experiments/BigARTM_run_example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Simple draft experiment using BigARTM Python API"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here one can find a code to understand how to work with [BigARTM](http://bigartm.org) and build shallow and hierarchical models.\n",
    "\n",
    "Before you can run script, you should collect dataset and install BigARTM.\n",
    "You can find instructions in [README](https://github.com/src-d/snippet-ranger).\n",
    "\n",
    "To install BigARTM you can use script from [./helpers/install_bigartm.sh](https://github.com/src-d/snippet-ranger/tree/master/helpers/install_bigartm.sh)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:bigARTM_experiment:artm version: 0.8.3\n"
     ]
    }
   ],
   "source": [
    "import glob\n",
    "import os\n",
    "import logging\n",
    "\n",
    "import artm\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "logging.basicConfig(level=logging.INFO)\n",
    "log = logging.getLogger(\"bigARTM_experiment\")\n",
    "log.info(\"artm version: {}\".format(artm.version()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "seed = 42\n",
    "# BigARTM optimisation params\n",
    "num_topics = 50  # topics numbers\n",
    "cache_theta = True\n",
    "num_document_passes = 1\n",
    "num_collection_passes_output = 1\n",
    "num_collection_passes_full = 35"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Please change it to your data directory.\n",
    "data_dir = \"/storage/konstantin/data/\"  \n",
    "fc = \"_fc\"\n",
    "libname = \"matplotlib\" + fc\n",
    "# from where load vowpal_wabbit dataset\n",
    "vw_path = os.path.join(data_dir, \"vowpal_wabbit/vw_{}.txt\".format(libname))\n",
    "\n",
    "# where to save batches for BigARTM\n",
    "batches_dir = os.path.join(data_dir, \"batches/{}/\".format(libname))\n",
    "# where to save BigARTM model\n",
    "artm_model_dir = os.path.join(data_dir, \"models/hartm_{}/\".format(libname))  \n",
    "# where to save dictionary for BigARTM\n",
    "dict_path = os.path.join(batches_dir, \"artm.dict\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:bigARTM_experiment:Batches directory found. Use batches from it and dictionary from dict_path.\n"
     ]
    }
   ],
   "source": [
    "if not os.path.exists(batches_dir):\n",
    "    log.info(\"Batches directory does not exists, create new batches from vowpal_wabbit format\")\n",
    "    batch_vectorizer = artm.BatchVectorizer(\n",
    "        data_path=vw_path, data_format=\"vowpal_wabbit\", collection_name=libname, target_folder=batches_dir)\n",
    "    batch_vectorizer.dictionary.save(dict_path)\n",
    "    dictionary = batch_vectorizer.dictionary\n",
    "else:\n",
    "    log.info(\"Batches directory found. Use batches from it and dictionary from dict_path.\")\n",
    "    batch_vectorizer = artm.BatchVectorizer(\n",
    "        data_path=batches_dir, data_format=\"batches\", collection_name=libname)\n",
    "    dictionary = artm.Dictionary(name=None, dictionary_path=dict_path, data_path=batches_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:bigARTM_experiment:num_batches: 311\n",
      "INFO:bigARTM_experiment:batch_size : 1000\n",
      "INFO:bigARTM_experiment:num_topics : 50\n"
     ]
    }
   ],
   "source": [
    "# important info output\n",
    "log.info(\"num_batches: %d\", batch_vectorizer.num_batches)\n",
    "log.info(\"batch_size : %d\", batch_vectorizer.batch_size)\n",
    "log.info(\"num_topics : %d\", num_topics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# whitch scores BigARTM model should calculate during optimization\n",
    "scores = [\n",
    "    artm.PerplexityScore(name='PerplexityScore', dictionary=dictionary),\n",
    "    artm.SparsityPhiScore(name='SparsityPhiScore'),\n",
    "    artm.SparsityThetaScore(name='SparsityThetaScore'),\n",
    "    artm.TopicKernelScore(name='TopicKernelScore', probability_mass_threshold=0.3),\n",
    "    artm.TopTokensScore(name='TopTokensScore', num_tokens=10),\n",
    "         ]\n",
    "\n",
    "# Add regularizers to BigARTM model. \n",
    "# Check out http://bigartm.readthedocs.io/en/latest/tutorials/regularizers_descr.html\n",
    "regularizers = [\n",
    "    artm.SmoothSparseThetaRegularizer(name='SparseTheta', tau=-0.001),\n",
    "    artm.SmoothSparsePhiRegularizer(name='SparsePhi', tau=-0.001),\n",
    "               ]\n",
    "\n",
    "model_artm = artm.ARTM(num_topics=num_topics, \n",
    "                       dictionary=dictionary, \n",
    "                       scores=scores, \n",
    "                       regularizers=regularizers,\n",
    "                       num_document_passes=num_document_passes,\n",
    "                       cache_theta=cache_theta,\n",
    "                       show_progress_bars=True,\n",
    "                       theta_columns_naming='title')\n",
    "\n",
    "model_artm.initialize(dictionary=dictionary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def print_measures(model_artm):\n",
    "    \"\"\"\n",
    "    Helper function to output score info\n",
    "    \"\"\"\n",
    "    for key in model_artm.score_tracker:\n",
    "        if hasattr(model_artm.score_tracker[key], \"last_value\"):\n",
    "            print('{}: {}{:10.3f}'.format(key, \" \" * ( 20 - len(key)), model_artm.score_tracker[key].last_value))\n",
    "\n",
    "    print('Kernel contrast:      {:10.3f}'.format(\n",
    "        model_artm.score_tracker['TopicKernelScore'].last_average_contrast))\n",
    "\n",
    "    print('Kernel purity:        {0:10.3f}'.format(\n",
    "        model_artm.score_tracker['TopicKernelScore'].last_average_purity))\n",
    "    \n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3a5b7002f1454de0ba9aff3f42f4f2cd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Pass: ', max=35), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2bc374652aeb435c884cfe5e7ab6f710",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d06aa87d72cd47d09ff99fba9391b947",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ddca4a5f0e03412ab00f9916d98b54e0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9cb0944dbe7d451fa6b99d04cf02fae7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5043e40e62fd43ecb1abc44b62dddc16",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "48ec67774c0a4102bc235167dfa1c639",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b8dd15735d934bf9a9769df2bdc9d32a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "427d9d113b03489091ca620a1ee1c123",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "649958ce31de44579a23e59a5b270b1b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8606199f55344118bd8c5c6b7e54e1af",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "42a56aa0c3f14c2ea250f483e2ba8ea7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e55d46a289044e55aa1f28a2e17c895a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "39f94e1232224c5fa353fcc3292a3abc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "241aaf53923948e0b723fa43084fece6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9f81a3ca4fc342c38876776fb6fde328",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f9c010c9e185427898c86f8d3129f358",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1aab6f1e14f341c0be079b3aa14c9aad",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "16bbb877e9174f489c1f8e0dba2b5e3e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4e06b80f038f4c62b60c03d3a101d684",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0d5a47f04bb5483681d79b12a6745132",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a6e9bb13f02c41468d926dc392d7f124",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "04421ff61f8c407db1e173c3215ddce6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ba26468caf644e9abc2c67944991ef87",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7b12952718704e25a2a9c346939f9acc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "42cc9b4f5b2e4560a38784888065f508",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0dd86db0cd034a47bd7d7bc2f07dd6f9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7bea7923287f4bdbad96709842b75d92",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "65be3ad8be224f00ba898170271eaae9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "aa34a50d668b4f5b9787eca11a01f1b4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ef0b0969d66c401c8459e1f66c55523e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e8222c47c8f34d619642ffd7ae5b5a19",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7a954b28581f4dce9e419a6c9e8e5864",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6e2c53110d46403da55a6adae4bb99a0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0b5eeeb7c0df4233b4353f267f17d913",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6fd236ffdbc041a2a17b2401d74fa512",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=311), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "SparsityPhiScore:          0.979\n",
      "SparsityThetaScore:        0.769\n",
      "PerplexityScore:           6.553\n",
      "Kernel contrast:           0.964\n",
      "Kernel purity:             1.000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "def train_model(model_artm, print_measures_flag=False):\n",
    "    if print_measures_flag:\n",
    "        for i in range(num_collection_passes_full):\n",
    "            model_artm.fit_offline(batch_vectorizer=batch_vectorizer, num_collection_passes=num_collection_passes_output)\n",
    "            print_measures(model_artm)\n",
    "    else:\n",
    "        model_artm.fit_offline(batch_vectorizer=batch_vectorizer, num_collection_passes=num_collection_passes_full)\n",
    "        print_measures(model_artm)\n",
    "train_model(model_artm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get Perplexity Score plot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Can help to get if model has converged or not"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f60b174d780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = model_artm.score_tracker['PerplexityScore'].value\n",
    "x = range(len(y))\n",
    "plt.plot(x, np.log(y), 'r--', linewidth=2)\n",
    "plt.xlabel(\"Iterations count\")\n",
    "plt.ylabel(\"log of perplexity\")\n",
    "plt.grid(True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "topic_0:  ['plot']\n",
      "topic_1:  ['plot', 'semilogx', 'paint']\n",
      "topic_2:  ['scatter', 'arrow', 'normpdf', 'delay', 'spring', 'bone']\n",
      "topic_3:  ['legend', 'get', 'showtip', 'hidetip', 'revcmap']\n",
      "topic_4:  ['ylabel', 'xlabel', 'yscale', 'xscale', 'title', 'legend', 'plot', 'ylim']\n",
      "topic_5:  ['savefig', 'ion', 'violinplot', 'ioff', 'color', 'summer', 'winter', 'figure']\n",
      "topic_6:  ['write', 'specgram', 'circle', 'cvt', 'data', 'autumn']\n",
      "topic_7:  ['use', 'box', 'main', 'step', 'axhspan', 'scaled', 'center', 'polar', 'rgrids', 'movavg']\n",
      "topic_8:  ['join', 'matshow', 'mkdirs', 'figaspect', 'spectral']\n",
      "topic_9:  ['pop', 'connect', 'resize']\n",
      "topic_10:  ['remove', 'hold', 'ishold', 'detrend', 'sci', 'inner', 'triplot', 'barbs', 'tricontour', 'spy']\n",
      "topic_11:  ['sqrt', 'strftime', 'scale', 'start', 'end', 'reset', 'element', 'fix', 'interpgrid', 'accept']\n",
      "topic_12:  ['type', 'flush', 'param', 'iternext', 'axisinfo', 'mutatedx', 'myiter', 'prism']\n",
      "topic_13:  ['keys', 'values', 'cycler', 'drange', 'containsx', 'getp']\n",
      "topic_14:  ['range', 'push', 'twiny', 'arc']\n",
      "topic_15:  ['xticks', 'cla', 'yticks', 'hexbin', 'twinx', 'gradient', 'shrink', 'toarray', 'hpack', 'kern']\n",
      "topic_16:  ['clear', 'vlines', 'symbol', 'streamplot', 'unit', 'eventplot', 'findobj', 'matchfunc', 'null', 'locked']\n",
      "topic_17:  ['bar', 'shape', 'barh', 'relim', 'frange', 'bubble', 'sphere']\n",
      "topic_18:  ['text', 'ylim', 'xlim']\n",
      "topic_19:  ['imread', 'stop', 'imsave', 'process', 'tripcolor', 'saving', 'fromstr', 'tostr']\n",
      "topic_20:  ['gcf', 'clf', 'empty', 'destroy', 'ignore', 'csd', 'adjust']\n",
      "topic_21:  ['append', 'tostr', 'safezip']\n",
      "topic_22:  ['apply', 'ord', 'back', 'vpack', 'figlegend', 'dst']\n",
      "topic_23:  ['draw', 'boxplot', 'cubehelix']\n",
      "topic_24:  ['open', 'demean', 'dummy']\n",
      "topic_25:  ['set', 'transform', 'union', 'pause', 'clean', 'fedit']\n",
      "topic_26:  ['hist', 'func', 'group', 'frozen', 'axvspan', 'hlines', 'stem', 'tzname', 'xkcd', 'plane']\n",
      "topic_27:  ['unique', 'flatten', 'render', 'rotate', 'ginput', 'gray', 'shade', 'clamp', 'popall', 'reject']\n",
      "topic_28:  ['close', 'function', 'extract']\n",
      "topic_29:  ['run', 'report', 'select', 'constant', 'mainloop', 'enable', 'unselect', 'cleanup', 'inferno', 'reshow']\n",
      "topic_30:  ['title', 'subplot', 'labels', 'figure']\n",
      "topic_31:  ['dot', 'convert', 'quad', 'frame', 'verify', 'fixlist', 'diag']\n",
      "topic_32:  ['update', 'find', 'home']\n",
      "topic_33:  ['copy', 'slopes']\n",
      "topic_34:  ['range', 'power', 'griddata', 'skew', 'base']\n",
      "topic_35:  ['imshow', 'axis', 'pcolormesh', 'finddir', 'connected']\n",
      "topic_36:  ['join', 'setup', 'teardown']\n",
      "topic_37:  ['grid', 'joined', 'diag', 'edges', 'writeln', 'flipy']\n",
      "topic_38:  ['show', 'tricontourf', 'xcorr', 'funcy']\n",
      "topic_39:  ['save', 'contourf', 'seek', 'parse', 'forward', 'dedent', 'disconnect', 'subs', 'clim', 'context']\n",
      "topic_40:  ['format', 'zoom']\n",
      "topic_41:  ['figure', 'colorbar', 'pcolor', 'dist', 'inside', 'binom', 'colors']\n",
      "topic_42:  ['axes', 'findfont', 'quiver', 'loglog', 'figtext', 'grow', 'quiverkey', 'check', 'bump', 'hsv']\n",
      "topic_43:  ['output', 'knownfailureif', 'gci', 'utcoffset', 'padded', 'closeto', 'hatch', 'waitforbuttonpress']\n",
      "topic_44:  ['gca', 'figure']\n",
      "topic_45:  ['append']\n",
      "topic_46:  ['setp', 'fill', 'nonsingular', 'sca', 'finish', 'after', 'stroke', 'init', 'hitlist', 'before']\n",
      "topic_47:  ['contour', 'errorbar', 'clabel', 'blit', 'stackplot', 'closest', 'gauss', 'press', 'trigger', 'maxdict']\n",
      "topic_48:  ['list', 'pie', 'table', 'psd', 'refresh', 'name', 'pad', 'project', 'suppress', 'makekey']\n",
      "topic_49:  ['add', 'jet', 'pan', 'pilread', 'set']\n"
     ]
    }
   ],
   "source": [
    "# Get most valuable topic tokens to understand topic model better.\n",
    "for topic_name in model_artm.topic_names:\n",
    "    print(topic_name + ': ', model_artm.score_tracker['TopTokensScore'].last_tokens[topic_name])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f5ff40bb1d0>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABkwAAAaeCAYAAAATBBN2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f60a81c4d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot topic-token matrix.\n",
    "import seaborn as sns\n",
    "sns.set_style(\"whitegrid\")\n",
    "plt.figure(figsize=(30, 30))\n",
    "sns.heatmap(model_artm.phi_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save topic-token (phi) and token-snippet (theta) matrixes\n",
    "model_artm.phi_.to_csv(os.path.join(data_dir, \"matlab_phi_mat.tsv\"), sep='\\t',header=False, index=False)\n",
    "with open(os.path.join(data_dir, \"matlab_phi_indx.tsv\"), \"w\") as f:\n",
    "    f.write(\"\\n\".join(list(model_artm.phi_.index)) + '\\n')\n",
    "os.chmod(os.path.join(data_dir, \"matlab_phi_mat.tsv\"), 0o666)\n",
    "os.chmod(os.path.join(data_dir, \"matlab_phi_indx.tsv\"), 0o666)\n",
    "\n",
    "theta_matrix = model_artm.get_theta()\n",
    "theta_matrix.T.to_csv(os.path.join(data_dir, \"matlab_theta_mat.tsv\"), sep='\\t',header=False, index=False)\n",
    "with open(os.path.join(data_dir, \"matlab_theta_indx.tsv\"), \"w\") as f:\n",
    "    f.write(\"\\n\".join(list(theta_matrix.T.index)) + '\\n')\n",
    "os.chmod(os.path.join(data_dir, \"matlab_theta_mat.tsv\"), 0o666)\n",
    "os.chmod(os.path.join(data_dir, \"matlab_theta_indx.tsv\"), 0o666)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Transform new snippet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In case you need to get topic representation for new snippet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Self-sufficient independent inports\n",
    "import os\n",
    "from ast2vec import bblfsh_roles\n",
    "from snippet_ranger import utils\n",
    "from snippet_ranger.models.snippet import Snippet\n",
    "from snippet_ranger.model2.snippet2bow import SnippetModel2FuncCallsBOW\n",
    "from ast2vec import UASTModel, DocumentFrequencies\n",
    "from ast2vec.vw_dataset import convert_bow_to_vw\n",
    "from bblfsh import BblfshClient\n",
    "import artm\n",
    "\n",
    "def code2batch(code):\n",
    "    \"\"\"\n",
    "    Converts string with code directly to new BigARTM batch:\n",
    "    Code -> Snippet Model -> BOW Model -> Vowpal Wabbit -> BigARTM batch\n",
    "    Only way to transform new snippet.\n",
    "    You need bblfsh server v0.0.7 running with 0.8.2 Python driver. \n",
    "    \"\"\"\n",
    "    filepath = './temp/snippet.py'\n",
    "    os.makedirs(os.path.dirname(filepath), exist_ok=True)\n",
    "    with open(filepath, 'w') as f:\n",
    "        f.write(code)\n",
    "\n",
    "    bc = BblfshClient(\"172.17.0.1:9432\")\n",
    "    res = bc.parse(filepath, language='Python')\n",
    "\n",
    "    s = Snippet()\n",
    "    s.construct(\"code\", [\"code\"], [res.uast], [code],\n",
    "                positions_start=[0], positions_end=[code.count(\"\\n\")])\n",
    "\n",
    "    docfreq = '/storage/konstantin/data/dfs_fc/matplotlib_df.asdf'\n",
    "    df = DocumentFrequencies().load(docfreq)\n",
    "\n",
    "    s2b = SnippetModel2FuncCallsBOW(len(df), df, num_processes=1)\n",
    "    bow = s2b.convert_model(s)\n",
    "    snbfn = 'snippet_vw.txt'\n",
    "    batches_dir = 'snippet_batch'\n",
    "    dict_path = \"snippet_dict\"\n",
    "    convert_bow_to_vw(bow, snbfn)\n",
    "    try:\n",
    "        os.remove(dict_path)\n",
    "    except:\n",
    "        pass\n",
    "    batch_vectorizer = artm.BatchVectorizer(\n",
    "        data_path=snbfn, data_format=\"vowpal_wabbit\", collection_name=\"Snippet\",\n",
    "        target_folder=batches_dir)\n",
    "    dictionary = batch_vectorizer.dictionary\n",
    "    return batch_vectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:docfreq:Reading /storage/konstantin/data/dfs_fc/matplotlib_df.asdf...\n",
      "INFO:docfreq:Building the docfreq dictionary...\n",
      "INFO:snippet2bow:Choosing the vocabulary...\n",
      "INFO:bow2vw:Writing snippet_vw.txt\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6cb729c5dd5b4c12bf53f87d9fee8494",
       "version_major": 2,
       "version_minor": 0
      },
      "text/html": [
       "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n",
       "<p>\n",
       "  If you're reading this message in Jupyter Notebook or JupyterLab, it may mean\n",
       "  that the widgets JavaScript is still loading. If this message persists, it\n",
       "  likely means that the widgets JavaScript library is either not installed or\n",
       "  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n",
       "  Widgets Documentation</a> for setup instructions.\n",
       "</p>\n",
       "<p>\n",
       "  If you're reading this message in another notebook frontend (for example, a static\n",
       "  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n",
       "  it may mean that your frontend doesn't currently support widgets.\n",
       "</p>\n"
      ],
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Batch: ', max=1), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "          code/code_0_7\n",
      "topic_0        0.088108\n",
      "topic_1        0.083005\n",
      "topic_2        0.000000\n",
      "topic_3        0.000000\n",
      "topic_4        0.000000\n",
      "topic_5        0.000000\n",
      "topic_6        0.000000\n",
      "topic_7        0.000000\n",
      "topic_8        0.000000\n",
      "topic_9        0.000000\n",
      "topic_10       0.000000\n",
      "topic_11       0.000000\n",
      "topic_12       0.000000\n",
      "topic_13       0.000000\n",
      "topic_14       0.169631\n",
      "topic_15       0.000000\n",
      "topic_16       0.000000\n",
      "topic_17       0.000000\n",
      "topic_18       0.000000\n",
      "topic_19       0.000000\n",
      "topic_20       0.000000\n",
      "topic_21       0.000000\n",
      "topic_22       0.000000\n",
      "topic_23       0.000000\n",
      "topic_24       0.000000\n",
      "topic_25       0.000000\n",
      "topic_26       0.000000\n",
      "topic_27       0.000000\n",
      "topic_28       0.000000\n",
      "topic_29       0.000000\n",
      "topic_30       0.000000\n",
      "topic_31       0.000000\n",
      "topic_32       0.000000\n",
      "topic_33       0.000000\n",
      "topic_34       0.143744\n",
      "topic_35       0.000000\n",
      "topic_36       0.000000\n",
      "topic_37       0.000000\n",
      "topic_38       0.278849\n",
      "topic_39       0.000000\n",
      "topic_40       0.000000\n",
      "topic_41       0.236663\n",
      "topic_42       0.000000\n",
      "topic_43       0.000000\n",
      "topic_44       0.000000\n",
      "topic_45       0.000000\n",
      "topic_46       0.000000\n",
      "topic_47       0.000000\n",
      "topic_48       0.000000\n",
      "topic_49       0.000000\n"
     ]
    }
   ],
   "source": [
    "# check:\n",
    "code = \"\"\"\n",
    "from matplotlib import pyplot as plt\n",
    "plt.figure(figsize=(10,10))\n",
    "x = range(20)\n",
    "y = range(20)\n",
    "plt.plot(x,y)\n",
    "plt.show()\n",
    "\"\"\"\n",
    "sn_batch_vectorizer = code2batch(code)\n",
    "x = model_artm.transform(sn_batch_vectorizer)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# How to train hierarchical topic model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**On going...**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SparsityPhiScore:          0.713\n",
      "SparsityThetaScore:        0.618\n",
      "PerplexityScore:          45.412\n",
      "Kernel contrast:           0.994\n",
      "Kernel purity:             0.990\n",
      "\n",
      "[286.8885498046875, 81.32962799072266, 75.4485092163086, 69.44510650634766, 64.07815551757812, 59.80461502075195, 56.62196350097656, 54.332862854003906, 52.71855926513672, 51.5455322265625, 50.62627410888672, 49.85963439941406, 49.19702911376953, 48.626129150390625, 48.139373779296875, 47.72269821166992, 47.36608123779297, 47.04627990722656, 46.733150482177734, 46.422706604003906, 46.14145278930664, 45.91961669921875, 45.764137268066406, 45.659912109375, 45.589324951171875, 45.539188385009766, 45.50178909301758, 45.47330093383789, 45.451759338378906, 45.43587875366211, 45.424705505371094, 45.41743087768555, 45.41328811645508, 45.41157150268555, 45.411617279052734]\n",
      "['format', 'figure', 'gca', 'imshow', 'dot', 'type', 'draw', 'grid', 'text', 'remove']\n",
      "['plot', 'close', 'savefig', 'open', 'legend', 'hist', 'title', 'ylabel', 'xlabel', 'values']\n",
      "['append', 'list', 'show', 'pop', 'keys', 'scatter', 'gcf', 'clf', 'boxplot', 'connect']\n",
      "['range', 'join', 'copy', 'sqrt', 'write', 'update', 'get', 'apply', 'save', 'scale']\n",
      "SparsityPhiScore:          0.895\n",
      "SparsityThetaScore:        0.848\n",
      "PerplexityScore:          20.348\n",
      "Kernel contrast:           0.991\n",
      "Kernel purity:             0.986\n",
      "\n",
      "[294.226318359375, 83.62002563476562, 74.64241027832031, 62.78110122680664, 51.599700927734375, 42.058048248291016, 34.76455307006836, 30.120901107788086, 27.365554809570312, 25.6710262298584, 24.55423355102539, 23.777936935424805, 23.226598739624023, 22.824100494384766, 22.521495819091797, 22.27754783630371, 22.06615447998047, 21.87464714050293, 21.69774055480957, 21.534704208374023, 21.389875411987305, 21.268522262573242, 21.167850494384766, 21.078472137451172, 20.989309310913086, 20.895681381225586, 20.806766510009766, 20.728445053100586, 20.652618408203125, 20.57744026184082, 20.511110305786133, 20.457101821899414, 20.41361427307129, 20.377979278564453, 20.348190307617188]\n",
      "['figure', 'gca', 'grid', 'unique', 'contourf', 'errorbar', 'cla', 'psd', 'loglog', 'draw']\n",
      "['values', 'setp', 'transform', 'add', 'ylim', 'xlim', 'pcolormesh', 'union', 'joined', 'render']\n",
      "['append', 'boxplot', 'gradient', 'axvspan', 'step', 'griddata', 'toarray', 'fix', 'finish', 'relim']\n",
      "['update', 'get', 'apply', 'save', 'resize', 'seek', 'convert', 'writeln', 'findfont', 'figtext']\n",
      "['plot', 'open', 'shape', 'destroy', 'end', 'element', 'yscale', 'xscale', 'quiverkey', 'suppress']\n",
      "['show', 'scatter', 'gcf', 'connect', 'clear', 'find', 'hexbin', 'ion', 'table', 'edges']\n",
      "['close', 'empty', 'contour', 'start', 'report', 'fill', 'group', 'clabel', 'quiver', 'knownfailureif']\n",
      "['imshow', 'set', 'axis', 'colorbar', 'subplot', 'imsave', 'scaled', 'center', 'closeto', 'show']\n",
      "['join', 'imread', 'open', 'forward', 'pad', 'demean', 'pan', 'range', 'onselect', 'slopes']\n",
      "['format', 'bar', 'run', 'draw', 'flatten', 'func', 'vlines', 'push', 'grow', 'vpack']\n",
      "['pop', 'clf', 'hold', 'barh', 'ishold', 'twinx', 'specgram', 'sca', 'detrend', 'pause']\n",
      "['copy', 'sqrt', 'write', 'scale', 'power', 'flush', 'ord', 'frozen', 'pie', 'diag']\n",
      "['dot', 'type', 'text', 'remove', 'axes', 'function', 'cycler', 'refresh', 'symbol', 'jet']\n",
      "['list', 'keys', 'use', 'strftime', 'rotate', 'main', 'delta', 'findobj', 'normpdf', 'matchfunc']\n",
      "['range', 'draw', 'blit', 'disconnect', 'utcoffset', 'cvt', 'closest', 'set', 'push', 'shade']\n",
      "['savefig', 'legend', 'hist', 'title', 'ylabel', 'xlabel', 'xticks', 'figure', 'subplot', 'yticks']\n",
      "SparsityPhiScore:          0.952\n",
      "SparsityThetaScore:        0.930\n",
      "PerplexityScore:           6.438\n",
      "Kernel contrast:           0.915\n",
      "Kernel purity:             0.895\n",
      "\n",
      "[289.85211181640625, 78.14535522460938, 60.90772247314453, 37.67391586303711, 21.091224670410156, 14.068900108337402, 10.954139709472656, 9.432174682617188, 8.657767295837402, 8.183581352233887, 7.851684093475342, 7.605587959289551, 7.400688648223877, 7.2288641929626465, 7.091105937957764, 6.976451873779297, 6.879032611846924, 6.796517372131348, 6.727474689483643, 6.670317649841309, 6.623518466949463, 6.585663318634033, 6.555349826812744, 6.531247138977051, 6.512126922607422, 6.496860504150391, 6.4845356941223145, 6.4744672775268555, 6.466120719909668, 6.4591288566589355, 6.453248500823975, 6.448313236236572, 6.4441962242126465, 6.440803527832031, 6.438027858734131]\n",
      "['grid', 'contourf', 'psd', 'loglog', 'ginput', 'hlines', 'hitlist', 'inside', 'bump', 'fixlist']\n",
      "['joined', 'stop', 'output', 'matshow', 'setup', 'subs', 'clim', 'barbs', 'padded', 'home']\n",
      "['scatter', 'hexbin', 'context', 'gcf', 'step', 'bone', 'show', 'add', 'pcolor', 'find']\n",
      "['get', 'apply', 'convert', 'writeln', 'findfont', 'enable', 'inner', 'param', 'drange', 'check']\n",
      "['plot', 'destroy', 'end', 'element', 'quiverkey', 'suppress', 'toval', 'axisinfo', 'funcy', 'comment']\n",
      "['gcf', 'find', 'ion', 'pcolor', 'ioff', 'gray', 'quad', 'finddir', 'color', 'hot']\n",
      "['contour', 'report', 'clabel', 'quiver', 'ignore', 'mkdirs', 'spy', 'skew', 'frange', 'waitforbuttonpress']\n",
      "['set', 'colorbar', 'imsave', 'scaled', 'center', 'labels', 'closeto', 'inferno', 'imshow', 'magma']\n",
      "['join', 'imread', 'forward', 'pan', 'slopes', 'hsv', 'colors', 'failer', 'pad', 'open']\n",
      "['bar', 'run', 'vlines', 'push', 'grow', 'vpack', 'format', 'flatten', 'draw', 'func']\n",
      "['hold', 'barh', 'ishold', 'specgram', 'sca', 'detrend', 'pause', 'gci', 'sci', 'arrow']\n",
      "['scale', 'pie', 'diag', 'select', 'violinplot', 'process', 'unselect', 'interpgrid', 'hpack', 'fromstr']\n",
      "['text', 'function', 'cycler', 'jet', 'dist', 'figaspect', 'shade', 'trigger', 'spectral', 'popall']\n",
      "['keys', 'delta', 'containsx', 'getp', 'list', 'dst', 'use', 'normpdf', 'add', 'rotate']\n",
      "['draw', 'push', 'shade']\n",
      "['title', 'xticks', 'zoom', 'yticks', 'ylabel', 'xlabel', 'legend', 'xlim', 'savefig', 'ylim']\n",
      "['savefig', 'nonsingular', 'streamplot', 'null', 'xlim', 'infodict', 'legend', 'figure', 'xlabel', 'ylabel']\n",
      "['remove', 'bubble', 'adjust', 'type', 'refresh', 'dot', 'axes', 'closest', 'cycler', 'text']\n",
      "['copy', 'extract', 'shrink', 'dedent', 'hpack', 'sqrt', 'kern', 'write', 'scale', 'power']\n",
      "['pop', 'clf', 'twinx', 'pick', 'acorr', 'figlegend', 'step', 'reshow', 'sca', 'hold']\n",
      "['transform', 'ylim', 'xlim', 'render', 'mainloop', 'stem', 'xkcd', 'maxdict', 'add', 'clim']\n",
      "['append', 'relim', 'gauss', 'tzname', 'cmd', 'show']\n",
      "['type', 'axes', 'refresh', 'symbol', 'saving', 'text', 'closest', 'axis', 'twiny', 'dot']\n",
      "['boxplot', 'arc', 'cmd', 'makekey', 'append', 'reshow', 'plane', 'show', 'relim', 'step']\n",
      "['copy', 'frozen', 'write', 'binom', 'flush', 'power', 'sqrt', 'extract', 'scale', 'hpack']\n",
      "['values', 'union', 'clean', 'transform', 'clamp', 'fedit', 'add', 'pcolormesh', 'stop', 'render']\n",
      "['savefig', 'isvector', 'figure', 'hist', 'clf', 'legend', 'xlabel', 'xticks', 'ylabel', 'title']\n",
      "['flatten', 'func', 'format', 'hot', 'vpack', 'shade', 'bar', 'vlines', 'run', 'draw']\n",
      "['close', 'fill', 'plot', 'clf', 'figure']\n",
      "['ylabel', 'xlabel', 'tripcolor', 'title', 'mutatedx', 'kern', 'xlim', 'ylim', 'legend', 'clf']\n",
      "['subplot', 'closeto', 'show', 'figure', 'step']\n",
      "['dot', 'text', 'closest', 'axis', 'type', 'loglog', 'axes', 'dist']\n",
      "['empty', 'close', 'accept', 'tostr', 'onmove', 'fill', 'ignore', 'base', 'clf', 'start']\n",
      "['legend', 'stackplot', 'polar', 'rgrids', 'layers', 'savefig', 'xlabel', 'title', 'ylabel', 'xlim']\n",
      "['range', 'closest', 'shade', 'draw']\n",
      "['setp', 'add', 'pcolormesh', 'parse', 'semilogx', 'stem', 'render', 'ylim', 'xlim', 'clim']\n",
      "['join', 'pad', 'demean', 'onselect', 'open', 'labels', 'imread', 'forward']\n",
      "['plot', 'yscale', 'xscale', 'xlim', 'ylim', 'xlabel', 'ylabel', 'open', 'clf', 'legend']\n",
      "['append', 'step', 'tzname', 'griddata', 'show', 'finish', 'arc', 'fix']\n",
      "['update', 'save', 'resize', 'seek', 'figtext', 'circle', 'get', 'showtip', 'inner', 'convert']\n",
      "['format', 'run', 'vlines', 'flatten', 'draw', 'bar', 'vpack']\n",
      "['xlabel', 'ylabel', 'yticks', 'box', 'xticks', 'title', 'legend', 'xlim', 'figure', 'ylim']\n",
      "['write', 'flush', 'ord', 'csd', 'cvt', 'process', 'sqrt', 'dedent', 'inner', 'copy']\n",
      "['hist', 'xlim', 'semilogx', 'ylim', 'clf', 'title', 'figure', 'xticks', 'savefig', 'xlabel']\n",
      "['figure', 'reset', 'twiny', 'gca', 'grid', 'hlines', 'psd']\n",
      "['append', 'gradient', 'axvspan', 'griddata', 'toarray', 'delay', 'wrapper', 'boxplot', 'step', 'show']\n",
      "['append', 'finish', 'after', 'stroke', 'init', 'iternext', 'before', 'plane', 'hatch', 'flipy']\n",
      "['gca', 'errorbar', 'closest', 'cubehelix', 'figure', 'reset', 'unique', 'grid', 'contourf', 'twiny']\n",
      "['list', 'use', 'normpdf', 'keys', 'findobj', 'add', 'matchfunc', 'getp']\n",
      "['open', 'plot', 'pilread']\n",
      "['sqrt', 'power', 'scale', 'diag', 'process', 'copy', 'flush', 'write', 'extract', 'ord']\n",
      "['show', 'table', 'edges', 'name', 'constant', 'scatter', 'gcf', 'pcolor', 'step', 'connect']\n",
      "['format', 'run', 'shade', 'imshow', 'func', 'draw', 'hot', 'flatten', 'push', 'bar']\n",
      "['list', 'strftime', 'rotate', 'main', 'findobj', 'matchfunc', 'reject', 'connected', 'keys', 'reshow']\n",
      "['append', 'fix', 'prctile', 'sphere', 'gauss', 'wrapper', 'relim']\n",
      "['plot', 'shape', 'open', 'clf', 'press', 'element', 'end', 'yscale', 'xscale']\n",
      "['unique', 'cla', 'unit', 'eventplot', 'cleanup', 'gca', 'figure', 'errorbar', 'contourf', 'grid']\n",
      "['imshow', 'axis', 'hot', 'colorbar', 'show', 'step', 'figure', 'shade', 'subplot', 'reshow']\n",
      "['close', 'start', 'group', 'knownfailureif', 'data', 'autumn', 'summer', 'winter', 'contour', 'fill']\n",
      "['range', 'blit', 'set']\n",
      "['plot', 'press', 'open', 'xscale', 'clf', 'yscale', 'shape']\n",
      "['show', 'ion', 'scatter', 'pcolor', 'step', 'pause', 'gray', 'find', 'gcf', 'connect']\n",
      "['range', 'disconnect', 'utcoffset', 'cvt', 'set', 'blit', 'closest']\n",
      "['connect', 'clear', 'scatter', 'find', 'show', 'add', 'finddir', 'color', 'hexbin', 'hot']\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import shutil\n",
    "import artm\n",
    "from artm.dictionary import Dictionary\n",
    "\n",
    "nums_topics = [4, 16, 64]\n",
    "\n",
    "hARTM = artm.hARTM(seed=seed,\n",
    "                   dictionary=dictionary, \n",
    "                   scores=scores, \n",
    "                   regularizers=regularizers,\n",
    "                   num_document_passes=num_document_passes,\n",
    "                   cache_theta=cache_theta,\n",
    "                   theta_columns_naming='title')\n",
    "\n",
    "# train from level to level\n",
    "for num_topics in nums_topics:\n",
    "    level = hARTM.add_level(num_topics=num_topics, parent_level_weight=1)\n",
    "    level.initialize(dictionary)\n",
    "    train_model(level)\n",
    "\n",
    "    print(level.score_tracker['PerplexityScore'].value)\n",
    "\n",
    "    top_tokens = level.score_tracker['TopTokensScore'].last_tokens\n",
    "    for topic_name in level.topic_names:\n",
    "        print(top_tokens[topic_name])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model\n",
    "model_dir = os.path.join(data_dir, 'hARTM')\n",
    "if os.path.exists(model_dir):\n",
    "    shutil.rmtree(model_dir)\n",
    "os.makedirs(model_dir, exist_ok=True)\n",
    "hARTM.save(model_dir)"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}