Testing_Mallet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Mallet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another algorithm for topic modeling is implemented in the java-based software Mallet. For this to work **you need to download and install Mallet** from http://mallet.cs.umass.edu/download.php.\n",
    "Mallet uses plain text as input, so none of the preprocessing steps of this package are available for Mallet topic modeling as of yet."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loading modules from DARIAH-Topics library\n",
    "First, we have to get access to the functionalities of the library by importing them. For using its functions we use the prefix of the toolbox's submodules (pre, visual and mallet)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from dariah_topics import preprocessing as pre\n",
    "from dariah_topics import visualization as visual\n",
    "from dariah_topics import mallet as mal"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Activating inline output in Jupyter notebook\n",
    "The following line will just tell the notebook to show graphics in the output frames."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Setting the parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define path to corpus folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path_to_corpus = \"corpus_txt\"\n",
    "#path_to_corpus = \"/home/sina/DARIAH/Topics/corpus_txt/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Path to mallet folder \n",
    "\n",
    "Now we must tell the library where to find the local instance of mallet. If you managed to install Mallet, it is sufficient set `path_to_mallet = \"mallet\"`, if you just store Mallet in a local folder, you have to specify the path to the binary explictly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#path_to_mallet = \"/home/sina/mallet-2.0.8/bin/mallet\"\n",
    "path_to_mallet = \"/home/severin/Documents/dariah/app/mallet/bin/mallet\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Output folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "outfolder = \"tutorial_supplementals/mallet_output\"\n",
    "binary = \"tutorial_supplementals/mallet_output/binary.mallet\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The Mallet corpus model\n",
    "\n",
    "Finally, we can give all these folder paths to a Mallet function that handles all the preprocessing steps and creates a Mallet-specific corpus model object."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO mallet: Running MALLET with ['/home/severin/Documents/dariah/app/mallet/bin/mallet', 'import-dir', '--input', 'corpus_txt', '--remove-stopwords', '--keep-sequence', '--output', 'tutorial_supplementals/mallet_output/binary.mallet'] ...\n",
      "INFO mallet: Saving MALLET binary to tutorial_supplementals/mallet_output/binary.mallet ...\n"
     ]
    }
   ],
   "source": [
    "mallet_model = mal.create_mallet_binary(path_to_mallet=path_to_mallet, path_to_file=False,\n",
    "                                        path_to_corpus=path_to_corpus, output_file=binary,\n",
    "                                        encoding=None, token_regex=None, preserve_case=False,\n",
    "                                        remove_stopwords=True, stoplist=None, extra_stopwords=None,\n",
    "                                        stop_pattern_file=None, skip_header=False, skip_html=False,\n",
    "                                        replacement_files=None, deletion_files=None, gram_sizes=None,\n",
    "                                        keep_sequence=True, keep_sequence_bigrams=False,\n",
    "                                        binary_features=False, save_text_in_source=False,\n",
    "                                        print_output=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Model creation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Warning: this step can take quite a while!** Meaning something between some seconds and some hours depending on corpus size and the number of passes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO mallet: Accessing Mallet with ['/home/severin/Documents/dariah/app/mallet/bin/mallet', 'train-topics', '--input', 'tutorial_supplementals/mallet_output/binary.mallet', '--num-topics', '3', '--num-iterations', '500', '--num-threads', '1', '--num-top-words', '10', '--num-icm-iterations', '0', '--random-seed', '0', '--optimize-interval', '0', '--optimize-burn-in', '200', '--use-symmetric-alpha', '--alpha', '5.0', '--beta', '0.01', '--output-topic-keys', 'tutorial_supplementals/mallet_output/topic_keys.txt', '--output-doc-topics', 'tutorial_supplementals/mallet_output/doc_topics.txt', '--doc-topics-threshold', '0.0', '--topic-word-weights-file', 'tutorial_supplementals/mallet_output/topic_word_weights.txt', '--word-topic-counts-file', 'tutorial_supplementals/mallet_output/word_topic_counts.txt', '--diagnostics-file', 'tutorial_supplementals/mallet_output/diagnostics.xml', '--xml-topic-report', 'tutorial_supplementals/mallet_output/topic_report.xml', '--xml-topic-phrase-report', 'tutorial_supplementals/mallet_output/topic_phrase_report.xml', '--output-model', 'tutorial_supplementals/mallet_output/mallet.model', '--output-model-interval', '0', '--output-state', 'tutorial_supplementals/mallet_output/state.gz', '--output-state-interval', '0', '--inferencer-filename', 'tutorial_supplementals/mallet_output/inferencer', '--evaluator-filename', 'tutorial_supplementals/mallet_output/evaluator'] ...\n"
     ]
    }
   ],
   "source": [
    "num_topics = 3\n",
    "output_folder = mal.create_mallet_model(path_to_mallet = path_to_mallet, \n",
    "                                        path_to_binary = mallet_model, \n",
    "                                        folder_for_output = outfolder,\n",
    "                                        num_top_words = 10,\n",
    "                                        num_iterations = 500,\n",
    "                                        num_topics = num_topics\n",
    "                                        )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.4. Create document-topic matrix\n",
    "\n",
    "The generated model object can now be translated into a human-readable document-topic matrix (that is a actually a pandas data frame) that constitutes our principle exchange format for topic modeling results. For generating the matrix from a Gensim model, we can use the following function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc_topic = mal.show_doc_topic_matrix(\"tutorial_supplementals/mallet_output\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Visualization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can see the topics in the model with the following function:\n",
    "\n",
    "**Hint:** Depending on the number of topics chosen in step 2, you might have to adjust *num_topics* in this step accordingly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Key 1</th>\n",
       "      <th>Key 2</th>\n",
       "      <th>Key 3</th>\n",
       "      <th>Key 4</th>\n",
       "      <th>Key 5</th>\n",
       "      <th>Key 6</th>\n",
       "      <th>Key 7</th>\n",
       "      <th>Key 8</th>\n",
       "      <th>Key 9</th>\n",
       "      <th>Key 10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Topic 1</th>\n",
       "      <td>man</td>\n",
       "      <td>holmes</td>\n",
       "      <td>back</td>\n",
       "      <td>sir</td>\n",
       "      <td>time</td>\n",
       "      <td>eyes</td>\n",
       "      <td>night</td>\n",
       "      <td>face</td>\n",
       "      <td>good</td>\n",
       "      <td>head</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Topic 2</th>\n",
       "      <td>ravager</td>\n",
       "      <td>point</td>\n",
       "      <td>matter</td>\n",
       "      <td>smallest</td>\n",
       "      <td>dog</td>\n",
       "      <td>proper</td>\n",
       "      <td>fact</td>\n",
       "      <td>moore</td>\n",
       "      <td>slippers</td>\n",
       "      <td>general</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Topic 3</th>\n",
       "      <td>conan</td>\n",
       "      <td>great</td>\n",
       "      <td>feet</td>\n",
       "      <td>city</td>\n",
       "      <td>made</td>\n",
       "      <td>door</td>\n",
       "      <td>floor</td>\n",
       "      <td>walls</td>\n",
       "      <td>land</td>\n",
       "      <td>back</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Key 1   Key 2   Key 3     Key 4 Key 5   Key 6  Key 7  Key 8  \\\n",
       "Topic 1      man  holmes    back       sir  time    eyes  night   face   \n",
       "Topic 2  ravager   point  matter  smallest   dog  proper   fact  moore   \n",
       "Topic 3    conan   great    feet      city  made    door  floor  walls   \n",
       "\n",
       "            Key 9   Key 10  \n",
       "Topic 1      good     head  \n",
       "Topic 2  slippers  general  \n",
       "Topic 3      land     back  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num_topics = 3\n",
    "mal.show_topics_keys(\"tutorial_supplementals/mallet_output\", num_topics = num_topics)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1. Distribution of topics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Distribution of topics over all documents\n",
    "\n",
    "The distribution of topics over all documents can now be visualized in a heat map:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa6adcf0c>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "heatmap = visual.doc_topic_heatmap(doc_topic)\n",
    "heatmap.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Distribution of topics in a single documents\n",
    "\n",
    "To take closer look on the topics in a single text, we can use the follwing function that shows all the topics in a text and their respective proportions. To select the document, we have to give its index to the function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'matplotlib.pyplot' from '/home/severin/.local/lib/python3.5/site-packages/matplotlib/pyplot.py'>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZln6P8tDja60Jcj9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa6c6dfac>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#document_nr = 7\n",
    "visual.plot_doc_topics(doc_topic, 7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "def make_wordle_from_mallet(word_weights_file,topics,words,outfolder, \n",
    "                            font_path, dpi):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "def read_mallet_output(word_weights_file):\n",
    "        \"\"\"Reads Mallet output (topics with words and word weights) into dataframe.\"\"\" \n",
    "        word_scores = pd.read_table(word_weights_file, header=None, sep=\"\\t\")\n",
    "        word_scores = word_scores.sort(columns=[0,2], axis=0, ascending=[True, False])\n",
    "        word_scores_grouped = word_scores.groupby(0)\n",
    "        return word_scores_grouped"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'DataFrame' object has no attribute 'sort'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-13-71aadb80548b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mword_weights_file\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"tutorial_supplementals/mallet_output/topic_word_weights.txt\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mword_scores_grouped\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mread_mallet_output\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mword_weights_file\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mtopic_nr\u001b[0m  \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mword_scores_grouped\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_group\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtopic_nr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-12-f9f805db8c59>\u001b[0m in \u001b[0;36mread_mallet_output\u001b[0;34m(word_weights_file)\u001b[0m\n\u001b[1;32m      3\u001b[0m         \u001b[0;34m\"\"\"Reads Mallet output (topics with words and word weights) into dataframe.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m         \u001b[0mword_scores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mread_table\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mword_weights_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheader\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msep\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"\\t\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m         \u001b[0mword_scores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mword_scores\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msort\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mascending\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m         \u001b[0mword_scores_grouped\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mword_scores\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgroupby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mword_scores_grouped\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/severin/.local/lib/python3.5/site-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36m__getattr__\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m   2968\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_info_axis\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2969\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2970\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mobject\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__getattribute__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2971\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2972\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__setattr__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'DataFrame' object has no attribute 'sort'"
     ]
    }
   ],
   "source": [
    "word_weights_file = \"tutorial_supplementals/mallet_output/topic_word_weights.txt\"\n",
    "word_scores_grouped = read_mallet_output(word_weights_file)\n",
    "topic_nr  = 0\n",
    "word_scores_grouped.get_group(topic_nr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_wordlewords(word_scores_grouped, number_of_top_words, topic_nr):\n",
    "    \"\"\"Transform Mallet output for wordle generation.\"\"\"\n",
    "    topic_word_scores = word_scores_grouped.get_group(topic_nr)\n",
    "    top_topic_word_scores = topic_word_scores.iloc[0:number_of_top_words]\n",
    "    topic_words = top_topic_word_scores.loc[:,1].tolist()\n",
    "    print(topic_words)\n",
    "    word_scores = top_topic_word_scores.loc[:,2].tolist()\n",
    "    print(word_scores)\n",
    "    wordlewords = \"\"\n",
    "    j = 0\n",
    "    for word in topic_words:\n",
    "        word = word\n",
    "        score = word_scores[j]\n",
    "        j += 1\n",
    "        wordlewords = wordlewords + ((word + \" \") + str(score))\n",
    "    return wordlewords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_top_words = 7\n",
    "topic_nr = 2\n",
    "wordlewords = get_wordlewords(word_scores_grouped, number_of_top_words, topic_nr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "wordlewords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def save_topicRanks(topicWordFile, outfolder, filename):\n",
    "    \"\"\"Save a list of topics with their rank by topic score.\"\"\"\n",
    "    print(\"Launched save_topicRanks.\")\n",
    "    with open(topicWordFile, \"r\") as infile:\n",
    "        topicRanks = pd.read_csv(infile, sep=\"\\t\", header=None)\n",
    "        topicRanks = topicRanks.drop(2, axis=1)\n",
    "        topicRanks.rename(columns={0:\"Number\"}, inplace=True)\n",
    "        topicRanks.rename(columns={1:\"Score\"}, inplace=True)\n",
    "        #topicRanks.sort(columns=[\"Score\"], ascending=False, inplace=True)\n",
    "        topicRanks[\"Rank\"] = topicRanks[\"Score\"].rank(ascending=False)\n",
    "        #print(topicRanks.head())\n",
    "        ## Saving the file.\n",
    "        if not os.path.exists(outfolder):\n",
    "            os.makedirs(outfolder)\n",
    "        outfile = outfolder + filename\n",
    "        with open(outfile, \"w\") as outfile: \n",
    "            topicRanks.to_csv(outfile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "outfolder = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output/\"\n",
    "topicWordFile = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output/topic_word_weights.txt\"\n",
    "filename = \"test.csv\"\n",
    "save_topicRanks(topicWordFile, outfolder, filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_topicRank(topic, topicRanksFile):\n",
    "    #print(\"getting topic rank.\")\n",
    "    with open(topicRanksFile, \"r\") as infile:\n",
    "        topicRanks = pd.read_csv(infile, sep=\",\", index_col=0)\n",
    "        rank = int(topicRanks.iloc[topic][\"Rank\"])\n",
    "    return rank"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "topicRanksFile = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output/test.csv\"\n",
    "topic = 2\n",
    "get_topicRank(topic, topicRanksFile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_wordle_from_mallet(word_weights_file, \n",
    "                            numOfTopics,\n",
    "                            number_of_top_words,\n",
    "                            outfolder,\n",
    "                            topicRanksFile,\n",
    "                            dpi):\n",
    "    \"\"\"Generate wordles from Mallet output, using the wordcloud module.\"\"\"\n",
    "    print(\"\\nLaunched make_wordle_from_mallet.\")\n",
    "    import os\n",
    "    for topic in range(0,numOfTopics):\n",
    "    ## Gets the text for one topic.\n",
    "        number_of_top_words = 7\n",
    "        word_scores_grouped = read_mallet_output(word_weights_file)\n",
    "        text = get_wordlewords(word_scores_grouped, number_of_top_words, topic)\n",
    "        wordcloud = WordCloud(width=600, height=400, background_color=\"white\", margin=4).generate(text)\n",
    "        default_colors = wordcloud.to_array()\n",
    "        rank = get_topicRank(topic, topicRanksFile)\n",
    "        figure_title = \"topic \"+ str(topic) + \" (\"+str(rank)+\"/\"+str(numOfTopics)+\")\"       \n",
    "        #plt.imshow(wordcloud.recolor(color_func=get_color_scale, random_state=3))\n",
    "        plt.imshow(default_colors)\n",
    "        plt.imshow(wordcloud)\n",
    "        plt.title(figure_title, fontsize=30)\n",
    "        plt.axis(\"off\")\n",
    "                \n",
    "        ## Saving the image file.\n",
    "        if not os.path.exists(outfolder):\n",
    "            os.makedirs(outfolder)\n",
    "            figure_filename = \"wordle_tp\"+\"{:03d}\".format(topic) + \".png\"\n",
    "            plt.savefig(outfolder + figure_filename, dpi=dpi)\n",
    "            return plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from wordcloud import WordCloud\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "\n",
    "word_weights_file = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output/topic_word_weights.txt\"\n",
    "numOfTopics = 1\n",
    "outfolder = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output\"\n",
    "dpi = 200\n",
    "topicRanksFile = \"/home/sina/DARIAH/Topics/tutorial_supplementals/mallet_output/test.csv\"\n",
    "font_path = \"/home/sina/DARIAH/Topics/tutorial_supplementals/font\"\n",
    "number_of_top_words = 10\n",
    "\n",
    "wordle = plot_wordle_from_mallet(word_weights_file, \n",
    "                                 numOfTopics,\n",
    "                                 number_of_top_words,\n",
    "                                 outfolder,\n",
    "                                 topicRanksFile,\n",
    "                                 dpi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}