stranasum-visual.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0490b8c7",
   "metadata": {},
   "source": [
    "# Sebastian Petrik - Stranasum - Visualization module"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2d10f6ed-ed5d-4725-94c9-f62038074815",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-05-17 13:42:23.403440: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-05-17 13:42:23.882648: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory\n",
      "2023-05-17 13:42:23.882693: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory\n",
      "2023-05-17 13:42:23.882698: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n"
     ]
    }
   ],
   "source": [
    "%pip install contractions --quiet\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import contractions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "522ff097-f6c0-4c26-90c2-8f29bda43471",
   "metadata": {},
   "outputs": [],
   "source": [
    "class TextProcessor:\n",
    "    \n",
    "    # Text cleanup\n",
    "    def clean_text(self, text: str):\n",
    "\n",
    "        # lowercase\n",
    "        text = str(text).lower()\n",
    "\n",
    "        # remove &-escaped characters\n",
    "        text = re.sub(r\"&.[1-9]+;\",\" \", str(text))\n",
    "\n",
    "        # remove escaped characters\n",
    "        text=re.sub(\"(\\\\t)\", ' ', str(text))\n",
    "        text=re.sub(\"(\\\\r)\", ' ', str(text))\n",
    "        text=re.sub(\"(\\\\n)\", ' ', str(text))\n",
    "\n",
    "        # remove double characters\n",
    "        text=re.sub(\"(__+)\", ' ', str(text))  #remove _ if it occurs more than one time consecutively\n",
    "        text=re.sub(\"(--+)\", ' ', str(text))   #remove - if it occurs more than one time consecutively\n",
    "        text=re.sub(\"(~~+)\", ' ', str(text))   #remove ~ if it occurs more than one time consecutively\n",
    "        text=re.sub(\"(\\+\\++)\", ' ', str(text))  #remove + if it occurs more than one time consecutively\n",
    "        text=re.sub(\"(\\.\\.+)\", ' ', str(text))  #remove . if it occurs more than one time consecutively\n",
    "        \n",
    "        # special - fix u.s. contraction in gigaword\n",
    "        text = re.sub(\"(u\\.s\\.)\", 'united states', str(text))\n",
    "        \n",
    "        # fix contractions to base form\n",
    "        text = contractions.fix(text)\n",
    "\n",
    "        #remove special tokens <>()|&©ø\"',;?~*!\n",
    "        text=re.sub(r\"[<>()|&©ø\\[\\]\\'\\\",;?~*!]\", ' ', str(text)).lower()\n",
    "\n",
    "        # CNN mail data cleanup\n",
    "        text=re.sub(\"(mailto:)\", ' ', str(text)) #remove mailto:\n",
    "        text=re.sub(r\"(\\\\x9\\d)\", ' ', str(text)) #remove \\x9* in text\n",
    "        text=re.sub(\"([iI][nN][cC]\\d+)\", 'INC_NUM', str(text)) #replace INC nums to INC_NUM\n",
    "        text=re.sub(\"([cC][mM]\\d+)|([cC][hH][gG]\\d+)\", 'CM_NUM', str(text)) #replace CM# and CHG# to CM_NUM\n",
    "\n",
    "        # url replacement into base form\n",
    "        try:\n",
    "            url = re.search(r'((https*:\\/*)([^\\/\\s]+))(.[^\\s]+)', str(text))\n",
    "            repl_url = url.group(3)\n",
    "            text = re.sub(r'((https*:\\/*)([^\\/\\s]+))(.[^\\s]+)',repl_url, str(text))\n",
    "        except:\n",
    "            pass\n",
    "\n",
    "\n",
    "        # handle dot at the end of words\n",
    "        text=re.sub(\"(\\.\\s+)\", ' ', str(text)) # remove\n",
    "        \n",
    "        text=re.sub(\"(\\-\\s+)\", ' ', str(text)) #remove - at end of words(not between)\n",
    "        text=re.sub(\"(\\:\\s+)\", ' ', str(text)) #remove : at end of words(not between)\n",
    "\n",
    "        #remove multiple spaces\n",
    "        text=re.sub(\"(\\s+)\",' ',str(text))\n",
    "\n",
    "        # apply lowercase again\n",
    "        text = text.lower().strip()\n",
    "        \n",
    "        # remove trailing dot, we will apply end of sequence anyway\n",
    "        text = re.sub(\"(\\.)$\", '', str(text)).strip()\n",
    "        \n",
    "        # gigaword - UNK token\n",
    "        text = re.sub(\"unk\", '', str(text).strip())\n",
    "        \n",
    "        # gigaword - change numbers to hashtags\n",
    "        text = re.sub(\"\\d\", \"#\", str(text).strip())\n",
    "\n",
    "        return text\n",
    "\n",
    "    def apply_special_tokens(self, text):\n",
    "        text = str(text).strip()\n",
    "        text = \"<sos> \" + str(text).strip() + \" <eos>\"\n",
    "        return text\n",
    "\n",
    "    def remove_special_tokens(self, text):\n",
    "        text = text.lower()\n",
    "        text = text.replace(\"<sos>\", \"\").replace(\"<eos>\", \"\")\n",
    "        text = text.strip()\n",
    "        return text\n",
    "\n",
    "class Summarizer:\n",
    "    \n",
    "    # Initialize using SummarizationModule or loaded graph tf module\n",
    "    def __init__(self, module: tf.Module):\n",
    "        self.module = module\n",
    "        self.processor = TextProcessor()\n",
    "        \n",
    "    def summarize(self, text: str):\n",
    "        prepared = self.processor.apply_special_tokens(self.processor.clean_text(text))\n",
    "        output_text, output_tensor, weights = self.module.predict(tf.constant([prepared]))\n",
    "        return prepared, self.processor.remove_special_tokens(bytes.decode(output_text.numpy())), output_tensor, weights\n",
    "    \n",
    "    # Shorthand for text output only\n",
    "    def __call__(self, text: str):\n",
    "        return self.summarize(text)[1]\n",
    "    \n",
    "module = tf.saved_model.load(\"./summarizer_module\")\n",
    "summarizer = Summarizer(module)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "id": "799c4da2-042f-4001-9dff-d75a46883736",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'estonia bans forest fires'"
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "euro = \"\"\"european union leaders gave slovenia a green light friday to join the eurozone next year , launching a new wave of expansion for the currently ##-nation single currency club .\"\"\"\n",
    "afgan = \"journalists and legislators in afghanistan have been outraged by the intelligence service 's new media guidelines that bar interviews with taliban leaders and criticism of foreign troops .\"\n",
    "gradu = \"hundreds of graduates from america 's most prestigious universities are temporarily setting aside their high-powered career ambitions to teach at inner-city schools , as part of a program to help children from low-income families .\"\n",
    "fires = \"estonians will miss their traditional midsummer bonfires in state-owned forests this year after the authorities on tuesday banned the practice because of risk of forest fires .\"\n",
    "\n",
    "article = fires\n",
    "prepared, summary, output, attention_weights = summarizer.summarize(article)\n",
    "summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "id": "79bcd446-2820-40d7-ad2e-bfa42d285969",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'<sos> journalists and legislators in afghanistan have been outraged by the intelligence service s new media guidelines that bar interviews with taliban leaders and criticism of foreign troops <eos>'"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prepared"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "91eac584-833f-4c02-b337-d045e11e93b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Shape: `(batch=1, num_heads, seq_len_q, seq_len_k)`.\n",
    "\n",
    "\n",
    "# example\n",
    "# attention = attention_heads[0]\n",
    "# attention.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "d432f767-c9a0-45d8-93cd-88d46cafbc7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# attention_heads[0][1, :(article_l)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "id": "f7c4866a-706b-41ab-b583-b8822a1d5af2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x2000 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# aggregated attention for all heads\n",
    "def extract_agg_attention(weights, article_len):\n",
    "    \n",
    "    attention_heads = tf.squeeze(attention_weights, 0)\n",
    "    attention = sum(attention_heads)[:, :(article_len)] # restrict by article len\n",
    "    \n",
    "    # custom clipping and scaling for visualisation\n",
    "    attention = tf.clip_by_value(\n",
    "        attention,\n",
    "        clip_value_min=0,\n",
    "        clip_value_max=1.6\n",
    "    )\n",
    "    \n",
    "    attention = (attention) / tf.reduce_max(attention)\n",
    "    \n",
    "    return attention\n",
    "\n",
    "def plot_attention_head(article_tok, sum_tok, attention):\n",
    "    # The model didn't generate `<START>` in the output. Skip it.\n",
    "    sum_tok_reduced = sum_tok[1:]\n",
    "\n",
    "    ax = plt.gca()\n",
    "    ax.matshow(tf.transpose(attention))\n",
    "    ax.set_yticks(range(len(article_tok)))\n",
    "    ax.set_xticks(range(len(sum_tok_reduced)))\n",
    "\n",
    "    # labels = [label.decode('utf-8') for label in in_tokens.numpy()]\n",
    "    ax.set_yticklabels(article_tok)\n",
    "\n",
    "    # labels = [label.decode('utf-8') for label in translated_tokens.numpy()]\n",
    "    ax.set_xticklabels(sum_tok_reduced, rotation=90)\n",
    "    \n",
    "def plot_agg_attention(article_tok, sum_tok, attention_weights):\n",
    "    plt.rcParams['figure.figsize'] = [20, 20]\n",
    "    plt.rc('xtick', labelsize=15) \n",
    "    plt.rc('ytick', labelsize=15) \n",
    "    plot_attention_head(\n",
    "        article_tok,\n",
    "        sum_tok,\n",
    "        extract_agg_attention(attention_weights, len(article_tok))\n",
    "    )\n",
    "    \n",
    "    \n",
    "article_tok = prepared.split(\" \")\n",
    "sum_tok = ['<sos>'] + summary.split(\" \") + ['<eos>'] # add back the tokens\n",
    "\n",
    "# plot_attention_head(article_tok, sum_tok, attention[:, :(article_l)])\n",
    "plot_agg_attention(article_tok, sum_tok, attention_weights)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c067048b-bf94-45d5-b59a-d33fb41338df",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}