n",
    "tester.test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "23de1716",
   "metadata": {
    "ExecuteTime": {
     "start_time": "2021-05-05T06:24:05.173Z"
    },
    "collapsed": true,
    "hidden": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0.00/412M [00:00<?, ?B/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "http://212.129.155.247/embedding/bert-chinese-wwm.zip not found in cache, downloading to C:\\Users\\小钒\\AppData\\Local\\Temp\\tmp6f73sd8y\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 412M/412M [04:06<00:00, 1.67MB/s]   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finish download from http://212.129.155.247/embedding/bert-chinese-wwm.zip\n",
      "Copy file to C:\\Users\\小钒\\.fastNLP\\embedding\\bert-chinese-wwm\n",
      "loading vocabulary file C:\\Users\\小钒\\.fastNLP\\embedding\\bert-chinese-wwm\\vocab.txt\n",
      "Load pre-trained BERT parameters from file C:\\Users\\小钒\\.fastNLP\\embedding\\bert-chinese-wwm\\chinese_wwm_pytorch.bin.\n",
      "input fields after batch(if batch size is 2):\n",
      "\ttarget: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "\tchars: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2, 106]) \n",
      "\tseq_len: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "target fields after batch(if batch size is 2):\n",
      "\ttarget: (1)type:torch.Tensor (2)dtype:torch.int64, (3)shape:torch.Size([2]) \n",
      "\n",
      "training epochs started 2021-05-05-14-28-21-762938\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "022db1732c2648c990dcec35d3230bc8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, layout=Layout(flex='2'), max=1800.0), HTML(value='')), layout=Layout(d…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "Performance on test is:\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "36ca496ee1dc4bb48ac583687acf4aa7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, layout=Layout(flex='2'), max=19.0), HTML(value='')), layout=Layout(dis…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ERROR:root:Internal Python error in the inspect module.\n",
      "Below is the traceback from this internal error.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3418, in run_code\n",
      "    exec(code_obj, self.user_global_ns, self.user_ns)\n",
      "  File \"<ipython-input-13-c9cbd5421797>\", line 30, in <module>\n",
      "    tester.test()\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 175, in test\n",
      "    pred_dict = self._data_forward(self._predict_func, batch_x)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 223, in _data_forward\n",
      "    y = self._predict_func_wrapper(**x)\n",
      "  File \"<ipython-input-11-9938381614ea>\", line 19, in forward\n",
      "    chars = self.embed(chars)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 137, in forward\n",
      "    outputs = self.model(words)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 485, in forward\n",
      "    output_all_encoded_layers=True)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 515, in forward\n",
      "    output_all_encoded_layers=output_all_encoded_layers)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 384, in forward\n",
      "    hidden_states = layer_module(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 362, in forward\n",
      "    attention_output = self.attention(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 320, in forward\n",
      "    self_output = self.self(input_tensor, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 280, in forward\n",
      "    attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))\n",
      "KeyboardInterrupt\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'KeyboardInterrupt' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1170, in get_records\n",
      "    return _fixed_getinnerframes(etb, number_of_lines_of_context, tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 316, in wrapped\n",
      "    return f(*args, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 350, in _fixed_getinnerframes\n",
      "    records = fix_frame_records_filenames(inspect.getinnerframes(etb, context))\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1495, in getinnerframes\n",
      "    frameinfo = (tb.tb_frame,) + getframeinfo(tb, context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1453, in getframeinfo\n",
      "    filename = getsourcefile(frame) or getfile(frame)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 693, in getsourcefile\n",
      "    if os.path.exists(filename):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\genericpath.py\", line 19, in exists\n",
      "    os.stat(path)\n",
      "KeyboardInterrupt\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ERROR:root:Internal Python error in the inspect module.\n",
      "Below is the traceback from this internal error.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3418, in run_code\n",
      "    exec(code_obj, self.user_global_ns, self.user_ns)\n",
      "  File \"<ipython-input-13-c9cbd5421797>\", line 30, in <module>\n",
      "    tester.test()\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 175, in test\n",
      "    pred_dict = self._data_forward(self._predict_func, batch_x)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 223, in _data_forward\n",
      "    y = self._predict_func_wrapper(**x)\n",
      "  File \"<ipython-input-11-9938381614ea>\", line 19, in forward\n",
      "    chars = self.embed(chars)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 137, in forward\n",
      "    outputs = self.model(words)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 485, in forward\n",
      "    output_all_encoded_layers=True)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 515, in forward\n",
      "    output_all_encoded_layers=output_all_encoded_layers)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 384, in forward\n",
      "    hidden_states = layer_module(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 362, in forward\n",
      "    attention_output = self.attention(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 320, in forward\n",
      "    self_output = self.self(input_tensor, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 280, in forward\n",
      "    attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))\n",
      "KeyboardInterrupt\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'KeyboardInterrupt' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3338, in run_ast_nodes\n",
      "    if (await self.run_code(code, result,  async_=asy)):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3435, in run_code\n",
      "    self.showtraceback(running_compiled_code=True)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2048, in showtraceback\n",
      "    value, tb, tb_offset=tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1437, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1337, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1194, in structured_traceback\n",
      "    tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1151, in format_exception_as_a_whole\n",
      "    last_unique, recursion_repeat = find_recursion(orig_etype, evalue, records)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 451, in find_recursion\n",
      "    return len(records), 0\n",
      "TypeError: object of type 'NoneType' has no len()\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'TypeError' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1170, in get_records\n",
      "    return _fixed_getinnerframes(etb, number_of_lines_of_context, tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 316, in wrapped\n",
      "    return f(*args, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 350, in _fixed_getinnerframes\n",
      "    records = fix_frame_records_filenames(inspect.getinnerframes(etb, context))\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1495, in getinnerframes\n",
      "    frameinfo = (tb.tb_frame,) + getframeinfo(tb, context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1453, in getframeinfo\n",
      "    filename = getsourcefile(frame) or getfile(frame)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 696, in getsourcefile\n",
      "    if getattr(getmodule(object, filename), '__loader__', None) is not None:\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 739, in getmodule\n",
      "    f = getabsfile(module)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 708, in getabsfile\n",
      "    _filename = getsourcefile(object) or getfile(object)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 693, in getsourcefile\n",
      "    if os.path.exists(filename):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\genericpath.py\", line 19, in exists\n",
      "    os.stat(path)\n",
      "KeyboardInterrupt\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ERROR:root:Internal Python error in the inspect module.\n",
      "Below is the traceback from this internal error.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3418, in run_code\n",
      "    exec(code_obj, self.user_global_ns, self.user_ns)\n",
      "  File \"<ipython-input-13-c9cbd5421797>\", line 30, in <module>\n",
      "    tester.test()\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 175, in test\n",
      "    pred_dict = self._data_forward(self._predict_func, batch_x)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\core\\tester.py\", line 223, in _data_forward\n",
      "    y = self._predict_func_wrapper(**x)\n",
      "  File \"<ipython-input-11-9938381614ea>\", line 19, in forward\n",
      "    chars = self.embed(chars)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 137, in forward\n",
      "    outputs = self.model(words)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\embeddings\\bert_embedding.py\", line 485, in forward\n",
      "    output_all_encoded_layers=True)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 515, in forward\n",
      "    output_all_encoded_layers=output_all_encoded_layers)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 384, in forward\n",
      "    hidden_states = layer_module(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 362, in forward\n",
      "    attention_output = self.attention(hidden_states, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 320, in forward\n",
      "    self_output = self.self(input_tensor, attention_mask)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\torch\\nn\\modules\\module.py\", line 727, in _call_impl\n",
      "    result = self.forward(*input, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\fastNLP\\modules\\encoder\\bert.py\", line 280, in forward\n",
      "    attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))\n",
      "KeyboardInterrupt\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'KeyboardInterrupt' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3338, in run_ast_nodes\n",
      "    if (await self.run_code(code, result,  async_=asy)):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3435, in run_code\n",
      "    self.showtraceback(running_compiled_code=True)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2048, in showtraceback\n",
      "    value, tb, tb_offset=tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1437, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1337, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1194, in structured_traceback\n",
      "    tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1151, in format_exception_as_a_whole\n",
      "    last_unique, recursion_repeat = find_recursion(orig_etype, evalue, records)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 451, in find_recursion\n",
      "    return len(records), 0\n",
      "TypeError: object of type 'NoneType' has no len()\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'TypeError' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2923, in _run_cell\n",
      "    return runner(coro)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\async_helpers.py\", line 68, in _pseudo_sync_runner\n",
      "    coro.send(None)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3147, in run_cell_async\n",
      "    interactivity=interactivity, compiler=compiler, result=result)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 3357, in run_ast_nodes\n",
      "    self.showtraceback()\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2048, in showtraceback\n",
      "    value, tb, tb_offset=tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1437, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1337, in structured_traceback\n",
      "    self, etype, value, tb, tb_offset, number_of_lines_of_context\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1212, in structured_traceback\n",
      "    chained_exceptions_tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1151, in format_exception_as_a_whole\n",
      "    last_unique, recursion_repeat = find_recursion(orig_etype, evalue, records)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 451, in find_recursion\n",
      "    return len(records), 0\n",
      "TypeError: object of type 'NoneType' has no len()\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\interactiveshell.py\", line 2045, in showtraceback\n",
      "    stb = value._render_traceback_()\n",
      "AttributeError: 'TypeError' object has no attribute '_render_traceback_'\n",
      "\n",
      "During handling of the above exception, another exception occurred:\n",
      "\n",
      "Traceback (most recent call last):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 1170, in get_records\n",
      "    return _fixed_getinnerframes(etb, number_of_lines_of_context, tb_offset)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 316, in wrapped\n",
      "    return f(*args, **kwargs)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\site-packages\\IPython\\core\\ultratb.py\", line 350, in _fixed_getinnerframes\n",
      "    records = fix_frame_records_filenames(inspect.getinnerframes(etb, context))\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1495, in getinnerframes\n",
      "    frameinfo = (tb.tb_frame,) + getframeinfo(tb, context)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 1453, in getframeinfo\n",
      "    filename = getsourcefile(frame) or getfile(frame)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 696, in getsourcefile\n",
      "    if getattr(getmodule(object, filename), '__loader__', None) is not None:\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 739, in getmodule\n",
      "    f = getabsfile(module)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 708, in getabsfile\n",
      "    _filename = getsourcefile(object) or getfile(object)\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\inspect.py\", line 693, in getsourcefile\n",
      "    if os.path.exists(filename):\n",
      "  File \"D:\\Anaconda\\envs\\Deep\\lib\\genericpath.py\", line 19, in exists\n",
      "    os.stat(path)\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "'''使用BERT进行文本分类'''\n",
    "# 只需要切换一下Embedding即可\n",
    "from fastNLP.embeddings import BertEmbedding\n",
    "\n",
    "# 这里为了演示一下效果，所以默认Bert不更新权重\n",
    "bert_embed = BertEmbedding(char_vocab, model_dir_or_name='cn', auto_truncate=True, requires_grad=False)\n",
    "model = BiLSTMMaxPoolCls(bert_embed, len(data_bundle.get_vocab('target')))\n",
    "\n",
    "\n",
    "import torch\n",
    "from fastNLP import Trainer\n",
    "from fastNLP import CrossEntropyLoss\n",
    "from torch.optim import Adam\n",
    "from fastNLP import AccuracyMetric\n",
    "\n",
    "loss = CrossEntropyLoss()\n",
    "optimizer = Adam(model.parameters(), lr=2e-5)\n",
    "metric = AccuracyMetric()\n",
    "device = 0 if torch.cuda.is_available() else 'cpu'  # 如果有gpu的话在gpu上运行，训练速度会更快\n",
    "\n",
    "trainer = Trainer(train_data=data_bundle.get_dataset('train'), model=model, loss=loss,\n",
    "                  optimizer=optimizer, batch_size=16, dev_data=data_bundle.get_dataset('test'),\n",
    "                  metrics=metric, device=device, n_epochs=3)\n",
    "trainer.train()  # 开始训练，训练完成之后默认会加载在dev上表现最好的模型\n",
    "\n",
    "# 在测试集上测试一下模型的性能\n",
    "from fastNLP import Tester\n",
    "print(\"Performance on test is:\")\n",
    "tester = Tester(data=data_bundle.get_dataset('test'), model=model, metrics=metric, batch_size=64, device=device)\n",
    "tester.test()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a8810348",
   "metadata": {},
   "source": [
    "# FlashText - 比正则快M倍以上！Python替换字符串的新姿势"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "8f096937",
   "metadata": {},
   "source": [
    "FlashText 算法是由 Vikash Singh 于2017年发表的大规模关键词替换算法，这个算法的时间复杂度仅由文本长度（N）决定，算法时间复杂度为O（N）。\n",
    "\n",
    "而对于正则表达式的替换，算法时间复杂度还需要考虑被替换的关键词数量（M），因此时间复杂度为O（MxN）。\n",
    "\n",
    "简而言之，基于FlashText算法的字符串替换比正则表达式替换快M倍以上，这个M是需要替换的关键词数量，关键词越多，FlashText算法的优势就越明显\n",
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c5ce5da1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:25:53.486102Z",
     "start_time": "2021-05-25T08:25:53.466970Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['New York', 'Bay Area']\n"
     ]
    }
   ],
   "source": [
    "# 提取关键词\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('Big Apple', 'New York')\n",
    "keyword_processor.add_keyword('Bay Area')\n",
    "# 3. 处理目标句子并提取相应关键词\n",
    "keywords_found = keyword_processor.extract_keywords('I love Big Apple and Bay Area.')\n",
    "# 4. 结果\n",
    "print(keywords_found)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec7d672e",
   "metadata": {},
   "source": [
    "其中 add_keyword 的第一个参数代表需要被查找的关键词，第二个参数是给这个关键词一个别名，如果找到了则以别名显示。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ec5a3ff",
   "metadata": {},
   "source": [
    "⭐常用:替换关键词"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "dc88266b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T10:33:49.766109Z",
     "start_time": "2021-05-25T10:33:49.759110Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I love Big Apple and NCR region.\n",
      "<class 'str'>\n"
     ]
    }
   ],
   "source": [
    "# 替换关键词 - 如果你想要替换关键词，只需要调用处理器的 replace_keywords 函数：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('New Delhi', 'NCR region')\n",
    "# 3. 替换关键词\n",
    "new_sentence = keyword_processor.replace_keywords('I love Big Apple and new delhi.')\n",
    "# 4. 结果\n",
    "print(new_sentence)\n",
    "print(type(new_sentence))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ec49a049",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:30:59.754502Z",
     "start_time": "2021-05-25T08:30:59.748111Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Bay Area']\n"
     ]
    }
   ],
   "source": [
    "# 关键词大小写敏感 - 如果你需要精确提取，识别大小写字母，那么你可以在处理器初始化的时候设定 sensitive 参数:\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器, 注意设置大小写敏感(case_sensitive)为TRUE\n",
    "keyword_processor = KeywordProcessor(case_sensitive=True)\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('Big Apple', 'New York')\n",
    "keyword_processor.add_keyword('Bay Area')\n",
    "# 3. 处理目标句子并提取相应关键词\n",
    "keywords_found = keyword_processor.extract_keywords('I love big Apple and Bay Area.')\n",
    "# 4. 结果\n",
    "print(keywords_found)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "83fb52e4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:31:38.536026Z",
     "start_time": "2021-05-25T08:31:38.519956Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('New York', 7, 16), ('Bay Area', 21, 29)]\n"
     ]
    }
   ],
   "source": [
    "# 标记关键词位置 - 如果你需要获取关键词在句子中的位置，在 extract_keywords 的时候添加 span_info=True 参数即可：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('Big Apple', 'New York')\n",
    "keyword_processor.add_keyword('Bay Area')\n",
    "# 3. 处理目标句子并提取相应关键词, 并标记关键词的起始、终止位置\n",
    "keywords_found = keyword_processor.extract_keywords('I love big Apple and Bay Area.', span_info=True)\n",
    "# 4. 结果\n",
    "print(keywords_found)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "1ef710e1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:49:27.005997Z",
     "start_time": "2021-05-25T08:49:26.985343Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'j2ee': 'Java', 'colour': 'color'}"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取目前所有的关键词-如果你需要获取当前已经添加的所有关键词，只需要调用处理器的 get_all_keywords 函数：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('j2ee', 'Java')\n",
    "keyword_processor.add_keyword('colour', 'color')\n",
    "# 3. 获取所有关键词\n",
    "keyword_processor.get_all_keywords()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9abb23db",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:32:10.743237Z",
     "start_time": "2021-05-25T08:32:10.721884Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['product management', 'java']"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 批量添加关键词 - 批量添加关键词有两种方法，一种是通过词典，一种是通过数组：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. （第一种）通过字典批量添加关键词\n",
    "keyword_dict = {\n",
    "    \"java\": [\"java_2e\", \"java programing\"],\n",
    "    \"product management\": [\"PM\", \"product manager\"]\n",
    "}\n",
    "keyword_processor.add_keywords_from_dict(keyword_dict)\n",
    "# 2. （第二种）通过数组批量添加关键词\n",
    "keyword_processor.add_keywords_from_list([\"java\", \"python\"])\n",
    "# 3. 第一种的提取效果如下\n",
    "keyword_processor.extract_keywords('I am a product manager for a java_2e platform')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a3fe639f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:32:34.946834Z",
     "start_time": "2021-05-25T08:32:34.921997Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['product management', 'java']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['product management']"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 单一或批量删除关键词 - 除关键词也非常简单，和添加类似：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 通过字典批量添加关键词\n",
    "keyword_dict = {\n",
    "    \"java\": [\"java_2e\", \"java programing\"],\n",
    "    \"product management\": [\"PM\", \"product manager\"]\n",
    "}\n",
    "keyword_processor.add_keywords_from_dict(keyword_dict)\n",
    "# 3. 提取效果如下\n",
    "print(keyword_processor.extract_keywords('I am a product manager for a java_2e platform'))\n",
    "# ['product management', 'java']\n",
    "# 4. 单个删除关键词\n",
    "keyword_processor.remove_keyword('java_2e')\n",
    "# 5. 批量删除关键词，也是可以通过词典或者数组的形式\n",
    "keyword_processor.remove_keywords_from_dict({\"product management\": [\"PM\"]})\n",
    "keyword_processor.remove_keywords_from_list([\"java programing\"])\n",
    "# 6. 删除了java programing关键词后的效果如下\n",
    "keyword_processor.extract_keywords('I am a product manager for a java_2e platform')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9029d5ce",
   "metadata": {},
   "source": [
    "⭐额外信息-可用于根据keyword抽取关系"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8c40fe51",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:47:05.463863Z",
     "start_time": "2021-05-25T08:47:05.449872Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Monument', 'Taj Mahal'), ('Location', 'Delhi')]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 支持额外信息-前面提到在添加关键词的时候第二个参数为其别名，其实你不仅可以指示别名，还可以将额外信息放到第二个参数中：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "kp = KeywordProcessor()\n",
    "# 2. 添加关键词并附带额外信息\n",
    "kp.add_keyword('Taj Mahal', ('Monument', 'Taj Mahal'))\n",
    "kp.add_keyword('Delhi', ('Location', 'Delhi'))\n",
    "# 3. 效果如下\n",
    "kp.extract_keywords('Taj Mahal is in Delhi.')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "752e8652",
   "metadata": {},
   "source": [
    "这样，在提取关键词的时候，你还能拿到其他一些你想要在得到此关键词时输出的信息"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "19965ba3",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T08:50:28.069014Z",
     "start_time": "2021-05-25T08:50:28.060011Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Big Apple']\n",
      "[]\n"
     ]
    }
   ],
   "source": [
    "# 支持特殊单词边界 - Flashtext 检测的单词边界一般局限于 \\w [A-Za-z0-9_] 外的任意字符，但是如果你想添加某些特殊字符作为单词的一部分也是可以实现的：\n",
    "from flashtext import KeywordProcessor\n",
    "# 1. 初始化关键字处理器\n",
    "keyword_processor = KeywordProcessor()\n",
    "# 2. 添加关键词\n",
    "keyword_processor.add_keyword('Big Apple')\n",
    "# 3. 正常效果\n",
    "print(keyword_processor.extract_keywords('I love Big Apple/Bay Area.'))\n",
    "# ['Big Apple']\n",
    "# 4. 将 '/' 作为单词一部分\n",
    "keyword_processor.add_non_word_boundary('/')\n",
    "# 5. 优化后的效果\n",
    "print(keyword_processor.extract_keywords('I love Big Apple/Bay Area.'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4914f6e",
   "metadata": {},
   "source": [
    "# FoolNLTK - 基于BiLSTM模型训练而成，功能包含分词，词性标注，实体识别。并支持用户自定义词典，可训练自己的模型及批量处理文本"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c5bb1bc",
   "metadata": {},
   "source": [
    "作者号称“可能不是最快的开源中文分词，但很可能是最准的开源中文分词”"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c29850e8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-25T11:03:27.112130Z",
     "start_time": "2021-05-25T11:03:27.099123Z"
    },
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "import foolnltk as fool\n",
    "text = \"一个肖仰华在上海\"\n",
    "print(fool.cut(text))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "85349bba",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:Deep]",
   "language": "python",
   "name": "deep"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}