Text_Rank.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Graph Construction\n",
    "\n",
    "Let's look at how to build a graph from a raw text.\n",
    "\n",
    "This work is based on \n",
    "[\"TextRank: Bringing Order into Text\"](http://web.eecs.umich.edu/~mihalcea/papers/mihalcea.emnlp04.pdf),\n",
    "Rada Mihalcea, Paul Tarau, _Empirical Methods in Natural Language Processing_ (2004)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "First we perform some basic housekeeping for Jupyter and then load `spaCy` ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import logging\n",
    "import sys\n",
    "import warnings\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)\n",
    "logger = logging.getLogger(\"PyTR\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:transformers.file_utils:PyTorch version 1.3.0 available.\n",
      "INFO:transformers.modeling_xlnet:Better speed can be achieved with apex installed from https://www.github.com/nvidia/apex .\n"
     ]
    }
   ],
   "source": [
    "import spacy\n",
    "nlp = spacy.load(\"en_core_web_sm\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, to get started, we'll create some text to use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#text = \"When Ada was twelve years old, this future 'Lady Fairy', as Charles Babbage affectionately called her, decided she wanted to fly. Ada Byron went about the project methodically, thoughtfully, with imagination and passion. Her first step, in February 1828, was to construct wings. She investigated different material and sizes. She considered various materials for the wings: paper, oilsilk, wires, and feathers. She examined the anatomy of birds to determine the right proportion between the wings and the body. She decided to write a book, Flyology, illustrating, with plates, some of her findings. She decided what equipment she would need; for example, a compass, to 'cut across the country by the most direct road', so that she could surmount mountains, rivers, and valleys. Her final step was to integrate steam with the 'art of flying.\"\n",
    "\n",
    "text = \"Compatibility of systems of linear constraints over the set of natural numbers. Criteria of compatibility of a system of linear Diophantine equations, strict inequations, and nonstrict inequations are considered. Upper bounds for components of a minimal set of solutions and algorithms of construction of minimal generating sets of solutions for all types of systems are given. These criteria and the corresponding algorithms for constructing a minimal supporting set of solutions can be used in solving all the considered types systems and systems of mixed types.\"\n",
    "\n",
    "doc = nlp(text)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How many sentences are in the parsed document and where are their boundaries?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "> 0 13\n",
      "> 13 33\n",
      "> 33 61\n",
      "> 61 91\n"
     ]
    }
   ],
   "source": [
    "for sent in doc.sents:\n",
    "    print(\">\", sent.start, sent.end)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What are the raw _noun chunks_ in the parsed document, as well as its _named entities_?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Compatibility\n",
      "systems\n",
      "linear constraints\n",
      "the set\n",
      "natural numbers\n",
      "Criteria\n",
      "compatibility\n",
      "a system\n",
      "linear Diophantine equations\n",
      "strict inequations\n",
      "nonstrict inequations\n",
      "Upper bounds\n",
      "components\n",
      "a minimal set\n",
      "solutions\n",
      "algorithms\n",
      "construction\n",
      "minimal generating sets\n",
      "solutions\n",
      "all types\n",
      "systems\n",
      "These criteria\n",
      "the corresponding algorithms\n",
      "a minimal supporting set\n",
      "solutions\n",
      "all the considered types systems\n",
      "systems\n",
      "mixed types\n"
     ]
    }
   ],
   "source": [
    "for chunk in doc.noun_chunks:\n",
    "    print(chunk.text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Diophantine ORG 21 22\n"
     ]
    }
   ],
   "source": [
    "for ent in doc.ents:\n",
    "    print(ent.text, ent.label_, ent.start, ent.end)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Given those details about the parsed document, next we use [NetworkX](https://networkx.github.io/) to manage an in-memory graph..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import networkx as nx\n",
    "\n",
    "def increment_edge (graph, node0, node1):\n",
    "    if logger:\n",
    "        logger.debug(\"link {} {}\".format(node0, node1))\n",
    "    \n",
    "    if graph.has_edge(node0, node1):\n",
    "        graph[node0][node1][\"weight\"] += 1.0\n",
    "    else:\n",
    "        graph.add_edge(node0, node1, weight=1.0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then construct a graph, sentence by sentence, based on the [spaCy part-of-speech tags](https://spacy.io/api/annotation#pos-en) tags:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "POS_KEPT = [\"ADJ\", \"NOUN\", \"PROPN\", \"VERB\"]\n",
    "\n",
    "def link_sentence (doc, sent, lemma_graph, seen_lemma):\n",
    "    visited_tokens = []\n",
    "    visited_nodes = []\n",
    "\n",
    "    for i in range(sent.start, sent.end):\n",
    "        token = doc[i]\n",
    "\n",
    "        if token.pos_ in POS_KEPT:\n",
    "            key = (token.lemma_, token.pos_)\n",
    "\n",
    "            if key not in seen_lemma:\n",
    "                seen_lemma[key] = set([token.i])\n",
    "            else:\n",
    "                seen_lemma[key].add(token.i)\n",
    "\n",
    "            node_id = list(seen_lemma.keys()).index(key)\n",
    "\n",
    "            if not node_id in lemma_graph:\n",
    "                lemma_graph.add_node(node_id)\n",
    "\n",
    "            if logger:\n",
    "                logger.debug(\"visit {} {}\".format(visited_tokens, visited_nodes))\n",
    "                logger.debug(\"range {}\".format(list(range(len(visited_tokens) - 1, -1, -1))))\n",
    "            \n",
    "            for prev_token in range(len(visited_tokens) - 1, -1, -1):\n",
    "                if logger:\n",
    "                    logger.debug(\"prev_tok {} {}\".format(prev_token, (token.i - visited_tokens[prev_token])))\n",
    "                \n",
    "                if (token.i - visited_tokens[prev_token]) <= 3:\n",
    "                    increment_edge(lemma_graph, node_id, visited_nodes[prev_token])\n",
    "                else:\n",
    "                    break\n",
    "\n",
    "            if logger:\n",
    "                logger.debug(\" -- {} {} {} {} {} {}\".format(token.i, token.text, token.lemma_, token.pos_, visited_tokens, visited_nodes))\n",
    "\n",
    "            visited_tokens.append(token.i)\n",
    "            visited_nodes.append(node_id)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now iterate through the sentences to construct the lemma graph..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:PyTR:visit [] []\n",
      "DEBUG:PyTR:range []\n",
      "DEBUG:PyTR: -- 0 Compatibility compatibility NOUN [] []\n",
      "DEBUG:PyTR:visit [0] [0]\n",
      "DEBUG:PyTR:range [0]\n",
      "DEBUG:PyTR:prev_tok 0 2\n",
      "DEBUG:PyTR:link 1 0\n",
      "DEBUG:PyTR: -- 2 systems system NOUN [0] [0]\n",
      "DEBUG:PyTR:visit [0, 2] [0, 1]\n",
      "DEBUG:PyTR:range [1, 0]\n",
      "DEBUG:PyTR:prev_tok 1 2\n",
      "DEBUG:PyTR:link 2 1\n",
      "DEBUG:PyTR:prev_tok 0 4\n",
      "DEBUG:PyTR: -- 4 linear linear ADJ [0, 2] [0, 1]\n",
      "DEBUG:PyTR:visit [0, 2, 4] [0, 1, 2]\n",
      "DEBUG:PyTR:range [2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 2 1\n",
      "DEBUG:PyTR:link 3 2\n",
      "DEBUG:PyTR:prev_tok 1 3\n",
      "DEBUG:PyTR:link 3 1\n",
      "DEBUG:PyTR:prev_tok 0 5\n",
      "DEBUG:PyTR: -- 5 constraints constraint NOUN [0, 2, 4] [0, 1, 2]\n",
      "DEBUG:PyTR:visit [0, 2, 4, 5] [0, 1, 2, 3]\n",
      "DEBUG:PyTR:range [3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 3 3\n",
      "DEBUG:PyTR:link 4 3\n",
      "DEBUG:PyTR:prev_tok 2 4\n",
      "DEBUG:PyTR: -- 8 set set NOUN [0, 2, 4, 5] [0, 1, 2, 3]\n",
      "DEBUG:PyTR:visit [0, 2, 4, 5, 8] [0, 1, 2, 3, 4]\n",
      "DEBUG:PyTR:range [4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 4 2\n",
      "DEBUG:PyTR:link 5 4\n",
      "DEBUG:PyTR:prev_tok 3 5\n",
      "DEBUG:PyTR: -- 10 natural natural ADJ [0, 2, 4, 5, 8] [0, 1, 2, 3, 4]\n",
      "DEBUG:PyTR:visit [0, 2, 4, 5, 8, 10] [0, 1, 2, 3, 4, 5]\n",
      "DEBUG:PyTR:range [5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 5 1\n",
      "DEBUG:PyTR:link 6 5\n",
      "DEBUG:PyTR:prev_tok 4 3\n",
      "DEBUG:PyTR:link 6 4\n",
      "DEBUG:PyTR:prev_tok 3 6\n",
      "DEBUG:PyTR: -- 11 numbers number NOUN [0, 2, 4, 5, 8, 10] [0, 1, 2, 3, 4, 5]\n",
      "DEBUG:PyTR:visit [] []\n",
      "DEBUG:PyTR:range []\n",
      "DEBUG:PyTR: -- 13 Criteria criterion NOUN [] []\n",
      "DEBUG:PyTR:visit [13] [7]\n",
      "DEBUG:PyTR:range [0]\n",
      "DEBUG:PyTR:prev_tok 0 2\n",
      "DEBUG:PyTR:link 0 7\n",
      "DEBUG:PyTR: -- 15 compatibility compatibility NOUN [13] [7]\n",
      "DEBUG:PyTR:visit [13, 15] [7, 0]\n",
      "DEBUG:PyTR:range [1, 0]\n",
      "DEBUG:PyTR:prev_tok 1 3\n",
      "DEBUG:PyTR:link 1 0\n",
      "DEBUG:PyTR:prev_tok 0 5\n",
      "DEBUG:PyTR: -- 18 system system NOUN [13, 15] [7, 0]\n",
      "DEBUG:PyTR:visit [13, 15, 18] [7, 0, 1]\n",
      "DEBUG:PyTR:range [2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 2 2\n",
      "DEBUG:PyTR:link 2 1\n",
      "DEBUG:PyTR:prev_tok 1 5\n",
      "DEBUG:PyTR: -- 20 linear linear ADJ [13, 15, 18] [7, 0, 1]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20] [7, 0, 1, 2]\n",
      "DEBUG:PyTR:range [3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 3 1\n",
      "DEBUG:PyTR:link 8 2\n",
      "DEBUG:PyTR:prev_tok 2 3\n",
      "DEBUG:PyTR:link 8 1\n",
      "DEBUG:PyTR:prev_tok 1 6\n",
      "DEBUG:PyTR: -- 21 Diophantine Diophantine PROPN [13, 15, 18, 20] [7, 0, 1, 2]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21] [7, 0, 1, 2, 8]\n",
      "DEBUG:PyTR:range [4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 4 1\n",
      "DEBUG:PyTR:link 9 8\n",
      "DEBUG:PyTR:prev_tok 3 2\n",
      "DEBUG:PyTR:link 9 2\n",
      "DEBUG:PyTR:prev_tok 2 4\n",
      "DEBUG:PyTR: -- 22 equations equation NOUN [13, 15, 18, 20, 21] [7, 0, 1, 2, 8]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21, 22] [7, 0, 1, 2, 8, 9]\n",
      "DEBUG:PyTR:range [5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 5 2\n",
      "DEBUG:PyTR:link 10 9\n",
      "DEBUG:PyTR:prev_tok 4 3\n",
      "DEBUG:PyTR:link 10 8\n",
      "DEBUG:PyTR:prev_tok 3 4\n",
      "DEBUG:PyTR: -- 24 strict strict ADJ [13, 15, 18, 20, 21, 22] [7, 0, 1, 2, 8, 9]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21, 22, 24] [7, 0, 1, 2, 8, 9, 10]\n",
      "DEBUG:PyTR:range [6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 6 1\n",
      "DEBUG:PyTR:link 11 10\n",
      "DEBUG:PyTR:prev_tok 5 3\n",
      "DEBUG:PyTR:link 11 9\n",
      "DEBUG:PyTR:prev_tok 4 4\n",
      "DEBUG:PyTR: -- 25 inequations inequation NOUN [13, 15, 18, 20, 21, 22, 24] [7, 0, 1, 2, 8, 9, 10]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21, 22, 24, 25] [7, 0, 1, 2, 8, 9, 10, 11]\n",
      "DEBUG:PyTR:range [7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 7 3\n",
      "DEBUG:PyTR:link 12 11\n",
      "DEBUG:PyTR:prev_tok 6 4\n",
      "DEBUG:PyTR: -- 28 nonstrict nonstrict NOUN [13, 15, 18, 20, 21, 22, 24, 25] [7, 0, 1, 2, 8, 9, 10, 11]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21, 22, 24, 25, 28] [7, 0, 1, 2, 8, 9, 10, 11, 12]\n",
      "DEBUG:PyTR:range [8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 8 1\n",
      "DEBUG:PyTR:link 11 12\n",
      "DEBUG:PyTR:prev_tok 7 4\n",
      "DEBUG:PyTR: -- 29 inequations inequation NOUN [13, 15, 18, 20, 21, 22, 24, 25, 28] [7, 0, 1, 2, 8, 9, 10, 11, 12]\n",
      "DEBUG:PyTR:visit [13, 15, 18, 20, 21, 22, 24, 25, 28, 29] [7, 0, 1, 2, 8, 9, 10, 11, 12, 11]\n",
      "DEBUG:PyTR:range [9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 9 2\n",
      "DEBUG:PyTR:link 13 11\n",
      "DEBUG:PyTR:prev_tok 8 3\n",
      "DEBUG:PyTR:link 13 12\n",
      "DEBUG:PyTR:prev_tok 7 6\n",
      "DEBUG:PyTR: -- 31 considered consider VERB [13, 15, 18, 20, 21, 22, 24, 25, 28, 29] [7, 0, 1, 2, 8, 9, 10, 11, 12, 11]\n",
      "DEBUG:PyTR:visit [] []\n",
      "DEBUG:PyTR:range []\n",
      "DEBUG:PyTR: -- 33 Upper upper ADJ [] []\n",
      "DEBUG:PyTR:visit [33] [14]\n",
      "DEBUG:PyTR:range [0]\n",
      "DEBUG:PyTR:prev_tok 0 1\n",
      "DEBUG:PyTR:link 15 14\n",
      "DEBUG:PyTR: -- 34 bounds bound NOUN [33] [14]\n",
      "DEBUG:PyTR:visit [33, 34] [14, 15]\n",
      "DEBUG:PyTR:range [1, 0]\n",
      "DEBUG:PyTR:prev_tok 1 2\n",
      "DEBUG:PyTR:link 16 15\n",
      "DEBUG:PyTR:prev_tok 0 3\n",
      "DEBUG:PyTR:link 16 14\n",
      "DEBUG:PyTR: -- 36 components component NOUN [33, 34] [14, 15]\n",
      "DEBUG:PyTR:visit [33, 34, 36] [14, 15, 16]\n",
      "DEBUG:PyTR:range [2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 2 3\n",
      "DEBUG:PyTR:link 17 16\n",
      "DEBUG:PyTR:prev_tok 1 5\n",
      "DEBUG:PyTR: -- 39 minimal minimal ADJ [33, 34, 36] [14, 15, 16]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39] [14, 15, 16, 17]\n",
      "DEBUG:PyTR:range [3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 3 1\n",
      "DEBUG:PyTR:link 4 17\n",
      "DEBUG:PyTR:prev_tok 2 4\n",
      "DEBUG:PyTR: -- 40 set set NOUN [33, 34, 36, 39] [14, 15, 16, 17]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40] [14, 15, 16, 17, 4]\n",
      "DEBUG:PyTR:range [4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 4 2\n",
      "DEBUG:PyTR:link 18 4\n",
      "DEBUG:PyTR:prev_tok 3 3\n",
      "DEBUG:PyTR:link 18 17\n",
      "DEBUG:PyTR:prev_tok 2 6\n",
      "DEBUG:PyTR: -- 42 solutions solution NOUN [33, 34, 36, 39, 40] [14, 15, 16, 17, 4]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42] [14, 15, 16, 17, 4, 18]\n",
      "DEBUG:PyTR:range [5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 5 2\n",
      "DEBUG:PyTR:link 19 18\n",
      "DEBUG:PyTR:prev_tok 4 4\n",
      "DEBUG:PyTR: -- 44 algorithms algorithm NOUN [33, 34, 36, 39, 40, 42] [14, 15, 16, 17, 4, 18]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44] [14, 15, 16, 17, 4, 18, 19]\n",
      "DEBUG:PyTR:range [6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 6 2\n",
      "DEBUG:PyTR:link 20 19\n",
      "DEBUG:PyTR:prev_tok 5 4\n",
      "DEBUG:PyTR: -- 46 construction construction NOUN [33, 34, 36, 39, 40, 42, 44] [14, 15, 16, 17, 4, 18, 19]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46] [14, 15, 16, 17, 4, 18, 19, 20]\n",
      "DEBUG:PyTR:range [7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 7 2\n",
      "DEBUG:PyTR:link 17 20\n",
      "DEBUG:PyTR:prev_tok 6 4\n",
      "DEBUG:PyTR: -- 48 minimal minimal ADJ [33, 34, 36, 39, 40, 42, 44, 46] [14, 15, 16, 17, 4, 18, 19, 20]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48] [14, 15, 16, 17, 4, 18, 19, 20, 17]\n",
      "DEBUG:PyTR:range [8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 8 1\n",
      "DEBUG:PyTR:link 21 17\n",
      "DEBUG:PyTR:prev_tok 7 3\n",
      "DEBUG:PyTR:link 21 20\n",
      "DEBUG:PyTR:prev_tok 6 5\n",
      "DEBUG:PyTR: -- 49 generating generating NOUN [33, 34, 36, 39, 40, 42, 44, 46, 48] [14, 15, 16, 17, 4, 18, 19, 20, 17]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48, 49] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21]\n",
      "DEBUG:PyTR:range [9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 9 1\n",
      "DEBUG:PyTR:link 4 21\n",
      "DEBUG:PyTR:prev_tok 8 2\n",
      "DEBUG:PyTR:link 4 17\n",
      "DEBUG:PyTR:prev_tok 7 4\n",
      "DEBUG:PyTR: -- 50 sets set NOUN [33, 34, 36, 39, 40, 42, 44, 46, 48, 49] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4]\n",
      "DEBUG:PyTR:range [10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 10 2\n",
      "DEBUG:PyTR:link 18 4\n",
      "DEBUG:PyTR:prev_tok 9 3\n",
      "DEBUG:PyTR:link 18 21\n",
      "DEBUG:PyTR:prev_tok 8 4\n",
      "DEBUG:PyTR: -- 52 solutions solution NOUN [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18]\n",
      "DEBUG:PyTR:range [11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 11 3\n",
      "DEBUG:PyTR:link 22 18\n",
      "DEBUG:PyTR:prev_tok 10 5\n",
      "DEBUG:PyTR: -- 55 types type NOUN [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52, 55] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18, 22]\n",
      "DEBUG:PyTR:range [12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 12 2\n",
      "DEBUG:PyTR:link 1 22\n",
      "DEBUG:PyTR:prev_tok 11 5\n",
      "DEBUG:PyTR: -- 57 systems system NOUN [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52, 55] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18, 22]\n",
      "DEBUG:PyTR:visit [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52, 55, 57] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18, 22, 1]\n",
      "DEBUG:PyTR:range [13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:PyTR:prev_tok 13 2\n",
      "DEBUG:PyTR:link 23 1\n",
      "DEBUG:PyTR:prev_tok 12 4\n",
      "DEBUG:PyTR: -- 59 given give VERB [33, 34, 36, 39, 40, 42, 44, 46, 48, 49, 50, 52, 55, 57] [14, 15, 16, 17, 4, 18, 19, 20, 17, 21, 4, 18, 22, 1]\n",
      "DEBUG:PyTR:visit [] []\n",
      "DEBUG:PyTR:range []\n",
      "DEBUG:PyTR: -- 62 criteria criterion NOUN [] []\n",
      "DEBUG:PyTR:visit [62] [7]\n",
      "DEBUG:PyTR:range [0]\n",
      "DEBUG:PyTR:prev_tok 0 3\n",
      "DEBUG:PyTR:link 24 7\n",
      "DEBUG:PyTR: -- 65 corresponding corresponding ADJ [62] [7]\n",
      "DEBUG:PyTR:visit [62, 65] [7, 24]\n",
      "DEBUG:PyTR:range [1, 0]\n",
      "DEBUG:PyTR:prev_tok 1 1\n",
      "DEBUG:PyTR:link 19 24\n",
      "DEBUG:PyTR:prev_tok 0 4\n",
      "DEBUG:PyTR: -- 66 algorithms algorithm NOUN [62, 65] [7, 24]\n",
      "DEBUG:PyTR:visit [62, 65, 66] [7, 24, 19]\n",
      "DEBUG:PyTR:range [2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 2 2\n",
      "DEBUG:PyTR:link 25 19\n",
      "DEBUG:PyTR:prev_tok 1 3\n",
      "DEBUG:PyTR:link 25 24\n",
      "DEBUG:PyTR:prev_tok 0 6\n",
      "DEBUG:PyTR: -- 68 constructing construct VERB [62, 65, 66] [7, 24, 19]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68] [7, 24, 19, 25]\n",
      "DEBUG:PyTR:range [3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 3 2\n",
      "DEBUG:PyTR:link 17 25\n",
      "DEBUG:PyTR:prev_tok 2 4\n",
      "DEBUG:PyTR: -- 70 minimal minimal ADJ [62, 65, 66, 68] [7, 24, 19, 25]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70] [7, 24, 19, 25, 17]\n",
      "DEBUG:PyTR:range [4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 4 1\n",
      "DEBUG:PyTR:link 26 17\n",
      "DEBUG:PyTR:prev_tok 3 3\n",
      "DEBUG:PyTR:link 26 25\n",
      "DEBUG:PyTR:prev_tok 2 5\n",
      "DEBUG:PyTR: -- 71 supporting support VERB [62, 65, 66, 68, 70] [7, 24, 19, 25, 17]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71] [7, 24, 19, 25, 17, 26]\n",
      "DEBUG:PyTR:range [5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 5 1\n",
      "DEBUG:PyTR:link 4 26\n",
      "DEBUG:PyTR:prev_tok 4 2\n",
      "DEBUG:PyTR:link 4 17\n",
      "DEBUG:PyTR:prev_tok 3 4\n",
      "DEBUG:PyTR: -- 72 set set NOUN [62, 65, 66, 68, 70, 71] [7, 24, 19, 25, 17, 26]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72] [7, 24, 19, 25, 17, 26, 4]\n",
      "DEBUG:PyTR:range [6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 6 2\n",
      "DEBUG:PyTR:link 18 4\n",
      "DEBUG:PyTR:prev_tok 5 3\n",
      "DEBUG:PyTR:link 18 26\n",
      "DEBUG:PyTR:prev_tok 4 4\n",
      "DEBUG:PyTR: -- 74 solutions solution NOUN [62, 65, 66, 68, 70, 71, 72] [7, 24, 19, 25, 17, 26, 4]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74] [7, 24, 19, 25, 17, 26, 4, 18]\n",
      "DEBUG:PyTR:range [7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 7 3\n",
      "DEBUG:PyTR:link 27 18\n",
      "DEBUG:PyTR:prev_tok 6 5\n",
      "DEBUG:PyTR: -- 77 used use VERB [62, 65, 66, 68, 70, 71, 72, 74] [7, 24, 19, 25, 17, 26, 4, 18]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77] [7, 24, 19, 25, 17, 26, 4, 18, 27]\n",
      "DEBUG:PyTR:range [8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 8 2\n",
      "DEBUG:PyTR:link 28 27\n",
      "DEBUG:PyTR:prev_tok 7 5\n",
      "DEBUG:PyTR: -- 79 solving solve VERB [62, 65, 66, 68, 70, 71, 72, 74, 77] [7, 24, 19, 25, 17, 26, 4, 18, 27]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28]\n",
      "DEBUG:PyTR:range [9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 9 3\n",
      "DEBUG:PyTR:link 13 28\n",
      "DEBUG:PyTR:prev_tok 8 5\n",
      "DEBUG:PyTR: -- 82 considered consider VERB [62, 65, 66, 68, 70, 71, 72, 74, 77, 79] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13]\n",
      "DEBUG:PyTR:range [10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 10 1\n",
      "DEBUG:PyTR:link 22 13\n",
      "DEBUG:PyTR:prev_tok 9 4\n",
      "DEBUG:PyTR: -- 83 types type NOUN [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22]\n",
      "DEBUG:PyTR:range [11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 11 1\n",
      "DEBUG:PyTR:link 1 22\n",
      "DEBUG:PyTR:prev_tok 10 2\n",
      "DEBUG:PyTR:link 1 13\n",
      "DEBUG:PyTR:prev_tok 9 5\n",
      "DEBUG:PyTR: -- 84 systems system NOUN [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1]\n",
      "DEBUG:PyTR:range [12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 12 2\n",
      "DEBUG:PyTR:link 1 1\n",
      "DEBUG:PyTR:prev_tok 11 3\n",
      "DEBUG:PyTR:link 1 22\n",
      "DEBUG:PyTR:prev_tok 10 4\n",
      "DEBUG:PyTR: -- 86 systems system NOUN [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84, 86] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1, 1]\n",
      "DEBUG:PyTR:range [13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 13 2\n",
      "DEBUG:PyTR:link 29 1\n",
      "DEBUG:PyTR:prev_tok 12 4\n",
      "DEBUG:PyTR: -- 88 mixed mixed ADJ [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84, 86] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1, 1]\n",
      "DEBUG:PyTR:visit [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84, 86, 88] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1, 1, 29]\n",
      "DEBUG:PyTR:range [14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]\n",
      "DEBUG:PyTR:prev_tok 14 1\n",
      "DEBUG:PyTR:link 22 29\n",
      "DEBUG:PyTR:prev_tok 13 3\n",
      "DEBUG:PyTR:link 22 1\n",
      "DEBUG:PyTR:prev_tok 12 5\n",
      "DEBUG:PyTR: -- 89 types type NOUN [62, 65, 66, 68, 70, 71, 72, 74, 77, 79, 82, 83, 84, 86, 88] [7, 24, 19, 25, 17, 26, 4, 18, 27, 28, 13, 22, 1, 1, 29]\n",
      "DEBUG:PyTR:{('compatibility', 'NOUN'): {0, 15}, ('system', 'NOUN'): {2, 18, 84, 86, 57}, ('linear', 'ADJ'): {4, 20}, ('constraint', 'NOUN'): {5}, ('set', 'NOUN'): {8, 40, 72, 50}, ('natural', 'ADJ'): {10}, ('number', 'NOUN'): {11}, ('criterion', 'NOUN'): {13, 62}, ('Diophantine', 'PROPN'): {21}, ('equation', 'NOUN'): {22}, ('strict', 'ADJ'): {24}, ('inequation', 'NOUN'): {25, 29}, ('nonstrict', 'NOUN'): {28}, ('consider', 'VERB'): {82, 31}, ('upper', 'ADJ'): {33}, ('bound', 'NOUN'): {34}, ('component', 'NOUN'): {36}, ('minimal', 'ADJ'): {48, 70, 39}, ('solution', 'NOUN'): {42, 52, 74}, ('algorithm', 'NOUN'): {66, 44}, ('construction', 'NOUN'): {46}, ('generating', 'NOUN'): {49}, ('type', 'NOUN'): {89, 83, 55}, ('give', 'VERB'): {59}, ('corresponding', 'ADJ'): {65}, ('construct', 'VERB'): {68}, ('support', 'VERB'): {71}, ('use', 'VERB'): {77}, ('solve', 'VERB'): {79}, ('mixed', 'ADJ'): {88}}\n"
     ]
    }
   ],
   "source": [
    "lemma_graph = nx.Graph()\n",
    "seen_lemma = {}\n",
    "\n",
    "for sent in doc.sents:\n",
    "    link_sentence(doc, sent, lemma_graph, seen_lemma)\n",
    "    #break # only test one sentence\n",
    "\n",
    "if logger:\n",
    "    logger.debug(seen_lemma)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualize the lemma graph, and for that first we need to collect a dictionary of the labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: 'compatibility',\n",
       " 1: 'system',\n",
       " 2: 'linear',\n",
       " 3: 'constraint',\n",
       " 4: 'set',\n",
       " 5: 'natural',\n",
       " 6: 'number',\n",
       " 7: 'criterion',\n",
       " 8: 'diophantine',\n",
       " 9: 'equation',\n",
       " 10: 'strict',\n",
       " 11: 'inequation',\n",
       " 12: 'nonstrict',\n",
       " 13: 'consider',\n",
       " 14: 'upper',\n",
       " 15: 'bound',\n",
       " 16: 'component',\n",
       " 17: 'minimal',\n",
       " 18: 'solution',\n",
       " 19: 'algorithm',\n",
       " 20: 'construction',\n",
       " 21: 'generating',\n",
       " 22: 'type',\n",
       " 23: 'give',\n",
       " 24: 'corresponding',\n",
       " 25: 'construct',\n",
       " 26: 'support',\n",
       " 27: 'use',\n",
       " 28: 'solve',\n",
       " 29: 'mixed'}"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels = {}\n",
    "keys = list(seen_lemma.keys())\n",
    "\n",
    "for i in range(len(seen_lemma)):\n",
    "    labels[i] = keys[i][0].lower()\n",
    "\n",
    "labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then use `matplotlib` to visualize the lemma graph:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:matplotlib:$HOME=/home/ceteri\n",
      "DEBUG:matplotlib:CONFIGDIR=/home/ceteri/.config/matplotlib\n",
      "DEBUG:matplotlib:matplotlib data path: /home/ceteri/anaconda3/lib/python3.7/site-packages/matplotlib/mpl-data\n",
      "DEBUG:matplotlib:loaded rc file /home/ceteri/anaconda3/lib/python3.7/site-packages/matplotlib/mpl-data/matplotlibrc\n",
      "DEBUG:matplotlib:matplotlib version 3.1.0\n",
      "DEBUG:matplotlib:interactive is False\n",
      "DEBUG:matplotlib:platform is linux\n",
      "DEBUG:matplotlib:loaded modules: ['sys', 'builtins', '_frozen_importlib', '_imp', '_thread', '_warnings', '_weakref', 'zipimport', '_frozen_importlib_external', '_io', 'marshal', 'posix', 'encodings', 'codecs', '_codecs', 'encodings.aliases', 'encodings.utf_8', '_signal', '__main__', 'encodings.latin_1', 'io', 'abc', '_abc', '_bootlocale', '_locale', 'site', 'os', 'stat', '_stat', 'posixpath', 'genericpath', 'os.path', '_collections_abc', '_sitebuiltins', 'types', 'importlib', 'importlib._bootstrap', 'importlib._bootstrap_external', 'warnings', 'importlib.util', 'importlib.abc', 'importlib.machinery', 'contextlib', 'collections', 'operator', '_operator', 'keyword', 'heapq', '_heapq', 'itertools', 'reprlib', '_collections', 'functools', '_functools', 'mpl_toolkits', 'sphinxcontrib', 'runpy', 'pkgutil', 'weakref', '_weakrefset', 'ipykernel', 'ipykernel._version', 'ipykernel.connect', '__future__', 'json', 'json.decoder', 're', 'enum', 'sre_compile', '_sre', 'sre_parse', 'sre_constants', 'copyreg', 'json.scanner', '_json', 'json.encoder', 'subprocess', 'time', 'signal', 'errno', '_posixsubprocess', 'select', 'selectors', 'collections.abc', 'math', 'threading', 'traceback', 'linecache', 'tokenize', 'token', 'IPython', 'IPython.core', 'IPython.core.getipython', 'IPython.core.release', 'IPython.core.application', 'atexit', 'copy', 'glob', 'fnmatch', 'logging', 'string', '_string', 'shutil', 'zlib', 'bz2', '_compression', '_bz2', 'lzma', '_lzma', 'pwd', 'grp', 'traitlets', 'traitlets.traitlets', 'inspect', 'dis', 'opcode', '_opcode', 'six', 'struct', '_struct', 'traitlets.utils', 'traitlets.utils.getargspec', 'traitlets.utils.importstring', 'ipython_genutils', 'ipython_genutils._version', 'ipython_genutils.py3compat', 'ipython_genutils.encoding', 'locale', 'platform', 'traitlets.utils.sentinel', 'traitlets.utils.bunch', 'traitlets._version', 'traitlets.config', 'traitlets.config.application', 'decorator', 'traitlets.config.configurable', 'traitlets.config.loader', 'argparse', 'gettext', 'ast', '_ast', 'ipython_genutils.path', 'random', 'hashlib', '_hashlib', '_blake2', '_sha3', 'bisect', '_bisect', '_random', 'ipython_genutils.text', 'textwrap', 'ipython_genutils.importstring', 'IPython.core.crashhandler', 'pprint', 'IPython.core.ultratb', 'pydoc', 'urllib', 'urllib.parse', 'IPython.core.debugger', 'bdb', 'IPython.utils', 'IPython.utils.PyColorize', 'IPython.utils.coloransi', 'IPython.utils.ipstruct', 'IPython.utils.colorable', 'pygments', 'pygments.util', 'IPython.utils.py3compat', 'IPython.utils.encoding', 'IPython.core.excolors', 'IPython.testing', 'IPython.testing.skipdoctest', 'pdb', 'cmd', 'code', 'codeop', 'IPython.core.display_trap', 'IPython.utils.path', 'IPython.utils.process', 'IPython.utils._process_posix', 'pexpect', 'pexpect.exceptions', 'pexpect.utils', 'pexpect.expect', 'pexpect.pty_spawn', 'pty', 'tty', 'termios', 'ptyprocess', 'ptyprocess.ptyprocess', 'fcntl', 'resource', 'ptyprocess.util', 'pexpect.spawnbase', 'pexpect.run', 'IPython.utils._process_common', 'shlex', 'IPython.utils.decorators', 'IPython.utils.data', 'IPython.utils.terminal', 'IPython.utils.sysinfo', 'IPython.utils._sysinfo', 'IPython.core.profiledir', 'IPython.paths', 'tempfile', 'IPython.utils.importstring', 'IPython.terminal', 'IPython.terminal.embed', 'IPython.core.compilerop', 'IPython.core.magic_arguments', 'IPython.core.error', 'IPython.utils.text', 'pathlib', 'ntpath', 'IPython.core.magic', 'getopt', 'IPython.core.oinspect', 'IPython.core.page', 'IPython.core.display', 'binascii', 'mimetypes', 'IPython.lib', 'IPython.lib.security', 'getpass', 'IPython.lib.pretty', 'datetime', '_datetime', 'IPython.utils.openpy', 'IPython.utils.dir2', 'IPython.utils.wildcard', 'pygments.lexers', 'pygments.lexers._mapping', 'pygments.modeline', 'pygments.plugin', 'pygments.lexers.python', 'pygments.lexer', 'pygments.filter', 'pygments.filters', 'pygments.token', 'pygments.regexopt', 'pygments.unistring', 'pygments.formatters', 'pygments.formatters._mapping', 'pygments.formatters.html', 'pygments.formatter', 'pygments.styles', 'IPython.core.inputtransformer2', 'typing', 'typing.io', 'typing.re', 'IPython.core.interactiveshell', 'asyncio', 'asyncio.base_events', 'concurrent', 'concurrent.futures', 'concurrent.futures._base', 'socket', '_socket', 'ssl', '_ssl', 'base64', 'asyncio.constants', 'asyncio.coroutines', 'asyncio.base_futures', 'asyncio.format_helpers', 'asyncio.log', 'asyncio.events', 'contextvars', '_contextvars', 'asyncio.base_tasks', '_asyncio', 'asyncio.futures', 'asyncio.protocols', 'asyncio.sslproto', 'asyncio.transports', 'asyncio.tasks', 'asyncio.locks', 'asyncio.runners', 'asyncio.queues', 'asyncio.streams', 'asyncio.subprocess', 'asyncio.unix_events', 'asyncio.base_subprocess', 'asyncio.selector_events', 'pickleshare', 'pickle', '_compat_pickle', '_pickle', 'IPython.core.prefilter', 'IPython.core.autocall', 'IPython.core.macro', 'IPython.core.splitinput', 'IPython.core.alias', 'IPython.core.builtin_trap', 'IPython.core.events', 'backcall', 'backcall.backcall', 'IPython.core.displayhook', 'IPython.core.displaypub', 'IPython.core.extensions', 'IPython.core.formatters', 'IPython.utils.sentinel', 'IPython.core.history', 'sqlite3', 'sqlite3.dbapi2', '_sqlite3', 'IPython.core.logger', 'IPython.core.payload', 'IPython.core.usage', 'IPython.display', 'IPython.lib.display', 'html', 'html.entities', 'IPython.utils.io', 'IPython.utils.capture', 'IPython.utils.strdispatch', 'IPython.core.hooks', 'IPython.utils.syspathcontext', 'IPython.utils.tempdir', 'IPython.utils.contexts', 'IPython.core.async_helpers', 'IPython.terminal.interactiveshell', 'prompt_toolkit', 'prompt_toolkit.application', 'prompt_toolkit.application.application', 'prompt_toolkit.buffer', 'prompt_toolkit.application.current', 'prompt_toolkit.eventloop', 'prompt_toolkit.eventloop.base', 'prompt_toolkit.log', 'prompt_toolkit.eventloop.coroutine', 'prompt_toolkit.eventloop.defaults', 'prompt_toolkit.utils', 'six.moves', 'wcwidth', 'wcwidth.wcwidth', 'wcwidth.table_wide', 'wcwidth.table_zero', 'prompt_toolkit.cache', 'prompt_toolkit.eventloop.future', 'prompt_toolkit.eventloop.context', 'prompt_toolkit.eventloop.async_generator', 'queue', '_queue', 'six.moves.queue', 'prompt_toolkit.eventloop.event', 'prompt_toolkit.application.run_in_terminal', 'prompt_toolkit.auto_suggest', 'prompt_toolkit.filters', 'prompt_toolkit.filters.base', 'prompt_toolkit.filters.app', 'prompt_toolkit.enums', 'prompt_toolkit.filters.utils', 'prompt_toolkit.filters.cli', 'prompt_toolkit.clipboard', 'prompt_toolkit.clipboard.base', 'prompt_toolkit.selection', 'prompt_toolkit.clipboard.in_memory', 'prompt_toolkit.completion', 'prompt_toolkit.completion.base', 'prompt_toolkit.completion.filesystem', 'prompt_toolkit.completion.word_completer', 'prompt_toolkit.completion.fuzzy_completer', 'prompt_toolkit.document', 'prompt_toolkit.history', 'prompt_toolkit.search', 'prompt_toolkit.key_binding', 'prompt_toolkit.key_binding.key_bindings', 'prompt_toolkit.keys', 'prompt_toolkit.key_binding.vi_state', 'prompt_toolkit.validation', 'prompt_toolkit.input', 'prompt_toolkit.input.base', 'prompt_toolkit.input.defaults', 'prompt_toolkit.input.typeahead', 'prompt_toolkit.key_binding.bindings', 'prompt_toolkit.key_binding.bindings.page_navigation', 'prompt_toolkit.key_binding.bindings.scroll', 'prompt_toolkit.key_binding.defaults', 'prompt_toolkit.key_binding.bindings.basic', 'prompt_toolkit.key_binding.key_processor', 'prompt_toolkit.key_binding.bindings.named_commands', 'prompt_toolkit.key_binding.bindings.completion', 'prompt_toolkit.key_binding.bindings.emacs', 'prompt_toolkit.key_binding.bindings.vi', 'prompt_toolkit.input.vt100_parser', 'prompt_toolkit.input.ansi_escape_sequences', 'prompt_toolkit.key_binding.digraphs', 'prompt_toolkit.key_binding.bindings.mouse', 'prompt_toolkit.layout', 'prompt_toolkit.layout.containers', 'prompt_toolkit.layout.controls', 'prompt_toolkit.formatted_text', 'prompt_toolkit.formatted_text.base', 'prompt_toolkit.formatted_text.html', 'xml', 'xml.dom', 'xml.dom.domreg', 'xml.dom.minidom', 'xml.dom.minicompat', 'xml.dom.xmlbuilder', 'xml.dom.NodeFilter', 'prompt_toolkit.formatted_text.ansi', 'prompt_toolkit.output', 'prompt_toolkit.output.base', 'prompt_toolkit.layout.screen', 'prompt_toolkit.output.defaults', 'prompt_toolkit.output.color_depth', 'prompt_toolkit.output.vt100', 'prompt_toolkit.styles', 'prompt_toolkit.styles.base', 'prompt_toolkit.styles.defaults', 'prompt_toolkit.styles.style', 'prompt_toolkit.styles.named_colors', 'prompt_toolkit.styles.pygments', 'prompt_toolkit.styles.style_transformation', 'colorsys', 'array', 'prompt_toolkit.formatted_text.pygments', 'prompt_toolkit.formatted_text.utils', 'prompt_toolkit.lexers', 'prompt_toolkit.lexers.base', 'prompt_toolkit.lexers.pygments', 'prompt_toolkit.mouse_events', 'prompt_toolkit.layout.processors', 'prompt_toolkit.layout.utils', 'prompt_toolkit.layout.dimension', 'prompt_toolkit.layout.margins', 'prompt_toolkit.layout.layout', 'prompt_toolkit.layout.menus', 'prompt_toolkit.renderer', 'prompt_toolkit.layout.mouse_handlers', 'prompt_toolkit.key_binding.bindings.cpr', 'prompt_toolkit.key_binding.emacs_state', 'prompt_toolkit.layout.dummy', 'prompt_toolkit.application.dummy', 'prompt_toolkit.shortcuts', 'prompt_toolkit.shortcuts.dialogs', 'prompt_toolkit.key_binding.bindings.focus', 'prompt_toolkit.widgets', 'prompt_toolkit.widgets.base', 'prompt_toolkit.widgets.toolbars', 'prompt_toolkit.widgets.dialogs', 'prompt_toolkit.widgets.menus', 'prompt_toolkit.shortcuts.prompt', 'prompt_toolkit.key_binding.bindings.auto_suggest', 'prompt_toolkit.key_binding.bindings.open_in_editor', 'prompt_toolkit.shortcuts.utils', 'prompt_toolkit.shortcuts.progress_bar', 'prompt_toolkit.shortcuts.progress_bar.base', 'prompt_toolkit.shortcuts.progress_bar.formatters', 'prompt_toolkit.patch_stdout', 'pygments.style', 'IPython.terminal.debugger', 'IPython.core.completer', 'unicodedata', 'IPython.core.latex_symbols', 'IPython.utils.generics', 'jedi', 'jedi.api', 'parso', 'parso.parser', 'parso.tree', 'parso._compatibility', 'parso.utils', 'parso.pgen2', 'parso.pgen2.generator', 'parso.pgen2.grammar_parser', 'parso.python', 'parso.python.tokenize', 'parso.python.token', 'parso.grammar', 'parso.python.diff', 'difflib', 'parso.python.parser', 'parso.python.tree', 'parso.python.prefix', 'parso.cache', 'gc', 'parso.python.errors', 'parso.normalizer', 'parso.python.pep8', 'parso.file_io', 'jedi._compatibility', 'jedi.parser_utils', 'jedi.debug', 'colorama', 'colorama.initialise', 'colorama.ansitowin32', 'colorama.ansi', 'colorama.winterm', 'colorama.win32', 'ctypes', '_ctypes', 'ctypes._endian', 'jedi.settings', 'jedi.cache', 'jedi.api.classes', 'jedi.evaluate', 'jedi.evaluate.utils', 'jedi.evaluate.imports', 'jedi.evaluate.sys_path', 'jedi.evaluate.cache', 'jedi.evaluate.base_context', 'jedi.common', 'jedi.common.context', 'jedi.evaluate.helpers', 'jedi.common.utils', 'jedi.evaluate.compiled', 'jedi.evaluate.compiled.context', 'jedi.evaluate.filters', 'jedi.evaluate.flow_analysis', 'jedi.evaluate.recursion', 'jedi.evaluate.lazy_context', 'jedi.evaluate.compiled.access', 'jedi.evaluate.compiled.getattr_static', 'jedi.evaluate.compiled.fake', 'jedi.evaluate.analysis', 'jedi.evaluate.context', 'jedi.evaluate.context.module', 'jedi.evaluate.context.klass', 'jedi.evaluate.context.function', 'jedi.evaluate.docstrings', 'jedi.evaluate.pep0484', 'jedi.evaluate.arguments', 'jedi.evaluate.context.iterable', 'jedi.evaluate.param', 'jedi.evaluate.context.asynchronous', 'jedi.evaluate.parser_cache', 'jedi.evaluate.context.instance', 'jedi.evaluate.syntax_tree', 'jedi.evaluate.finder', 'jedi.api.keywords', 'pydoc_data', 'pydoc_data.topics', 'jedi.api.interpreter', 'jedi.evaluate.compiled.mixed', 'jedi.api.helpers', 'jedi.api.completion', 'jedi.api.environment', 'filecmp', 'jedi.evaluate.compiled.subprocess', 'jedi.evaluate.compiled.subprocess.functions', 'jedi.api.exceptions', 'jedi.api.project', 'jedi.evaluate.usages', 'IPython.terminal.ptutils', 'IPython.terminal.shortcuts', 'IPython.terminal.magics', 'IPython.lib.clipboard', 'IPython.terminal.pt_inputhooks', 'IPython.terminal.prompts', 'IPython.terminal.ipapp', 'IPython.core.magics', 'IPython.core.magics.auto', 'IPython.core.magics.basic', 'IPython.core.magics.code', 'urllib.request', 'email', 'http', 'http.client', 'email.parser', 'email.feedparser', 'email.errors', 'email._policybase', 'email.header', 'email.quoprimime', 'email.base64mime', 'email.charset', 'email.encoders', 'quopri', 'email.utils', 'email._parseaddr', 'calendar', 'email.message', 'uu', 'email._encoded_words', 'email.iterators', 'urllib.error', 'urllib.response', 'IPython.core.magics.config', 'IPython.core.magics.display', 'IPython.core.magics.execution', 'timeit', 'cProfile', '_lsprof', 'profile', 'pstats', 'IPython.utils.module_paths', 'IPython.utils.timing', 'IPython.core.magics.extension', 'IPython.core.magics.history', 'IPython.core.magics.logging', 'IPython.core.magics.namespace', 'IPython.core.magics.osm', 'IPython.core.magics.packaging', 'IPython.core.magics.pylab', 'IPython.core.pylabtools', 'IPython.core.magics.script', 'IPython.lib.backgroundjobs', 'IPython.core.shellapp', 'IPython.extensions', 'IPython.extensions.storemagic', 'IPython.utils.frame', 'jupyter_client', 'jupyter_client._version', 'jupyter_client.connect', 'zmq', 'zmq.backend', 'zmq.backend.select', 'zmq.backend.cython', 'zmq.backend.cython.constants', 'cython_runtime', 'zmq.backend.cython.error', '_cython_0_29_5', 'zmq.backend.cython.message', 'zmq.error', 'zmq.backend.cython.context', 'zmq.backend.cython.socket', 'zmq.backend.cython.utils', 'zmq.backend.cython._poll', 'zmq.backend.cython._version', 'zmq.backend.cython._device', 'zmq.backend.cython._proxy_steerable', 'zmq.sugar', 'zmq.sugar.constants', 'zmq.utils', 'zmq.utils.constant_names', 'zmq.sugar.context', 'zmq.sugar.attrsettr', 'zmq.sugar.socket', 'zmq.sugar.poll', 'zmq.utils.jsonapi', 'zmq.utils.strtypes', 'zmq.sugar.frame', 'zmq.sugar.tracker', 'zmq.sugar.version', 'zmq.sugar.stopwatch', 'jupyter_client.localinterfaces', 'jupyter_core', 'jupyter_core.version', 'jupyter_core.paths', 'jupyter_client.launcher', 'traitlets.log', 'jupyter_client.client', 'jupyter_client.channels', 'jupyter_client.channelsabc', 'jupyter_client.clientabc', 'jupyter_client.manager', 'jupyter_client.kernelspec', 'jupyter_client.managerabc', 'jupyter_client.blocking', 'jupyter_client.blocking.client', 'jupyter_client.blocking.channels', 'jupyter_client.multikernelmanager', 'uuid', 'ipykernel.kernelapp', 'tornado', 'tornado.ioloop', 'numbers', 'tornado.concurrent', 'tornado.log', 'logging.handlers', 'tornado.escape', 'tornado.util', 'tornado.speedups', 'curses', '_curses', 'zmq.eventloop', 'zmq.eventloop.ioloop', 'tornado.platform', 'tornado.platform.asyncio', 'tornado.gen', 'zmq.eventloop.zmqstream', 'ipykernel.iostream', 'imp', 'jupyter_client.session', 'hmac', 'jupyter_client.jsonutil', 'dateutil', 'dateutil._version', 'dateutil.parser', 'dateutil.parser._parser', 'decimal', '_decimal', 'dateutil.relativedelta', 'dateutil._common', 'dateutil.tz', 'dateutil.tz.tz', 'dateutil.tz._common', 'dateutil.tz._factories', 'dateutil.parser.isoparser', '_strptime', 'jupyter_client.adapter', 'ipykernel.heartbeat', 'ipykernel.ipkernel', 'IPython.utils.tokenutil', 'ipykernel.comm', 'ipykernel.comm.manager', 'ipykernel.comm.comm', 'ipykernel.kernelbase', 'tornado.queues', 'tornado.locks', 'ipykernel.jsonutil', 'ipykernel.zmqshell', 'IPython.core.payloadpage', 'ipykernel.displayhook', 'ipykernel.parentpoller', 'faulthandler', 'ipykernel.datapub', 'ipykernel.serialize', 'ipykernel.pickleutil', 'ipykernel.codeutil', 'IPython.core.completerlib', 'storemagic', 'ipywidgets', 'ipywidgets._version', 'ipywidgets.widgets', 'ipywidgets.widgets.widget', 'ipywidgets.widgets.domwidget', 'ipywidgets.widgets.trait_types', 'ipywidgets.widgets.util', 'ipywidgets.widgets.widget_layout', 'ipywidgets.widgets.widget_style', 'ipywidgets.widgets.valuewidget', 'ipywidgets.widgets.widget_core', 'ipywidgets.widgets.widget_bool', 'ipywidgets.widgets.widget_description', 'ipywidgets.widgets.widget_button', 'ipywidgets.widgets.widget_box', 'ipywidgets.widgets.docutils', 'ipywidgets.widgets.widget_float', 'ipywidgets.widgets.widget_int', 'ipywidgets.widgets.widget_color', 'ipywidgets.widgets.widget_date', 'ipywidgets.widgets.widget_output', 'ipywidgets.widgets.widget_selection', 'ipywidgets.widgets.widget_selectioncontainer', 'ipywidgets.widgets.widget_string', 'ipywidgets.widgets.widget_controller', 'ipywidgets.widgets.interaction', 'ipywidgets.widgets.widget_link', 'ipywidgets.widgets.widget_media', 'ipywidgets.widgets.widget_templates', 'ipywidgets.widgets.widget_upload', 'spacy', 'thinc', 'numpy', 'numpy._globals', 'numpy.__config__', 'numpy.version', 'numpy._distributor_init', 'numpy._mklinit', 'numpy.core', 'numpy.core.info', 'numpy.core.multiarray', 'numpy.core.overrides', 'numpy.core._multiarray_umath', 'numpy.compat', 'numpy.compat._inspect', 'numpy.compat.py3k', 'numpy.core.umath', 'numpy.core.numerictypes', 'numpy.core._string_helpers', 'numpy.core._type_aliases', 'numpy.core._dtype', 'numpy.core.numeric', 'numpy.core._internal', 'numpy.core.fromnumeric', 'numpy.core._methods', 'numpy.core.arrayprint', 'numpy.core.defchararray', 'numpy.core.records', 'numpy.core.memmap', 'numpy.core.function_base', 'numpy.core.machar', 'numpy.core.getlimits', 'numpy.core.shape_base', 'numpy.core.einsumfunc', 'numpy.core._add_newdocs', 'numpy.core._multiarray_tests', 'numpy.core._dtype_ctypes', 'numpy._pytesttester', 'numpy.lib', 'numpy.lib.info', 'numpy.lib.type_check', 'numpy.lib.ufunclike', 'numpy.lib.index_tricks', 'numpy.matrixlib', 'numpy.matrixlib.defmatrix', 'numpy.linalg', 'numpy.linalg.info', 'numpy.linalg.linalg', 'numpy.lib.twodim_base', 'numpy.linalg.lapack_lite', 'numpy.linalg._umath_linalg', 'numpy.lib.function_base', 'numpy.lib.utils', 'numpy.lib.histograms', 'numpy.lib.stride_tricks', 'numpy.lib.mixins', 'numpy.lib.nanfunctions', 'numpy.lib.shape_base', 'numpy.lib.scimath', 'numpy.lib.polynomial', 'numpy.lib.arraysetops', 'numpy.lib.npyio', 'numpy.lib.format', 'numpy.lib._datasource', 'numpy.lib._iotools', 'numpy.lib.financial', 'numpy.lib.arrayterator', 'numpy.lib.arraypad', 'numpy.lib._version', 'numpy.fft', 'numpy.fft.info', 'numpy.fft.fftpack', 'numpy.fft.fftpack_lite', 'numpy.fft.helper', 'mkl_fft', 'mkl_fft._pydfti', '_cython_0_29_7', 'mkl_fft._version', 'mkl_fft._numpy_fft', 'numpy.polynomial', 'numpy.polynomial.polynomial', 'numpy.polynomial.polyutils', 'numpy.polynomial._polybase', 'numpy.polynomial.chebyshev', 'numpy.polynomial.legendre', 'numpy.polynomial.hermite', 'numpy.polynomial.hermite_e', 'numpy.polynomial.laguerre', 'numpy.random', 'numpy.random.mtrand', 'mtrand', 'numpy.ctypeslib', 'numpy.ma', 'numpy.ma.core', 'numpy.ma.extras', 'numpy.testing', 'unittest', 'unittest.result', 'unittest.util', 'unittest.case', 'unittest.suite', 'unittest.loader', 'unittest.main', 'unittest.runner', 'unittest.signals', 'numpy.testing._private', 'numpy.testing._private.utils', 'numpy.testing._private.decorators', 'numpy.testing._private.nosetester', 'thinc.about', 'thinc.neural', 'thinc.neural._classes', 'thinc.neural._classes.model', 'srsly', 'srsly._json_api', 'gzip', 'srsly.ujson', 'srsly.ujson.ujson', 'srsly.util', 'srsly._msgpack_api', 'srsly.msgpack', 'srsly.msgpack._version', 'srsly.msgpack.exceptions', 'srsly.msgpack._packer', 'srsly.msgpack._ext_type', 'srsly.msgpack.util', 'srsly.msgpack._unpacker', 'srsly.msgpack._msgpack_numpy', 'srsly._pickle_api', 'srsly.cloudpickle', 'srsly.cloudpickle.cloudpickle', 'thinc.neural.util', 'thinc.neural.train', 'tqdm', 'tqdm._tqdm', 'tqdm._utils', 'tqdm._monitor', 'multiprocessing', 'multiprocessing.context', 'multiprocessing.process', 'multiprocessing.reduction', '__mp_main__', 'tqdm._tqdm_gui', 'tqdm._tqdm_pandas', 'tqdm._main', 'tqdm._version', 'thinc.neural.optimizers', '_cython_0_29_13', 'thinc.neural.ops', 'cymem', 'cymem.about', 'cymem.cymem', 'preshed', 'preshed.about', 'preshed.maps', 'thinc.linalg', 'blis', 'blis.cy', 'blis.py', 'murmurhash', 'murmurhash.about', 'murmurhash.mrmr', 'thinc.neural._aligned_alloc', 'thinc.compat', 'thinc.neural.mem', 'thinc.check', 'thinc.extra', 'thinc.extra.wrapt', 'thinc.extra.wrapt.wrappers', 'thinc.extra.wrapt._wrappers', 'thinc.extra.wrapt.decorators', 'thinc.extra.wrapt.importer', 'thinc.exceptions', 'wasabi', 'wasabi.printer', 'wasabi.tables', 'wasabi.util', 'wasabi.traceback', 'spacy.cli', 'spacy.cli.download', 'plac', 'plac_core', 'plac_ext', 'asyncore', 'asynchat', 'requests', 'urllib3', 'urllib3.connectionpool', 'urllib3.exceptions', 'urllib3.packages', 'urllib3.packages.ssl_match_hostname', 'urllib3.packages.six', 'urllib3.packages.six.moves', 'urllib3.packages.six.moves.http_client', 'urllib3.connection', 'urllib3.util', 'urllib3.util.connection', 'urllib3.util.wait', 'urllib3.contrib', 'urllib3.contrib._appengine_environ', 'urllib3.util.request', 'urllib3.util.response', 'urllib3.util.ssl_', 'urllib3.util.timeout', 'urllib3.util.retry', 'urllib3.util.url', 'urllib3._collections', 'urllib3.request', 'urllib3.filepost', 'urllib3.fields', 'urllib3.packages.six.moves.urllib', 'urllib3.packages.six.moves.urllib.parse', 'urllib3.response', 'urllib3.util.queue', 'urllib3.poolmanager', 'chardet', 'chardet.compat', 'chardet.universaldetector', 'chardet.charsetgroupprober', 'chardet.enums', 'chardet.charsetprober', 'chardet.escprober', 'chardet.codingstatemachine', 'chardet.escsm', 'chardet.latin1prober', 'chardet.mbcsgroupprober', 'chardet.utf8prober', 'chardet.mbcssm', 'chardet.sjisprober', 'chardet.mbcharsetprober', 'chardet.chardistribution', 'chardet.euctwfreq', 'chardet.euckrfreq', 'chardet.gb2312freq', 'chardet.big5freq', 'chardet.jisfreq', 'chardet.jpcntx', 'chardet.eucjpprober', 'chardet.gb2312prober', 'chardet.euckrprober', 'chardet.cp949prober', 'chardet.big5prober', 'chardet.euctwprober', 'chardet.sbcsgroupprober', 'chardet.sbcharsetprober', 'chardet.langcyrillicmodel', 'chardet.langgreekmodel', 'chardet.langbulgarianmodel', 'chardet.langthaimodel', 'chardet.langhebrewmodel', 'chardet.hebrewprober', 'chardet.langturkishmodel', 'chardet.version', 'requests.exceptions', 'urllib3.contrib.pyopenssl', 'OpenSSL', 'OpenSSL.crypto', 'cryptography', 'cryptography.__about__', 'cryptography.x509', 'cryptography.x509.certificate_transparency', 'cryptography.x509.base', 'cryptography.utils', 'cryptography.hazmat', 'cryptography.hazmat.primitives', 'cryptography.hazmat.primitives.asymmetric', 'cryptography.hazmat.primitives.asymmetric.dsa', 'cryptography.hazmat.primitives.asymmetric.ec', 'cryptography.hazmat._oid', 'cryptography.hazmat.primitives.asymmetric.rsa', 'cryptography.exceptions', 'cryptography.hazmat.backends', 'cryptography.hazmat.backends.interfaces', 'cryptography.x509.extensions', 'ipaddress', 'asn1crypto', 'asn1crypto.version', 'asn1crypto.keys', 'asn1crypto._elliptic_curve', 'asn1crypto._int', 'asn1crypto.util', 'asn1crypto._errors', 'asn1crypto._iri', 'encodings.idna', 'stringprep', 'asn1crypto._types', 'asn1crypto._ordereddict', 'asn1crypto._ffi', 'asn1crypto._perf', 'asn1crypto._perf._big_num_ctypes', 'ctypes.util', 'asn1crypto.algos', 'asn1crypto.core', 'asn1crypto._teletex_codec', 'asn1crypto.parser', 'cryptography.hazmat.primitives.constant_time', 'cryptography.hazmat.bindings', '_cffi_backend', '_constant_time.lib', '_constant_time', 'cryptography.hazmat.bindings._constant_time', 'cryptography.hazmat.primitives.serialization', 'cryptography.hazmat.primitives.serialization.base', 'cryptography.hazmat.primitives.serialization.ssh', 'cryptography.hazmat.primitives.asymmetric.ed25519', 'cryptography.x509.general_name', 'cryptography.x509.name', 'cryptography.x509.oid', 'cryptography.hazmat.primitives.hashes', 'OpenSSL._util', 'cryptography.hazmat.bindings.openssl', 'cryptography.hazmat.bindings.openssl.binding', '_openssl.lib', '_openssl', 'cryptography.hazmat.bindings._openssl', 'cryptography.hazmat.bindings.openssl._conditional', 'OpenSSL.SSL', 'OpenSSL.version', 'cryptography.hazmat.backends.openssl', 'cryptography.hazmat.backends.openssl.backend', 'cryptography.hazmat.backends.openssl.aead', 'cryptography.hazmat.backends.openssl.ciphers', 'cryptography.hazmat.primitives.ciphers', 'cryptography.hazmat.primitives.ciphers.base', 'cryptography.hazmat.primitives.ciphers.modes', 'cryptography.hazmat.backends.openssl.cmac', 'cryptography.hazmat.backends.openssl.decode_asn1', 'cryptography.hazmat.backends.openssl.dh', 'cryptography.hazmat.primitives.asymmetric.dh', 'cryptography.hazmat.backends.openssl.dsa', 'cryptography.hazmat.backends.openssl.utils', 'cryptography.hazmat.primitives.asymmetric.utils', 'cryptography.hazmat.backends.openssl.ec', 'cryptography.hazmat.backends.openssl.ed25519', 'cryptography.hazmat.backends.openssl.ed448', 'cryptography.hazmat.primitives.asymmetric.ed448', 'cryptography.hazmat.backends.openssl.encode_asn1', 'cryptography.hazmat.backends.openssl.hashes', 'cryptography.hazmat.backends.openssl.hmac', 'cryptography.hazmat.backends.openssl.ocsp', 'cryptography.hazmat.backends.openssl.x509', 'cryptography.x509.ocsp', 'cryptography.hazmat.backends.openssl.poly1305', 'cryptography.hazmat.backends.openssl.rsa', 'cryptography.hazmat.primitives.asymmetric.padding', 'cryptography.hazmat.backends.openssl.x25519', 'cryptography.hazmat.primitives.asymmetric.x25519', 'cryptography.hazmat.backends.openssl.x448', 'cryptography.hazmat.primitives.asymmetric.x448', 'cryptography.hazmat.primitives.ciphers.algorithms', 'cryptography.hazmat.primitives.kdf', 'cryptography.hazmat.primitives.kdf.scrypt', 'urllib3.packages.backports', 'urllib3.packages.backports.makefile', 'requests.__version__', 'requests.utils', 'zipfile', 'requests.certs', 'certifi', 'certifi.core', 'requests._internal_utils', 'requests.compat', 'http.cookiejar', 'http.cookies', 'requests.cookies', 'requests.structures', 'requests.packages', 'requests.packages.urllib3', 'requests.packages.urllib3.connectionpool', 'requests.packages.urllib3.exceptions', 'requests.packages.urllib3.packages', 'requests.packages.urllib3.packages.ssl_match_hostname', 'requests.packages.urllib3.packages.six', 'requests.packages.urllib3.packages.six.moves', 'requests.packages.urllib3.packages.six.moves.http_client', 'requests.packages.urllib3.connection', 'requests.packages.urllib3.util', 'requests.packages.urllib3.util.connection', 'requests.packages.urllib3.util.wait', 'requests.packages.urllib3.contrib', 'requests.packages.urllib3.contrib._appengine_environ', 'requests.packages.urllib3.util.request', 'requests.packages.urllib3.util.response', 'requests.packages.urllib3.util.ssl_', 'requests.packages.urllib3.util.timeout', 'requests.packages.urllib3.util.retry', 'requests.packages.urllib3.util.url', 'requests.packages.urllib3._collections', 'requests.packages.urllib3.request', 'requests.packages.urllib3.filepost', 'requests.packages.urllib3.fields', 'requests.packages.urllib3.packages.six.moves.urllib', 'requests.packages.urllib3.packages.six.moves.urllib.parse', 'requests.packages.urllib3.response', 'requests.packages.urllib3.util.queue', 'requests.packages.urllib3.poolmanager', 'requests.packages.urllib3.contrib.pyopenssl', 'requests.packages.urllib3.packages.backports', 'requests.packages.urllib3.packages.backports.makefile', 'idna', 'idna.package_data', 'idna.core', 'idna.idnadata', 'idna.intranges', 'requests.packages.idna', 'requests.packages.idna.package_data', 'requests.packages.idna.core', 'requests.packages.idna.idnadata', 'requests.packages.idna.intranges', 'requests.packages.chardet', 'requests.packages.chardet.compat', 'requests.packages.chardet.universaldetector', 'requests.packages.chardet.charsetgroupprober', 'requests.packages.chardet.enums', 'requests.packages.chardet.charsetprober', 'requests.packages.chardet.escprober', 'requests.packages.chardet.codingstatemachine', 'requests.packages.chardet.escsm', 'requests.packages.chardet.latin1prober', 'requests.packages.chardet.mbcsgroupprober', 'requests.packages.chardet.utf8prober', 'requests.packages.chardet.mbcssm', 'requests.packages.chardet.sjisprober', 'requests.packages.chardet.mbcharsetprober', 'requests.packages.chardet.chardistribution', 'requests.packages.chardet.euctwfreq', 'requests.packages.chardet.euckrfreq', 'requests.packages.chardet.gb2312freq', 'requests.packages.chardet.big5freq', 'requests.packages.chardet.jisfreq', 'requests.packages.chardet.jpcntx', 'requests.packages.chardet.eucjpprober', 'requests.packages.chardet.gb2312prober', 'requests.packages.chardet.euckrprober', 'requests.packages.chardet.cp949prober', 'requests.packages.chardet.big5prober', 'requests.packages.chardet.euctwprober', 'requests.packages.chardet.sbcsgroupprober', 'requests.packages.chardet.sbcharsetprober', 'requests.packages.chardet.langcyrillicmodel', 'requests.packages.chardet.langgreekmodel', 'requests.packages.chardet.langbulgarianmodel', 'requests.packages.chardet.langthaimodel', 'requests.packages.chardet.langhebrewmodel', 'requests.packages.chardet.hebrewprober', 'requests.packages.chardet.langturkishmodel', 'requests.packages.chardet.version', 'requests.models', 'requests.hooks', 'requests.auth', 'requests.status_codes', 'requests.api', 'requests.sessions', 'requests.adapters', 'urllib3.contrib.socks', 'socks', 'pkg_resources', 'plistlib', 'xml.parsers', 'xml.parsers.expat', 'pyexpat.errors', 'pyexpat.model', 'pyexpat', 'xml.parsers.expat.model', 'xml.parsers.expat.errors', 'pkg_resources.extern', 'pkg_resources._vendor', 'pkg_resources.extern.six', 'pkg_resources._vendor.six', 'pkg_resources.extern.six.moves', 'pkg_resources._vendor.six.moves', 'pkg_resources.py31compat', 'pkg_resources.extern.appdirs', 'pkg_resources._vendor.packaging.__about__', 'pkg_resources.extern.packaging', 'pkg_resources.extern.packaging.version', 'pkg_resources.extern.packaging._structures', 'pkg_resources.extern.packaging.specifiers', 'pkg_resources.extern.packaging._compat', 'pkg_resources.extern.packaging.requirements', 'pkg_resources.extern.pyparsing', 'pkg_resources.extern.six.moves.urllib', 'pkg_resources.extern.packaging.markers', 'sysconfig', 'spacy.cli.link', 'spacy.compat', 'spacy.util', 'jsonschema', 'jsonschema.exceptions', 'attr', 'attr.converters', 'attr._make', 'attr._config', 'attr._compat', 'attr.exceptions', 'attr.filters', 'attr.validators', 'attr._funcs', 'jsonschema._utils', 'jsonschema.compat', 'jsonschema._format', 'jsonschema._types', 'pyrsistent', 'pyrsistent._pmap', 'pyrsistent._compat', 'pyrsistent._pvector', 'pyrsistent._transformations', 'pvectorc', 'pyrsistent._pset', 'pyrsistent._pbag', 'pyrsistent._plist', 'pyrsistent._pdeque', 'pyrsistent._checked_types', 'pyrsistent._field_common', 'pyrsistent._precord', 'pyrsistent._pclass', 'pyrsistent._immutable', 'pyrsistent._helpers', 'pyrsistent._toolz', 'jsonschema.validators', 'jsonschema._legacy_validators', 'jsonschema._validators', 'spacy.symbols', 'spacy.errors', 'spacy.about', 'spacy.cli.info', 'spacy.cli.package', 'spacy.cli.profile', 'thinc.extra.datasets', 'csv', '_csv', 'thinc.extra._vendorized', 'thinc.extra._vendorized.keras_data_utils', 'tarfile', 'thinc.extra._vendorized.keras_generic_utils', 'spacy.cli.train', 'spacy._ml', 'thinc.v2v', 'thinc.neural._classes.affine', 'thinc.describe', 'thinc.neural._classes.relu', 'thinc.neural._classes.maxout', 'thinc.neural._classes.softmax', 'thinc.neural._classes.selu', 'thinc.i2v', 'thinc.neural._classes.hash_embed', 'thinc.neural._lsuv', 'thinc.neural._classes.embed', 'thinc.neural._classes.static_vectors', 'thinc.extra.load_nlp', 'thinc.t2t', 'thinc.neural._classes.convolution', 'thinc.neural._classes.attention', 'thinc.neural._classes.rnn', 'thinc.api', 'thinc.neural._classes.function_layer', 'thinc.neural._classes.feed_forward', 'thinc.neural._classes.multiheaded_attention', 'thinc.t2v', 'thinc.neural.pooling', 'thinc.misc', 'thinc.neural._classes.batchnorm', 'thinc.neural._classes.layernorm', 'thinc.neural._classes.resnet', 'thinc.neural._classes.feature_extracter', 'thinc.linear', 'thinc.linear.linear', 'spacy.attrs', 'torch', 'torch._utils', 'torch._utils_internal', 'torch.version', 'torch._six', 'torch._C._onnx', 'torch._C._jit_tree_views', 'torch._C.cpp', 'torch._C.cpp.nn', 'torch._C', 'torch.random', 'torch.serialization', 'torch._tensor_str', 'torch.tensor', 'torch._namedtensor_internals', 'torch.utils', 'torch.utils.throughput_benchmark', 'torch.utils.hooks', 'torch.storage', 'torch.cuda', 'multiprocessing.util', 'torch.cuda._utils', 'torch.cuda.random', 'torch.cuda.sparse', 'torch.cuda.profiler', 'torch.cuda.nvtx', 'torch.cuda.streams', 'torch.sparse', 'torch.functional', 'torch.nn', 'torch.nn.modules', 'torch.nn.modules.module', 'torch.nn.parameter', 'torch.nn.modules.linear', 'torch.nn.functional', 'torch.nn._reduction', 'torch.nn.modules.utils', 'torch.nn.grad', 'torch.nn._VF', 'torch._jit_internal', 'torch.nn.init', 'torch.nn.modules.conv', 'torch.nn.modules.activation', 'torch.nn.modules.loss', 'torch.nn.modules.container', 'torch.nn.modules.pooling', 'torch.nn.modules.batchnorm', 'torch.nn.modules._functions', 'torch.autograd', 'torch.autograd.variable', 'torch.autograd.function', 'torch.autograd.gradcheck', 'torch.testing', 'torch.autograd.grad_mode', 'torch.autograd.anomaly_mode', 'torch.autograd.profiler', 'torch.nn.modules.instancenorm', 'torch.nn.modules.normalization', 'torch.nn.modules.dropout', 'torch.nn.modules.padding', 'torch.nn.modules.sparse', 'torch.nn.modules.rnn', 'torch.nn.utils', 'torch.nn.utils.rnn', 'torch.nn.utils.clip_grad', 'torch.nn.utils.weight_norm', 'torch.nn.utils.convert_parameters', 'torch.nn.utils.spectral_norm', 'torch.nn.utils.fusion', 'torch.nn.modules.pixelshuffle', 'torch.nn.modules.upsampling', 'torch.nn.modules.distance', 'torch.nn.modules.fold', 'torch.nn.modules.adaptive', 'torch.nn.modules.transformer', 'torch.nn.modules.flatten', 'torch.nn.parallel', 'torch.nn.parallel.parallel_apply', 'torch.nn.parallel.replicate', 'torch.cuda.comm', 'torch.cuda.nccl', 'torch.nn.parallel.data_parallel', 'torch.nn.parallel.scatter_gather', 'torch.nn.parallel._functions', 'torch.nn.parallel.distributed', 'torch.distributed', 'torch.distributed.autograd', 'torch.distributed.distributed_c10d', 'torch.distributed.rendezvous', 'torch.distributed.rpc', 'torch.distributed.internal_rpc_utils', 'torch.distributed.rpc_backend_registry', 'torch.nn.intrinsic', 'torch.nn.intrinsic.modules', 'torch.nn.intrinsic.modules.fused', 'torch.nn.quantized', 'torch.nn.quantized.modules', 'torch.nn.quantized.modules.activation', 'torch.nn.quantized.functional', 'torch.nn.quantized.modules.conv', 'torch.nn.intrinsic.qat', 'torch.nn.intrinsic.qat.modules', 'torch.nn.intrinsic.qat.modules.linear_relu', 'torch.nn.qat', 'torch.nn.qat.modules', 'torch.nn.qat.modules.linear', 'torch.nn.qat.modules.conv', 'torch.nn.intrinsic.qat.modules.conv_fused', 'torch._ops', 'torch.jit', 'torch.backends', 'torch.backends.cudnn', 'torch.jit.annotations', 'torch.jit._recursive', 'torch.jit.frontend', 'torch.nn.quantized.modules.utils', 'torch.nn.quantized.modules.linear', 'torch.nn.quantized.modules.functional_modules', 'torch.optim', 'torch.optim.adadelta', 'torch.optim.optimizer', 'torch.optim.adagrad', 'torch.optim.adam', 'torch.optim.adamw', 'torch.optim.sparse_adam', 'torch.optim.adamax', 'torch.optim.asgd', 'torch.optim.sgd', 'torch.optim.rprop', 'torch.optim.rmsprop', 'torch.optim.lbfgs', 'torch.optim.lr_scheduler', 'torch.multiprocessing', 'torch.multiprocessing.reductions', 'multiprocessing.resource_sharer', 'torch.multiprocessing.spawn', 'multiprocessing.connection', '_multiprocessing', 'torch.utils.backcompat', 'torch.onnx', 'torch.hub', 'torch.distributions', 'torch.distributions.bernoulli', 'torch.distributions.constraints', 'torch.distributions.exp_family', 'torch.distributions.distribution', 'torch.distributions.utils', 'torch.distributions.beta', 'torch.distributions.dirichlet', 'torch.distributions.binomial', 'torch.distributions.categorical', 'torch.distributions.cauchy', 'torch.distributions.chi2', 'torch.distributions.gamma', 'torch.distributions.constraint_registry', 'torch.distributions.transforms', 'torch.distributions.exponential', 'torch.distributions.fishersnedecor', 'torch.distributions.geometric', 'torch.distributions.gumbel', 'torch.distributions.uniform', 'torch.distributions.transformed_distribution', 'torch.distributions.half_cauchy', 'torch.distributions.half_normal', 'torch.distributions.normal', 'torch.distributions.independent', 'torch.distributions.kl', 'torch.distributions.laplace', 'torch.distributions.lowrank_multivariate_normal', 'torch.distributions.multivariate_normal', 'torch.distributions.one_hot_categorical', 'torch.distributions.pareto', 'torch.distributions.poisson', 'torch.distributions.log_normal', 'torch.distributions.logistic_normal', 'torch.distributions.multinomial', 'torch.distributions.negative_binomial', 'torch.distributions.relaxed_bernoulli', 'torch.distributions.relaxed_categorical', 'torch.distributions.studentT', 'torch.distributions.weibull', 'torch.backends.cuda', 'torch.backends.mkl', 'torch.backends.openmp', 'torch.backends.quantized', 'torch.quantization', 'torch.quantization.quantize', 'torch.quantization.default_mappings', 'torch.nn.intrinsic.quantized', 'torch.nn.intrinsic.quantized.modules', 'torch.nn.intrinsic.quantized.modules.linear_relu', 'torch.nn.intrinsic.quantized.modules.conv_relu', 'torch.nn.quantized.dynamic', 'torch.nn.quantized.dynamic.modules', 'torch.nn.quantized.dynamic.modules.linear', 'torch.nn.quantized.dynamic.modules.rnn', 'torch.quantization.stubs', 'torch.quantization.QConfig', 'torch.quantization.observer', 'torch.quantization.fake_quantize', 'torch.quantization.fuse_modules', 'torch.utils.data', 'torch.utils.data.sampler', 'torch.utils.data.distributed', 'torch.utils.data.dataset', 'torch.utils.data.dataloader', 'torch.utils.data._utils', 'torch.utils.data._utils.worker', 'torch.utils.data._utils.signal_handling', 'torch.utils.data._utils.pin_memory', 'torch.utils.data._utils.collate', 'torch.utils.data._utils.fetch', 'torch.__config__', 'torch.__future__', 'torch._torch_docs', 'torch._tensor_docs', 'torch._storage_docs', 'torch._classes', 'torch.quasirandom', 'thinc.extra.wrappers', 'torch.utils.dlpack', 'spacy.gold', 'spacy.strings', 'spacy.morphology', 'spacy.vocab', 'spacy.lexeme', 'spacy.tokens', 'spacy.tokens.doc', 'spacy.tokens.span', 'spacy.tokens.token', 'spacy.tokens.morphanalysis', 'spacy.parts_of_speech', 'spacy.tokens.underscore', 'spacy.tokens._retokenize', 'spacy.tokens._serialize', 'spacy.lemmatizer', 'spacy.lookups', 'preshed.bloom', 'spacy.vectors', 'spacy.syntax', 'spacy.syntax.stateclass', 'spacy.syntax.transition_system', 'thinc.extra.search', 'spacy.syntax._beam_utils', 'spacy._align', 'spacy.syntax.nonproj', 'spacy.cli.pretrain', 'spacy.cli.debug_data', 'spacy.cli.evaluate', 'spacy.displacy', 'spacy.displacy.render', 'spacy.displacy.templates', 'spacy.cli.convert', 'spacy.cli.converters', 'spacy.cli.converters.conllu2json', 'spacy.cli.converters.iob2json', 'spacy.cli.converters.conll_ner2json', 'spacy.lang', 'spacy.lang.xx', 'spacy.lang.tokenizer_exceptions', 'spacy.lang.norm_exceptions', 'spacy.language', 'spacy.tokenizer', 'spacy.pipeline', 'spacy.pipeline.pipes', 'spacy.syntax.nn_parser', 'spacy.syntax.arc_eager', 'spacy.syntax._parser_model', 'spacy.syntax.ner', 'spacy.pipeline.functions', 'spacy.matcher', 'spacy.matcher.matcher', 'spacy.matcher._schemas', 'spacy.matcher.phrasematcher', 'spacy.matcher.dependencymatcher', 'spacy.kb', 'spacy.pipeline.morphologizer', 'spacy.pipeline.entityruler', 'spacy.pipeline.hooks', 'thinc.neural._classes.difference', 'spacy.scorer', 'spacy.lang.punctuation', 'spacy.lang.char_classes', 'spacy.lang.tag_map', 'spacy.lang.lex_attrs', 'spacy.cli.converters.jsonl2json', 'spacy.cli.init_model', 'preshed.counter', 'ftfy', 'ftfy.bad_codecs', 'ftfy.fixes', 'ftfy.badness', 'ftfy.chardata', 'ftfy.bad_codecs.sloppy', 'encodings.cp1250', 'encodings.cp1251', 'encodings.cp1252', 'encodings.cp1253', 'encodings.cp1254', 'encodings.cp1255', 'encodings.cp1256', 'encodings.cp1257', 'encodings.cp1258', 'encodings.iso8859_3', 'encodings.iso8859_6', 'encodings.iso8859_7', 'encodings.iso8859_8', 'encodings.iso8859_11', 'encodings.cp874', 'encodings.iso8859_2', 'encodings.mac_roman', 'encodings.cp437', 'ftfy.formatting', 'spacy.cli.validate', 'spacy.glossary', 'en_core_web_sm', 'spacy.lang.en', 'spacy.lang.en.tokenizer_exceptions', 'spacy.lang.en.norm_exceptions', 'spacy.lang.en.tag_map', 'spacy.lang.en.stop_words', 'spacy.lang.en.lex_attrs', 'spacy.lang.en.morph_rules', 'spacy.lang.en.syntax_iterators', 'spacy_transformers', 'spacy_transformers.language', 'spacy_transformers.util', 'transformers', 'transformers.file_utils', 'boto3', 'boto3.session', 'botocore', 'botocore.session', 'botocore.configloader', 'botocore.compat', 'botocore.vendored', 'botocore.vendored.six', 'botocore.exceptions', 'botocore.vendored.requests', 'botocore.vendored.requests.packages', 'botocore.vendored.requests.packages.urllib3', 'botocore.vendored.requests.packages.urllib3.connectionpool', 'botocore.vendored.requests.packages.urllib3.exceptions', 'botocore.vendored.requests.packages.urllib3.packages', 'botocore.vendored.requests.packages.urllib3.packages.ssl_match_hostname', 'botocore.vendored.requests.packages.urllib3.packages.six', 'botocore.vendored.requests.packages.urllib3.packages.six.moves', 'botocore.vendored.requests.packages.urllib3.connection', 'botocore.vendored.requests.packages.urllib3.util', 'botocore.vendored.requests.packages.urllib3.util.connection', 'botocore.vendored.requests.packages.urllib3.util.request', 'botocore.vendored.requests.packages.urllib3.util.response', 'botocore.vendored.requests.packages.urllib3.util.ssl_', 'botocore.vendored.requests.packages.urllib3.util.timeout', 'botocore.vendored.requests.packages.urllib3.util.retry', 'botocore.vendored.requests.packages.urllib3.util.url', 'botocore.vendored.requests.packages.urllib3.request', 'botocore.vendored.requests.packages.urllib3.filepost', 'botocore.vendored.requests.packages.urllib3.fields', 'botocore.vendored.requests.packages.urllib3.response', 'botocore.vendored.requests.packages.urllib3._collections', 'botocore.vendored.requests.packages.urllib3.poolmanager', 'botocore.vendored.requests.packages.urllib3.contrib', 'botocore.vendored.requests.utils', 'cgi', 'botocore.vendored.requests.certs', 'botocore.vendored.requests.compat', 'botocore.vendored.requests.packages.chardet', 'botocore.vendored.requests.cookies', 'botocore.vendored.requests.structures', 'botocore.vendored.requests.exceptions', 'botocore.vendored.requests.models', 'botocore.vendored.requests.hooks', 'botocore.vendored.requests.auth', 'botocore.vendored.requests.status_codes', 'botocore.vendored.requests.api', 'botocore.vendored.requests.sessions', 'botocore.vendored.requests.adapters', 'botocore.vendored.six.moves', 'xml.etree', 'xml.etree.cElementTree', 'xml.etree.ElementTree', 'xml.etree.ElementPath', '_elementtree', 'botocore.credentials', 'botocore.config', 'botocore.endpoint', 'botocore.awsrequest', 'botocore.utils', 'botocore.httpsession', 'botocore.vendored.six.moves.urllib_parse', 'botocore.vendored.six.moves.urllib', 'botocore.vendored.six.moves.urllib.request', 'botocore.hooks', 'botocore.history', 'botocore.response', 'botocore.parsers', 'botocore.eventstream', 'botocore.client', 'botocore.waiter', 'jmespath', 'jmespath.parser', 'jmespath.lexer', 'jmespath.exceptions', 'jmespath.compat', 'jmespath.ast', 'jmespath.visitor', 'jmespath.functions', 'botocore.docs', 'botocore.docs.service', 'botocore.docs.utils', 'botocore.docs.client', 'botocore.docs.method', 'botocore.docs.params', 'botocore.docs.shape', 'botocore.docs.example', 'botocore.docs.sharedexample', 'botocore.docs.waiter', 'botocore.docs.paginator', 'botocore.docs.bcdoc', 'botocore.docs.bcdoc.restdoc', 'botocore.docs.bcdoc.docstringparser', 'html.parser', '_markupbase', 'botocore.docs.bcdoc.style', 'botocore.docs.docstring', 'botocore.auth', 'botocore.model', 'botocore.paginate', 'botocore.args', 'botocore.serialize', 'botocore.validate', 'botocore.signers', 'botocore.discovery', 'botocore.configprovider', 'botocore.errorfactory', 'botocore.handlers', 'botocore.retryhandler', 'botocore.translate', 'botocore.loaders', 'botocore.regions', 'botocore.monitoring', 'boto3.utils', 'boto3.exceptions', 'boto3.resources', 'boto3.resources.factory', 'boto3.resources.action', 'boto3.resources.params', 'boto3.resources.response', 'boto3.resources.model', 'boto3.docs', 'boto3.docs.service', 'boto3.docs.client', 'boto3.docs.resource', 'boto3.docs.base', 'boto3.docs.action', 'boto3.docs.method', 'boto3.docs.utils', 'boto3.docs.waiter', 'boto3.docs.collection', 'boto3.docs.subresource', 'boto3.docs.attr', 'boto3.docs.docstring', 'boto3.resources.base', 'boto3.resources.collection', 'transformers.data', 'transformers.data.processors', 'transformers.data.processors.utils', 'transformers.data.processors.glue', 'transformers.data.metrics', 'scipy', 'scipy._distributor_init', 'scipy.__config__', 'scipy.version', 'scipy._lib', 'scipy._lib._testutils', 'scipy._lib._version', 'scipy._lib.six', 'scipy._lib._ccallback', 'scipy._lib._ccallback_c', 'scipy.stats', 'scipy.stats.stats', 'scipy._lib._util', 'scipy.special', 'scipy.special.sf_error', 'scipy.special._ufuncs', '_cython_0_29_12', 'scipy.special._ufuncs_cxx', 'scipy.special.basic', 'scipy.special.specfun', 'scipy.special.orthogonal', 'scipy.linalg', 'scipy.linalg.linalg_version', 'scipy.linalg.misc', 'scipy.linalg.blas', 'scipy.linalg._fblas', 'scipy.linalg.lapack', 'scipy.linalg._flapack', 'scipy.linalg.basic', 'scipy.linalg.flinalg', 'scipy.linalg._flinalg', 'scipy.linalg.decomp', 'scipy.linalg.decomp_svd', 'scipy.linalg._solve_toeplitz', 'scipy.linalg.decomp_lu', 'scipy.linalg._decomp_ldl', 'scipy.linalg.decomp_cholesky', 'scipy.linalg.decomp_qr', 'scipy.linalg._decomp_qz', 'scipy.linalg.decomp_schur', 'scipy.linalg._decomp_polar', 'scipy.linalg.matfuncs', 'scipy.linalg.special_matrices', 'scipy.linalg._expm_frechet', 'scipy.linalg._matfuncs_sqrtm', 'scipy.linalg._solvers', 'scipy.linalg._procrustes', 'scipy.linalg._decomp_update', 'scipy.linalg.cython_blas', 'scipy.linalg.cython_lapack', 'scipy.linalg._sketches', 'scipy.sparse', 'scipy.sparse.base', 'scipy._lib._numpy_compat', 'scipy.sparse.sputils', 'scipy.sparse.csr', 'scipy.sparse._sparsetools', 'scipy.sparse.compressed', 'scipy.sparse.data', 'scipy.sparse.dia', 'scipy.sparse._index', 'scipy.sparse.csc', 'scipy.sparse.lil', 'scipy.sparse._csparsetools', 'scipy.sparse.dok', 'scipy.sparse.coo', 'scipy.sparse.bsr', 'scipy.sparse.construct', 'scipy.sparse.extract', 'scipy.sparse._matrix_io', 'scipy.sparse.csgraph', 'scipy.sparse.csgraph._laplacian', 'scipy.sparse.csgraph._shortest_path', 'scipy.sparse.csgraph._validation', 'scipy.sparse.csgraph._tools', 'scipy.sparse.csgraph._traversal', 'scipy.sparse.csgraph._min_spanning_tree', 'scipy.sparse.csgraph._reordering', 'numpy.dual', 'scipy.special._comb', 'scipy.special._logsumexp', 'scipy.special.spfun_stats', 'scipy.special._ellip_harm', 'scipy.special._ellip_harm_2', 'scipy.special.lambertw', 'scipy.special._spherical_bessel', 'scipy.stats.distributions', 'scipy.stats._distn_infrastructure', 'scipy._lib.doccer', 'scipy.stats._distr_params', 'scipy.optimize', 'scipy.optimize.optimize', 'scipy.optimize.linesearch', 'scipy.optimize.minpack2', 'scipy.optimize._minimize', 'scipy.optimize._trustregion_dogleg', 'scipy.optimize._trustregion', 'scipy.optimize._trustregion_ncg', 'scipy.optimize._trustregion_krylov', 'scipy.optimize._trlib', 'scipy.optimize._trlib._trlib', 'scipy._lib.messagestream', 'scipy.optimize._trustregion_exact', 'scipy.optimize._trustregion_constr', 'scipy.optimize._trustregion_constr.minimize_trustregion_constr', 'scipy.sparse.linalg', 'scipy.sparse.linalg.isolve', 'scipy.sparse.linalg.isolve.iterative', 'scipy.sparse.linalg.isolve._iterative', 'scipy.sparse.linalg.interface', 'scipy.sparse.linalg.isolve.utils', 'scipy._lib._threadsafety', 'scipy._lib.decorator', 'scipy.sparse.linalg.isolve.minres', 'scipy.sparse.linalg.isolve.lgmres', 'scipy.sparse.linalg.isolve._gcrotmk', 'scipy.sparse.linalg.isolve.lsqr', 'scipy.sparse.linalg.isolve.lsmr', 'scipy.sparse.linalg.dsolve', 'scipy.sparse.linalg.dsolve.linsolve', 'scipy.sparse.linalg.dsolve._superlu', 'scipy.sparse.linalg.dsolve._add_newdocs', 'scipy.sparse.linalg.eigen', 'scipy.sparse.linalg.eigen.arpack', 'scipy.sparse.linalg.eigen.arpack.arpack', 'scipy.sparse.linalg.eigen.arpack._arpack', 'scipy.sparse.linalg.eigen.lobpcg', 'scipy.sparse.linalg.eigen.lobpcg.lobpcg', 'scipy.sparse.linalg.matfuncs', 'scipy.sparse.linalg._onenormest', 'scipy.sparse.linalg._norm', 'scipy.sparse.linalg._expm_multiply', 'scipy.optimize._differentiable_functions', 'scipy.optimize._numdiff', 'scipy.optimize._group_columns', 'scipy.optimize._hessian_update_strategy', 'scipy.optimize._constraints', 'scipy.optimize._trustregion_constr.equality_constrained_sqp', 'scipy.optimize._trustregion_constr.projections', 'scipy.optimize._trustregion_constr.qp_subproblem', 'scipy.optimize._trustregion_constr.canonical_constraint', 'scipy.optimize._trustregion_constr.tr_interior_point', 'scipy.optimize._trustregion_constr.report', 'scipy.optimize.lbfgsb', 'scipy.optimize._lbfgsb', 'scipy.optimize.tnc', 'scipy.optimize.moduleTNC', 'scipy.optimize.cobyla', 'scipy.optimize._cobyla', 'scipy.optimize.slsqp', 'scipy.optimize._slsqp', 'scipy.optimize._root', 'scipy.optimize.minpack', 'scipy.optimize._minpack', 'scipy.optimize._lsq', 'scipy.optimize._lsq.least_squares', 'scipy.optimize._lsq.trf', 'scipy.optimize._lsq.common', 'scipy.optimize._lsq.dogbox', 'scipy.optimize._lsq.lsq_linear', 'scipy.optimize._lsq.trf_linear', 'scipy.optimize._lsq.givens_elimination', 'scipy.optimize._lsq.bvls', 'scipy.optimize._spectral', 'scipy.optimize.nonlin', 'scipy.optimize._root_scalar', 'scipy.optimize.zeros', 'scipy.optimize._zeros', 'scipy.optimize.nnls', 'scipy.optimize._nnls', 'scipy.optimize._basinhopping', 'scipy.optimize._linprog', 'scipy.optimize._linprog_ip', 'scipy.optimize._linprog_util', 'scipy.optimize._remove_redundancy', 'scipy.optimize._linprog_simplex', 'scipy.optimize._linprog_rs', 'scipy.optimize._bglu_dense', 'scipy.optimize._hungarian', 'scipy.optimize._differentialevolution', 'scipy.optimize._shgo', 'scipy.spatial', 'scipy.spatial.kdtree', 'scipy.spatial.ckdtree', 'scipy.spatial.qhull', 'scipy.spatial._spherical_voronoi', 'scipy.spatial._voronoi', 'scipy.spatial.distance', 'scipy.spatial._distance_wrap', 'scipy.spatial._hausdorff', 'scipy.spatial._plotutils', 'scipy.spatial._procrustes', 'scipy.spatial.transform', 'scipy.spatial.transform.rotation', 'scipy.spatial.transform._rotation_spline', 'scipy.optimize._shgo_lib', 'scipy.optimize._shgo_lib.sobol_seq', 'scipy.optimize._shgo_lib.triangulation', 'scipy.optimize._dual_annealing', 'scipy.integrate', 'scipy.integrate.quadrature', 'scipy.integrate.odepack', 'scipy.integrate._odepack', 'scipy.integrate.quadpack', 'scipy.integrate._quadpack', 'scipy.integrate._ode', 'scipy.integrate.vode', 'scipy.integrate._dop', 'scipy.integrate.lsoda', 'scipy.integrate._bvp', 'scipy.integrate._ivp', 'scipy.integrate._ivp.ivp', 'scipy.integrate._ivp.bdf', 'scipy.integrate._ivp.common', 'scipy.integrate._ivp.base', 'scipy.integrate._ivp.radau', 'scipy.integrate._ivp.rk', 'scipy.integrate._ivp.lsoda', 'scipy.misc', 'scipy.misc.doccer', 'scipy.misc.common', 'scipy.stats._constants', 'scipy.stats._continuous_distns', 'scipy.interpolate', 'scipy.interpolate.interpolate', 'scipy.interpolate.fitpack', 'scipy.interpolate._fitpack_impl', 'scipy.interpolate._fitpack', 'scipy.interpolate.dfitpack', 'scipy.interpolate._bsplines', 'scipy.interpolate._bspl', 'scipy.interpolate.polyint', 'scipy.interpolate._ppoly', 'scipy.interpolate.fitpack2', 'scipy.interpolate.interpnd', 'scipy.interpolate.rbf', 'scipy.interpolate._cubic', 'scipy.interpolate.ndgriddata', 'scipy.interpolate._pade', 'scipy.stats._stats', 'scipy.stats._tukeylambda_stats', 'scipy.stats._discrete_distns', 'scipy.stats.mstats_basic', 'scipy.stats._stats_mstats_common', 'scipy.stats._rvs_sampling', 'scipy.stats._hypotests', 'scipy.stats.morestats', 'scipy.stats.statlib', 'scipy.stats.contingency', 'scipy.stats._binned_statistic', 'scipy.stats.kde', 'scipy.stats.mvn', 'scipy.stats.mstats', 'scipy.stats.mstats_extras', 'scipy.stats._multivariate', 'sklearn', 'sklearn._config', 'sklearn.__check_build', 'sklearn.__check_build._check_build', 'sklearn.base', 'sklearn.utils', 'sklearn.utils.murmurhash', 'sklearn.utils.class_weight', 'sklearn.utils._joblib', 'joblib', 'joblib.memory', 'joblib.hashing', 'joblib._compat', 'joblib.func_inspect', 'joblib.logger', 'joblib.disk', 'joblib._memory_helpers', 'joblib._store_backends', 'joblib.backports', 'distutils', 'distutils.version', 'joblib.numpy_pickle', 'joblib.compressor', 'joblib.numpy_pickle_utils', 'joblib.numpy_pickle_compat', 'joblib.parallel', 'joblib._multiprocessing_helpers', 'multiprocessing.synchronize', 'multiprocessing.semaphore_tracker', 'multiprocessing.spawn', 'joblib.format_stack', 'joblib.my_exceptions', 'joblib._parallel_backends', 'joblib.pool', 'joblib._memmapping_reducer', 'mmap', 'multiprocessing.pool', 'joblib.executor', 'joblib.externals', 'joblib.externals.loky', 'joblib.externals.loky._base', 'joblib.externals.loky.backend', 'joblib.externals.loky.backend.context', 'joblib.externals.loky.backend.process', 'joblib.externals.loky.backend.compat', 'joblib.externals.loky.backend.compat_posix', 'joblib.externals.loky.backend.reduction', 'joblib.externals.loky.backend._posix_reduction', 'joblib.externals.cloudpickle', 'joblib.externals.cloudpickle.cloudpickle', 'joblib.externals.loky.reusable_executor', 'joblib.externals.loky.process_executor', 'joblib.externals.loky.backend.queues', 'multiprocessing.queues', 'joblib.externals.loky.backend.utils', 'psutil', 'psutil._common', 'psutil._compat', 'psutil._pslinux', 'psutil._psposix', 'psutil._psutil_linux', 'psutil._psutil_posix', 'concurrent.futures.process', 'joblib.externals.loky.cloudpickle_wrapper', 'sklearn.exceptions', 'sklearn.utils.deprecation', 'sklearn.utils.validation', 'sklearn.utils.fixes', 'sklearn.utils._show_versions', 'sklearn.metrics', 'sklearn.metrics.ranking', 'sklearn.utils.multiclass', 'sklearn.utils.extmath', 'sklearn.utils._logistic_sigmoid', 'sklearn.utils.sparsefuncs_fast', '_cython_0_29_8', 'sklearn.utils.sparsefuncs', 'sklearn.preprocessing', 'sklearn.preprocessing._function_transformer', 'sklearn.utils.testing', 'scipy.io', 'scipy.io.matlab', 'scipy.io.matlab.mio', 'scipy.io.matlab.miobase', 'scipy.io.matlab.byteordercodes', 'scipy.io.matlab.mio4', 'scipy.io.matlab.mio_utils', 'scipy.io.matlab.mio5', 'scipy.io.matlab.mio5_utils', 'scipy.io.matlab.streams', 'scipy.io.matlab.mio5_params', 'scipy.io.netcdf', 'scipy.io._fortran', 'scipy.io.mmio', 'scipy.io.idl', 'scipy.io.harwell_boeing', 'scipy.io.harwell_boeing.hb', 'scipy.io.harwell_boeing._fortran_format_parser', 'sklearn.utils._unittest_backport', 'nose', 'nose.core', 'nose.config', 'optparse', 'configparser', 'nose.util', 'nose.pyversion', 'nose.plugins', 'nose.plugins.base', 'nose.plugins.manager', 'nose.failure', 'nose.plugins.plugintest', 'nose.loader', 'nose.case', 'nose.importer', 'nose.selector', 'nose.suite', 'nose.proxy', 'nose.result', 'nose.exc', 'nose.plugins.skip', 'nose.plugins.errorclass', 'nose.plugins.deprecated', 'nose.tools', 'nose.tools.nontrivial', 'nose.tools.trivial', 'pytest', '_pytest', '_pytest._version', '_pytest.assertion', '_pytest.assertion.rewrite', 'atomicwrites', '_pytest._io', '_pytest._io.saferepr', '_pytest.assertion.util', '_pytest._code', '_pytest._code.code', 'pluggy', 'pluggy._version', 'pluggy.manager', 'pluggy._tracing', 'pluggy.callers', 'pluggy.hooks', 'importlib_metadata', 'zipp', 'importlib_metadata._compat', 'py', 'py._error', 'py.error', 'py._vendored_packages', 'py._vendored_packages.apipkg', 'py._version', 'py.test', 'py.process', 'py.apipkg', 'py.iniconfig', 'py.path', 'py.code', 'py.builtin', 'py.io', 'py.xml', 'py.log', 'py._path', 'py._path.local', 'py._path.common', '_pytest._code.source', '_pytest.outcomes', 'packaging', 'packaging.__about__', 'packaging.version', 'packaging._structures', '_pytest.pathlib', '_pytest.assertion.truncate', '_pytest.config', '_pytest.hookspec', '_pytest.deprecated', '_pytest.warning_types', '_pytest.config.exceptions', '_pytest.config.findpaths', '_pytest.debugging', 'doctest', '_pytest.fixtures', '_pytest.nodes', '_pytest.compat', '_pytest.mark', '_pytest.mark.legacy', '_pytest.mark.structures', '_pytest.freeze_support', '_pytest.main', '_pytest.runner', '_pytest.reports', '_pytest.python', '_pytest.python_api', 'more_itertools', 'more_itertools.more', 'more_itertools.recipes', '_pytest.recwarn', 'sklearn.preprocessing.data', 'sklearn.preprocessing._csr_polynomial_expansion', 'sklearn.preprocessing._encoders', 'sklearn.preprocessing.base', 'sklearn.preprocessing.label', 'sklearn.preprocessing._discretization', 'sklearn.preprocessing.imputation', 'sklearn.metrics.base', 'sklearn.metrics.classification', 'sklearn.metrics.cluster', 'sklearn.metrics.cluster.supervised', 'sklearn.metrics.cluster.expected_mutual_info_fast', 'sklearn.utils.lgamma', 'sklearn.metrics.cluster.unsupervised', 'sklearn.metrics.pairwise', 'sklearn.metrics.pairwise_fast', 'sklearn.utils._cython_blas', 'sklearn.metrics.cluster.bicluster', 'sklearn.metrics.regression', 'sklearn.metrics.scorer', 'transformers.tokenization_utils', 'transformers.tokenization_auto', 'transformers.tokenization_bert', 'transformers.tokenization_openai', 'transformers.tokenization_gpt2', 'regex', 'regex.regex', 'regex._regex_core', 'regex._regex', 'transformers.tokenization_transfo_xl', 'transformers.tokenization_xlnet', 'transformers.tokenization_xlm', 'sacremoses', 'sacremoses.corpus', 'sacremoses.tokenize', 'sacremoses.util', 'xml.sax', 'xml.sax.xmlreader', 'xml.sax.handler', 'xml.sax._exceptions', 'xml.sax.saxutils', 'sacremoses.truecase', 'sacremoses.normalize', 'transformers.tokenization_roberta', 'transformers.tokenization_distilbert', 'transformers.configuration_utils', 'transformers.configuration_auto', 'transformers.configuration_bert', 'transformers.configuration_openai', 'transformers.configuration_gpt2', 'transformers.configuration_transfo_xl', 'transformers.configuration_xlnet', 'transformers.configuration_xlm', 'transformers.configuration_roberta', 'transformers.configuration_distilbert', 'transformers.modeling_utils', 'transformers.modeling_auto', 'transformers.modeling_bert', 'transformers.modeling_openai', 'transformers.modeling_gpt2', 'transformers.modeling_transfo_xl', 'transformers.modeling_transfo_xl_utilities', 'transformers.modeling_xlnet', 'transformers.modeling_xlm', 'transformers.modeling_roberta', 'transformers.modeling_distilbert', 'transformers.optimization', 'spacy_transformers._tokenizers', 'sentencepiece', 'swig_runtime_data4', '_sentencepiece', 'spacy_transformers.activations', 'dataclasses', 'spacy_transformers.pipeline', 'spacy_transformers.pipeline.textcat', 'spacy_transformers.model_registry', 'spacy_transformers.wrapper', 'spacy_transformers.pipeline.tok2vec', 'spacy_transformers.pipeline.wordpiecer', 'networkx', 'networkx.release', 'networkx.exception', 'networkx.utils', 'networkx.utils.misc', 'networkx.utils.decorators', 'networkx.utils.random_sequence', 'networkx.utils.union_find', 'networkx.utils.rcm', 'networkx.utils.heaps', 'networkx.utils.contextmanagers', 'networkx.classes', 'networkx.classes.graph', 'networkx.classes.coreviews', 'networkx.classes.reportviews', 'networkx.convert', 'networkx.classes.digraph', 'networkx.classes.multigraph', 'networkx.classes.multidigraph', 'networkx.classes.ordered', 'networkx.classes.function', 'networkx.classes.filters', 'networkx.classes.graphviews', 'networkx.convert_matrix', 'networkx.relabel', 'networkx.generators', 'networkx.generators.atlas', 'networkx.generators.classic', 'networkx.generators.community', 'networkx.generators.degree_seq', 'networkx.generators.directed', 'networkx.generators.duplication', 'networkx.generators.ego', 'networkx.generators.expanders', 'networkx.generators.geometric', 'networkx.generators.intersection', 'networkx.algorithms', 'networkx.algorithms.assortativity', 'networkx.algorithms.assortativity.connectivity', 'networkx.algorithms.assortativity.correlation', 'networkx.algorithms.assortativity.mixing', 'networkx.algorithms.assortativity.pairs', 'networkx.algorithms.assortativity.neighbor_degree', 'networkx.algorithms.boundary', 'networkx.algorithms.bridges', 'networkx.algorithms.chains', 'networkx.algorithms.centrality', 'networkx.algorithms.centrality.betweenness', 'networkx.algorithms.centrality.betweenness_subset', 'networkx.algorithms.centrality.closeness', 'networkx.algorithms.centrality.subgraph_alg', 'networkx.algorithms.centrality.current_flow_closeness', 'networkx.algorithms.centrality.flow_matrix', 'networkx.algorithms.centrality.current_flow_betweenness', 'networkx.algorithms.centrality.current_flow_betweenness_subset', 'networkx.algorithms.centrality.degree_alg', 'networkx.algorithms.centrality.dispersion', 'networkx.algorithms.centrality.eigenvector', 'networkx.algorithms.centrality.harmonic', 'networkx.algorithms.centrality.katz', 'networkx.algorithms.centrality.load', 'networkx.algorithms.centrality.reaching', 'networkx.algorithms.centrality.percolation', 'networkx.algorithms.centrality.second_order', 'networkx.algorithms.centrality.voterank_alg', 'networkx.algorithms.chordal', 'networkx.algorithms.cluster', 'networkx.algorithms.clique', 'networkx.algorithms.communicability_alg', 'networkx.algorithms.components', 'networkx.algorithms.components.connected', 'networkx.algorithms.components.strongly_connected', 'networkx.algorithms.components.weakly_connected', 'networkx.algorithms.components.attracting', 'networkx.algorithms.components.biconnected', 'networkx.algorithms.components.semiconnected', 'networkx.algorithms.coloring', 'networkx.algorithms.coloring.greedy_coloring', 'networkx.algorithms.coloring.greedy_coloring_with_interchange', 'networkx.algorithms.coloring.equitable_coloring', 'networkx.algorithms.core', 'networkx.algorithms.covering', 'networkx.algorithms.cycles', 'networkx.algorithms.cuts', 'networkx.algorithms.dag', 'fractions', 'networkx.generators.trees', 'networkx.algorithms.distance_measures', 'networkx.algorithms.distance_regular', 'networkx.algorithms.dominance', 'networkx.algorithms.dominating', 'networkx.algorithms.efficiency', 'networkx.algorithms.euler', 'networkx.algorithms.graphical', 'networkx.algorithms.hierarchy', 'networkx.algorithms.hybrid', 'networkx.algorithms.link_analysis', 'networkx.algorithms.link_analysis.pagerank_alg', 'networkx.algorithms.link_analysis.hits_alg', 'networkx.algorithms.link_prediction', 'networkx.algorithms.lowest_common_ancestors', 'networkx.algorithms.isolate', 'networkx.algorithms.matching', 'networkx.algorithms.minors', 'networkx.algorithms.mis', 'networkx.algorithms.operators', 'networkx.algorithms.operators.all', 'networkx.algorithms.operators.binary', 'networkx.algorithms.operators.product', 'networkx.algorithms.operators.unary', 'networkx.algorithms.planarity', 'networkx.algorithms.planar_drawing', 'networkx.algorithms.reciprocity', 'networkx.algorithms.richclub', 'networkx.algorithms.shortest_paths', 'networkx.algorithms.shortest_paths.generic', 'networkx.algorithms.shortest_paths.unweighted', 'networkx.algorithms.shortest_paths.weighted', 'networkx.algorithms.shortest_paths.astar', 'networkx.algorithms.shortest_paths.dense', 'networkx.algorithms.similarity', 'networkx.algorithms.simple_paths', 'networkx.algorithms.smallworld', 'networkx.algorithms.smetric', 'networkx.algorithms.structuralholes', 'networkx.algorithms.sparsifiers', 'networkx.algorithms.swap', 'networkx.algorithms.traversal', 'networkx.algorithms.traversal.beamsearch', 'networkx.algorithms.traversal.breadth_first_search', 'networkx.algorithms.traversal.depth_first_search', 'networkx.algorithms.traversal.edgedfs', 'networkx.algorithms.traversal.edgebfs', 'networkx.algorithms.triads', 'networkx.algorithms.vitality', 'networkx.algorithms.voronoi', 'networkx.algorithms.wiener', 'networkx.algorithms.bipartite', 'networkx.algorithms.bipartite.basic', 'networkx.algorithms.bipartite.centrality', 'networkx.algorithms.bipartite.cluster', 'networkx.algorithms.bipartite.covering', 'networkx.algorithms.bipartite.matching', 'networkx.algorithms.bipartite.edgelist', 'networkx.algorithms.bipartite.matrix', 'networkx.algorithms.bipartite.projection', 'networkx.algorithms.bipartite.redundancy', 'networkx.algorithms.bipartite.spectral', 'networkx.algorithms.bipartite.generators', 'networkx.algorithms.node_classification', 'networkx.algorithms.node_classification.hmn', 'networkx.algorithms.node_classification.utils', 'networkx.algorithms.node_classification.lgc', 'networkx.algorithms.connectivity', 'networkx.algorithms.connectivity.connectivity', 'networkx.algorithms.flow', 'networkx.algorithms.flow.maxflow', 'networkx.algorithms.flow.boykovkolmogorov', 'networkx.algorithms.flow.utils', 'networkx.algorithms.flow.dinitz_alg', 'networkx.algorithms.flow.edmondskarp', 'networkx.algorithms.flow.preflowpush', 'networkx.algorithms.flow.shortestaugmentingpath', 'networkx.algorithms.flow.mincost', 'networkx.algorithms.flow.gomory_hu', 'networkx.algorithms.flow.capacityscaling', 'networkx.algorithms.flow.networksimplex', 'networkx.algorithms.connectivity.utils', 'networkx.algorithms.connectivity.cuts', 'networkx.algorithms.connectivity.edge_augmentation', 'networkx.algorithms.connectivity.edge_kcomponents', 'networkx.algorithms.connectivity.disjoint_paths', 'networkx.algorithms.connectivity.kcomponents', 'networkx.algorithms.connectivity.kcutsets', 'networkx.algorithms.connectivity.stoerwagner', 'networkx.algorithms.community', 'networkx.algorithms.community.asyn_fluid', 'networkx.algorithms.community.centrality', 'networkx.algorithms.community.community_generators', 'networkx.algorithms.community.kclique', 'networkx.algorithms.community.kernighan_lin', 'networkx.algorithms.community.community_utils', 'networkx.algorithms.community.label_propagation', 'networkx.algorithms.community.modularity_max', 'networkx.algorithms.community.quality', 'networkx.utils.mapped_queue', 'networkx.algorithms.isomorphism', 'networkx.algorithms.isomorphism.isomorph', 'networkx.algorithms.isomorphism.vf2userfunc', 'networkx.algorithms.isomorphism.isomorphvf2', 'networkx.algorithms.isomorphism.matchhelpers', 'networkx.algorithms.isomorphism.temporalisomorphvf2', 'networkx.algorithms.tournament', 'networkx.algorithms.tree', 'networkx.algorithms.tree.branchings', 'networkx.algorithms.tree.recognition', 'networkx.algorithms.tree.coding', 'networkx.algorithms.tree.mst', 'networkx.algorithms.tree.operations', 'networkx.generators.joint_degree_seq', 'networkx.generators.lattice', 'networkx.generators.line', 'networkx.generators.mycielski', 'networkx.generators.nonisomorphic_trees', 'networkx.generators.random_clustered', 'networkx.generators.random_graphs', 'networkx.generators.small', 'networkx.generators.social', 'networkx.generators.spectral_graph_forge', 'networkx.generators.stochastic', 'networkx.generators.triads', 'networkx.readwrite', 'networkx.readwrite.adjlist', 'networkx.readwrite.multiline_adjlist', 'networkx.readwrite.edgelist', 'networkx.readwrite.gpickle', 'networkx.readwrite.pajek', 'networkx.readwrite.leda', 'networkx.readwrite.sparse6', 'networkx.readwrite.graph6', 'networkx.readwrite.nx_yaml', 'networkx.readwrite.gml', 'networkx.readwrite.graphml', 'lxml', 'lxml.etree', '_cython_0_29_10', 'lxml._elementpath', 'networkx.readwrite.gexf', 'networkx.readwrite.nx_shp', 'networkx.readwrite.json_graph', 'networkx.readwrite.json_graph.node_link', 'networkx.readwrite.json_graph.adjacency', 'networkx.readwrite.json_graph.tree', 'networkx.readwrite.json_graph.jit', 'networkx.readwrite.json_graph.cytoscape', 'networkx.linalg', 'networkx.linalg.attrmatrix', 'networkx.linalg.spectrum', 'networkx.linalg.graphmatrix', 'networkx.linalg.laplacianmatrix', 'networkx.linalg.algebraicconnectivity', 'networkx.linalg.modularitymatrix', 'networkx.tests', 'networkx.tests.test', 'networkx.drawing', 'networkx.drawing.layout', 'networkx.drawing.nx_pylab', 'networkx.drawing.nx_agraph', 'networkx.drawing.nx_pydot', 'matplotlib', 'matplotlib.cbook', 'matplotlib.cbook.deprecation', 'matplotlib.rcsetup', 'matplotlib.fontconfig_pattern', 'pyparsing', 'matplotlib.colors', 'matplotlib._color_data', 'cycler', 'matplotlib._version', 'kiwisolver']\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:matplotlib:CACHEDIR=/home/ceteri/.cache/matplotlib\n",
      "DEBUG:matplotlib.font_manager:Using fontManager instance from /home/ceteri/.cache/matplotlib/fontlist-v310.json\n",
      "DEBUG:matplotlib.pyplot:Loaded backend module://ipykernel.pylab.backend_inline version unknown.\n",
      "DEBUG:matplotlib.pyplot:Loaded backend module://ipykernel.pylab.backend_inline version unknown.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{0: Text(-0.259197838584597, -0.15872403687931053, 'compatibility'),\n",
       " 1: Text(-0.11695784118870008, -0.3713979484271882, 'system'),\n",
       " 2: Text(-0.08852892875670923, -0.5079477285197758, 'linear'),\n",
       " 3: Text(0.09213049042976762, -0.19987158370488659, 'constraint'),\n",
       " 4: Text(0.25330122691920326, 0.19153379113591698, 'set'),\n",
       " 5: Text(0.5185972918823862, 0.12252144735187699, 'natural'),\n",
       " 6: Text(0.5244082115510904, 0.22784013636184175, 'number'),\n",
       " 7: Text(-0.36103422253728407, 0.1553942016003211, 'criterion'),\n",
       " 8: Text(-0.19336944344824034, -0.6505271449585693, 'diophantine'),\n",
       " 9: Text(-0.2660752416095871, -0.7391842672963954, 'equation'),\n",
       " 10: Text(-0.3654875399435523, -0.8243459637400486, 'strict'),\n",
       " 11: Text(-0.4597495497267383, -0.6805045883021201, 'inequation'),\n",
       " 12: Text(-0.5658576299933219, -0.6025220285231017, 'nonstrict'),\n",
       " 13: Text(-0.3599919297901866, -0.4244719440327338, 'consider'),\n",
       " 14: Text(0.5781434212564582, 1.0, 'upper'),\n",
       " 15: Text(0.6690017134927084, 0.9366815527346906, 'bound'),\n",
       " 16: Text(0.48609569153618537, 0.7651270668995663, 'component'),\n",
       " 17: Text(0.2497684884161864, 0.4200526735559854, 'minimal'),\n",
       " 18: Text(0.07213378209505379, 0.167964348930075, 'solution'),\n",
       " 19: Text(-0.05120399908985773, 0.41869565492829397, 'algorithm'),\n",
       " 20: Text(0.1495382301490408, 0.5338464624417432, 'construction'),\n",
       " 21: Text(0.23114521046953881, 0.33445069688834145, 'generating'),\n",
       " 22: Text(-0.08865640290120212, -0.2342772723644826, 'type'),\n",
       " 23: Text(0.05921254193267405, -0.6176359304937585, 'give'),\n",
       " 24: Text(-0.24414780128308222, 0.41552118932596926, 'corresponding'),\n",
       " 25: Text(-0.0033777258750930946, 0.5059147201561344, 'construct'),\n",
       " 26: Text(0.12109519202330948, 0.35071686490934506, 'support'),\n",
       " 27: Text(-0.21334881434958014, 0.03337602813970498, 'use'),\n",
       " 28: Text(-0.4212865379902862, -0.1753225353989069, 'solve'),\n",
       " 29: Text(0.05369995491441751, -0.392903862718526, 'mixed')}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:matplotlib.font_manager:findfont: Matching :family=sans-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=12.0.\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXNonUnicode' (STIXNonUniIta.ttf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeFourSym' (STIXSizFourSymReg.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeFourSym' (STIXSizFourSymBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeTwoSym' (STIXSizTwoSymBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmb10' (cmb10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Serif' (DejaVuSerif-Italic.ttf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeOneSym' (STIXSizOneSymBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans Mono' (DejaVuSansMono-Oblique.ttf) oblique normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans Display' (DejaVuSansDisplay.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeTwoSym' (STIXSizTwoSymReg.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXGeneral' (STIXGeneralBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans Mono' (DejaVuSansMono.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXNonUnicode' (STIXNonUniBolIta.ttf) italic normal bold normal>) = 11.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeOneSym' (STIXSizOneSymReg.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXGeneral' (STIXGeneralItalic.ttf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans Mono' (DejaVuSansMono-BoldOblique.ttf) oblique normal bold normal>) = 11.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans' (DejaVuSans-BoldOblique.ttf) oblique normal bold normal>) = 1.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXNonUnicode' (STIXNonUniBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXGeneral' (STIXGeneralBolIta.ttf) italic normal bold normal>) = 11.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeThreeSym' (STIXSizThreeSymReg.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans' (DejaVuSans-Oblique.ttf) oblique normal 400 normal>) = 1.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmmi10' (cmmi10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeThreeSym' (STIXSizThreeSymBol.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans Mono' (DejaVuSansMono-Bold.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans' (DejaVuSans-Bold.ttf) normal normal bold normal>) = 0.33499999999999996\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmex10' (cmex10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Serif' (DejaVuSerif-BoldItalic.ttf) italic normal bold normal>) = 11.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXGeneral' (STIXGeneral.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Serif Display' (DejaVuSerifDisplay.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmr10' (cmr10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Serif' (DejaVuSerif-Bold.ttf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXNonUnicode' (STIXNonUni.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Sans' (DejaVuSans.ttf) normal normal 400 normal>) = 0.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmsy10' (cmsy10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmtt10' (cmtt10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'DejaVu Serif' (DejaVuSerif.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'cmss10' (cmss10.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'STIXSizeFiveSym' (STIXSizFiveSymReg.ttf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Math' (MathJax_Math-Italic.otf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_SansSerif' (MathJax_SansSerif-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Math' (MathJax_Math-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Size2' (MathJax_Size2-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Main' (MathJax_Main-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Typewriter' (MathJax_Typewriter-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_SansSerif' (MathJax_SansSerif-Bold.otf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Size1' (MathJax_Size1-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Caligraphic' (MathJax_Caligraphic-Bold.otf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Vector' (MathJax_Vector-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Caligraphic' (MathJax_Caligraphic-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Main' (MathJax_Main-Italic.otf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Fraktur' (MathJax_Fraktur-Bold.otf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'FontAwesome' (fontawesome-webfont.ttf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_AMS' (MathJax_AMS-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Script' (MathJax_Script-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Size3' (MathJax_Size3-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Math' (MathJax_Math-BoldItalic.otf) italic normal bold normal>) = 11.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_WinChrome' (MathJax_WinChrome-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_SansSerif' (MathJax_SansSerif-Italic.otf) italic normal 400 normal>) = 11.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Main' (MathJax_Main-Bold.otf) normal normal bold normal>) = 10.335\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Fraktur' (MathJax_Fraktur-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'FontAwesome' (FontAwesome.otf) normal normal 400 normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Size4' (MathJax_Size4-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_Vector-Bold' (MathJax_Vector-Bold.otf) normal normal bold normal>) = 10.335\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:matplotlib.font_manager:findfont: score(<Font 'MathJax_WinIE6' (MathJax_WinIE6-Regular.otf) normal normal regular normal>) = 10.05\n",
      "DEBUG:matplotlib.font_manager:findfont: Matching :family=sans-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=12.0 to DejaVu Sans ('/home/ceteri/anaconda3/lib/python3.7/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:matplotlib.pyplot:Loaded backend module://ipykernel.pylab.backend_inline version unknown.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig = plt.figure(figsize=(9, 9))\n",
    "pos = nx.spring_layout(lemma_graph)\n",
    "\n",
    "nx.draw(lemma_graph, pos=pos, with_labels=False, font_weight=\"bold\")\n",
    "nx.draw_networkx_labels(lemma_graph, pos, labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now to run the algorithm, we use `PageRank` – which is approximately _eigenvalue centrality_ – to calculate ranks for each of the nodes in the lemma graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: 0.025190055141357165,\n",
       " 1: 0.09709174565608479,\n",
       " 2: 0.03658656272138432,\n",
       " 3: 0.022947339381092696,\n",
       " 4: 0.07548767636963792,\n",
       " 5: 0.01884004785394172,\n",
       " 6: 0.01884004785394172,\n",
       " 7: 0.019767433918161055,\n",
       " 8: 0.03093250736456608,\n",
       " 9: 0.031636552282656216,\n",
       " 10: 0.0250439175297852,\n",
       " 11: 0.03969617593153302,\n",
       " 12: 0.02513276636673567,\n",
       " 13: 0.0390375393827704,\n",
       " 14: 0.02428614673389346,\n",
       " 15: 0.02428614673389346,\n",
       " 16: 0.031629446298645975,\n",
       " 17: 0.06334806476862227,\n",
       " 18: 0.061826419749828485,\n",
       " 19: 0.03201021345587308,\n",
       " 20: 0.02404712231242087,\n",
       " 21: 0.029468555366439973,\n",
       " 22: 0.04816979699201436,\n",
       " 23: 0.010894627731045426,\n",
       " 24: 0.026930354700088012,\n",
       " 25: 0.03165915652710971,\n",
       " 26: 0.029382686223731833,\n",
       " 27: 0.019263849959229938,\n",
       " 28: 0.01982332786706688,\n",
       " 29: 0.016743716826448173}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ranks = nx.pagerank(lemma_graph)\n",
    "ranks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 0.09709174565608479 system\n",
      "4 0.07548767636963792 set\n",
      "17 0.06334806476862227 minimal\n",
      "18 0.061826419749828485 solution\n",
      "22 0.04816979699201436 type\n",
      "11 0.03969617593153302 inequation\n",
      "13 0.0390375393827704 consider\n",
      "2 0.03658656272138432 linear\n",
      "19 0.03201021345587308 algorithm\n",
      "25 0.03165915652710971 construct\n",
      "9 0.031636552282656216 equation\n",
      "16 0.031629446298645975 component\n",
      "8 0.03093250736456608 diophantine\n",
      "21 0.029468555366439973 generating\n",
      "26 0.029382686223731833 support\n",
      "24 0.026930354700088012 corresponding\n",
      "0 0.025190055141357165 compatibility\n",
      "12 0.02513276636673567 nonstrict\n",
      "10 0.0250439175297852 strict\n",
      "14 0.02428614673389346 upper\n",
      "15 0.02428614673389346 bound\n",
      "20 0.02404712231242087 construction\n",
      "3 0.022947339381092696 constraint\n",
      "28 0.01982332786706688 solve\n",
      "7 0.019767433918161055 criterion\n",
      "27 0.019263849959229938 use\n",
      "5 0.01884004785394172 natural\n",
      "6 0.01884004785394172 number\n",
      "29 0.016743716826448173 mixed\n",
      "23 0.010894627731045426 give\n"
     ]
    }
   ],
   "source": [
    "for node_id, rank in sorted(ranks.items(), key=lambda x: x[1], reverse=True):\n",
    "    print(node_id, rank, labels[node_id])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define a function to collect the top-ranked phrases from the lemma graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "\n",
    "def collect_phrases (chunk, phrases, counts):\n",
    "    chunk_len = chunk.end - chunk.start + 1\n",
    "    sq_sum_rank = 0.0\n",
    "    non_lemma = 0\n",
    "    compound_key = set([])\n",
    "\n",
    "    for i in range(chunk.start, chunk.end):\n",
    "        token = doc[i]\n",
    "        key = (token.lemma_, token.pos_)\n",
    "        \n",
    "        if key in seen_lemma:\n",
    "            node_id = list(seen_lemma.keys()).index(key)\n",
    "            rank = ranks[node_id]\n",
    "            sq_sum_rank += rank\n",
    "            compound_key.add(key)\n",
    "        \n",
    "            if logger:\n",
    "                logger.debug(\" {} {} {} {}\".format(token.lemma_, token.pos_, node_id, rank))\n",
    "        else:\n",
    "            non_lemma += 1\n",
    "    \n",
    "    # although the noun chunking is greedy, we discount the ranks using a\n",
    "    # point estimate based on the number of non-lemma tokens within a phrase\n",
    "    non_lemma_discount = chunk_len / (chunk_len + (2.0 * non_lemma) + 1.0)\n",
    "\n",
    "    # use root mean square (RMS) to normalize the contributions of all the tokens\n",
    "    phrase_rank = math.sqrt(sq_sum_rank / (chunk_len + non_lemma))\n",
    "    phrase_rank *= non_lemma_discount\n",
    "\n",
    "    # remove spurious punctuation\n",
    "    phrase = chunk.text.lower().replace(\"'\", \"\")\n",
    "\n",
    "    # create a unique key for the the phrase based on its lemma components\n",
    "    compound_key = tuple(sorted(list(compound_key)))\n",
    "    \n",
    "    if not compound_key in phrases:\n",
    "        phrases[compound_key] = set([ (phrase, phrase_rank) ])\n",
    "        counts[compound_key] = 1\n",
    "    else:\n",
    "        phrases[compound_key].add( (phrase, phrase_rank) )\n",
    "        counts[compound_key] += 1\n",
    "\n",
    "    if logger:\n",
    "        logger.debug(\"{} {} {} {} {} {}\".format(phrase_rank, chunk.text, chunk.start, chunk.end, chunk_len, counts[compound_key]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Collect the top-ranked phrases based on both the noun chunks and the named entities..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DEBUG:PyTR: compatibility NOUN 0 0.025190055141357165\n",
      "DEBUG:PyTR:0.07481838030465979 Compatibility 0 1 2 1\n",
      "DEBUG:PyTR: system NOUN 1 0.09709174565608479\n",
      "DEBUG:PyTR:0.14688751982088183 systems 2 3 2 1\n",
      "DEBUG:PyTR: linear ADJ 2 0.03658656272138432\n",
      "DEBUG:PyTR: constraint NOUN 3 0.022947339381092696\n",
      "DEBUG:PyTR:0.10565323773654285 linear constraints 4 6 3 1\n",
      "DEBUG:PyTR: set NOUN 4 0.07548767636963792\n",
      "DEBUG:PyTR:0.06868755180600318 the set 7 9 3 1\n",
      "DEBUG:PyTR: natural ADJ 5 0.01884004785394172\n",
      "DEBUG:PyTR: number NOUN 6 0.01884004785394172\n",
      "DEBUG:PyTR:0.08405366110543994 natural numbers 10 12 3 1\n",
      "DEBUG:PyTR: criterion NOUN 7 0.019767433918161055\n",
      "DEBUG:PyTR:0.06627792311867262 Criteria 13 14 2 1\n",
      "DEBUG:PyTR: compatibility NOUN 0 0.025190055141357165\n",
      "DEBUG:PyTR:0.07481838030465979 compatibility 15 16 2 2\n",
      "DEBUG:PyTR: system NOUN 1 0.09709174565608479\n",
      "DEBUG:PyTR:0.07789887100276421 a system 17 19 3 2\n",
      "DEBUG:PyTR: linear ADJ 2 0.03658656272138432\n",
      "DEBUG:PyTR: Diophantine PROPN 8 0.03093250736456608\n",
      "DEBUG:PyTR: equation NOUN 9 0.031636552282656216\n",
      "DEBUG:PyTR:0.1259559430077718 linear Diophantine equations 20 23 4 1\n",
      "DEBUG:PyTR: strict ADJ 10 0.0250439175297852\n",
      "DEBUG:PyTR: inequation NOUN 11 0.03969617593153302\n",
      "DEBUG:PyTR:0.11017607509798653 strict inequations 24 26 3 1\n",
      "DEBUG:PyTR: nonstrict NOUN 12 0.02513276636673567\n",
      "DEBUG:PyTR: inequation NOUN 11 0.03969617593153302\n",
      "DEBUG:PyTR:0.11025165160180313 nonstrict inequations 28 30 3 1\n",
      "DEBUG:PyTR: upper ADJ 14 0.02428614673389346\n",
      "DEBUG:PyTR: bound NOUN 15 0.02428614673389346\n",
      "DEBUG:PyTR:0.09543220119650414 Upper bounds 33 35 3 1\n",
      "DEBUG:PyTR: component NOUN 16 0.031629446298645975\n",
      "DEBUG:PyTR:0.08383773520404489 components 36 37 2 1\n",
      "DEBUG:PyTR: minimal ADJ 17 0.06334806476862227\n",
      "DEBUG:PyTR: set NOUN 4 0.07548767636963792\n",
      "DEBUG:PyTR:0.0952198714085986 a minimal set 38 41 4 1\n",
      "DEBUG:PyTR: solution NOUN 18 0.061826419749828485\n",
      "DEBUG:PyTR:0.1172143523159633 solutions 42 43 2 1\n",
      "DEBUG:PyTR: algorithm NOUN 19 0.03201021345587308\n",
      "DEBUG:PyTR:0.0843408606072513 algorithms 44 45 2 1\n",
      "DEBUG:PyTR: construction NOUN 20 0.02404712231242087\n",
      "DEBUG:PyTR:0.07310133349204889 construction 46 47 2 1\n",
      "DEBUG:PyTR: minimal ADJ 17 0.06334806476862227\n",
      "DEBUG:PyTR: generating NOUN 21 0.029468555366439973\n",
      "DEBUG:PyTR: set NOUN 4 0.07548767636963792\n",
      "DEBUG:PyTR:0.1640996265710316 minimal generating sets 48 51 4 1\n",
      "DEBUG:PyTR: solution NOUN 18 0.061826419749828485\n",
      "DEBUG:PyTR:0.1172143523159633 solutions 52 53 2 2\n",
      "DEBUG:PyTR: type NOUN 22 0.04816979699201436\n",
      "DEBUG:PyTR:0.05486904693906117 all types 54 56 3 1\n",
      "DEBUG:PyTR: system NOUN 1 0.09709174565608479\n",
      "DEBUG:PyTR:0.14688751982088183 systems 57 58 2 3\n",
      "DEBUG:PyTR: criterion NOUN 7 0.019767433918161055\n",
      "DEBUG:PyTR:0.035149176660130545 These criteria 61 63 3 2\n",
      "DEBUG:PyTR: corresponding ADJ 24 0.026930354700088012\n",
      "DEBUG:PyTR: algorithm NOUN 19 0.03201021345587308\n",
      "DEBUG:PyTR:0.06204175981712335 the corresponding algorithms 64 67 4 1\n",
      "DEBUG:PyTR: minimal ADJ 17 0.06334806476862227\n",
      "DEBUG:PyTR: support VERB 26 0.029382686223731833\n",
      "DEBUG:PyTR: set NOUN 4 0.07548767636963792\n",
      "DEBUG:PyTR:0.10465046837630346 a minimal supporting set 69 73 5 1\n",
      "DEBUG:PyTR: solution NOUN 18 0.061826419749828485\n",
      "DEBUG:PyTR:0.1172143523159633 solutions 74 75 2 3\n",
      "DEBUG:PyTR: consider VERB 13 0.0390375393827704\n",
      "DEBUG:PyTR: type NOUN 22 0.04816979699201436\n",
      "DEBUG:PyTR: system NOUN 1 0.09709174565608479\n",
      "DEBUG:PyTR:0.08278948056400119 all the considered types systems 80 85 6 1\n",
      "DEBUG:PyTR: system NOUN 1 0.09709174565608479\n",
      "DEBUG:PyTR:0.14688751982088183 systems 86 87 2 4\n",
      "DEBUG:PyTR: mixed ADJ 29 0.016743716826448173\n",
      "DEBUG:PyTR: type NOUN 22 0.04816979699201436\n",
      "DEBUG:PyTR:0.1103235416443912 mixed types 88 90 3 1\n",
      "DEBUG:PyTR: Diophantine PROPN 8 0.03093250736456608\n",
      "DEBUG:PyTR:0.082908929105731 Diophantine 21 22 2 1\n"
     ]
    }
   ],
   "source": [
    "phrases = {}\n",
    "counts = {}\n",
    "\n",
    "for chunk in doc.noun_chunks:\n",
    "    collect_phrases(chunk, phrases, counts)\n",
    "\n",
    "for ent in doc.ents:\n",
    "    collect_phrases(ent, phrases, counts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since noun chunks can be expressed in different ways (e.g., they may have articles or prepositions), we need to find a minimum span for each phrase based on combinations of lemmas..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import operator\n",
    "\n",
    "min_phrases = {}\n",
    "\n",
    "for compound_key, rank_tuples in phrases.items():\n",
    "    l = list(rank_tuples)\n",
    "    l.sort(key=operator.itemgetter(1), reverse=True)\n",
    "    \n",
    "    phrase, rank = l[0]\n",
    "    count = counts[compound_key]\n",
    "    \n",
    "    min_phrases[phrase] = (rank, count)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Yield the results of TextRank..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "minimal generating sets 1 0.1640996265710316\n",
      "systems 4 0.14688751982088183\n",
      "linear diophantine equations 1 0.1259559430077718\n",
      "solutions 3 0.1172143523159633\n",
      "mixed types 1 0.1103235416443912\n",
      "nonstrict inequations 1 0.11025165160180313\n",
      "strict inequations 1 0.11017607509798653\n",
      "linear constraints 1 0.10565323773654285\n",
      "a minimal supporting set 1 0.10465046837630346\n",
      "upper bounds 1 0.09543220119650414\n",
      "a minimal set 1 0.0952198714085986\n",
      "algorithms 1 0.0843408606072513\n",
      "natural numbers 1 0.08405366110543994\n",
      "components 1 0.08383773520404489\n",
      "diophantine 1 0.082908929105731\n",
      "all the considered types systems 1 0.08278948056400119\n",
      "compatibility 2 0.07481838030465979\n",
      "construction 1 0.07310133349204889\n",
      "the set 1 0.06868755180600318\n",
      "criteria 2 0.06627792311867262\n",
      "the corresponding algorithms 1 0.06204175981712335\n",
      "all types 1 0.05486904693906117\n"
     ]
    }
   ],
   "source": [
    "for phrase, (rank, count) in sorted(min_phrases.items(), key=lambda x: x[1][0], reverse=True):\n",
    "    print(phrase, count, rank)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Just for kicks, compare with raw results of the non-chunked lemma nodes..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "system 0.09709174565608479\n",
      "set 0.07548767636963792\n",
      "minimal 0.06334806476862227\n",
      "solution 0.061826419749828485\n",
      "type 0.04816979699201436\n",
      "inequation 0.03969617593153302\n",
      "consider 0.0390375393827704\n",
      "linear 0.03658656272138432\n",
      "algorithm 0.03201021345587308\n",
      "construct 0.03165915652710971\n",
      "equation 0.031636552282656216\n",
      "component 0.031629446298645975\n",
      "diophantine 0.03093250736456608\n",
      "generating 0.029468555366439973\n",
      "support 0.029382686223731833\n",
      "corresponding 0.026930354700088012\n",
      "compatibility 0.025190055141357165\n",
      "nonstrict 0.02513276636673567\n",
      "strict 0.0250439175297852\n",
      "upper 0.02428614673389346\n",
      "bound 0.02428614673389346\n",
      "construction 0.02404712231242087\n",
      "constraint 0.022947339381092696\n",
      "solve 0.01982332786706688\n",
      "criterion 0.019767433918161055\n",
      "use 0.019263849959229938\n",
      "natural 0.01884004785394172\n",
      "number 0.01884004785394172\n",
      "mixed 0.016743716826448173\n",
      "give 0.010894627731045426\n"
     ]
    }
   ],
   "source": [
    "for node_id, rank in sorted(ranks.items(), key=lambda x: x[1], reverse=True):\n",
    "    print(labels[node_id], rank)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}