{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Text Data Explanation Benchmarking: Abstractive Summarization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook demonstrates how to use the benchmark utility to benchmark the performance of an explainer for text data. In this demo, we showcase explanation performance for partition explainer on an Abstractive Summarization model. The metric used to evaluate is \"keep positive\". The masker used is Text Masker.\n",
    "\n",
    "The new benchmark utility uses the new API with MaskedModel as wrapper around user-imported model and evaluates masked values of inputs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nlp\n",
    "from transformers import AutoModelForSeq2SeqLM, AutoTokenizer\n",
    "\n",
    "import shap\n",
    "import shap.benchmark as benchmark"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Data and Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = AutoTokenizer.from_pretrained(\"sshleifer/distilbart-xsum-12-6\")\n",
    "model = AutoModelForSeq2SeqLM.from_pretrained(\"sshleifer/distilbart-xsum-12-6\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using custom data configuration default\n"
     ]
    }
   ],
   "source": [
    "dataset = nlp.load_dataset(\"xsum\", split=\"train\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "s = dataset[\"document\"][0:1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Explainer Object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "explainers.Partition is still in an alpha state, so use with caution...\n"
     ]
    }
   ],
   "source": [
    "explainer = shap.Explainer(model, tokenizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run SHAP Explanation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=48.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Partition explainer: 2it [00:43, 21.70s/it]                                       \n"
     ]
    }
   ],
   "source": [
    "shap_values = explainer(s)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define Metrics (Sort Order & Perturbation Method)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "sort_order = \"positive\"\n",
    "perturbation = \"keep\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Benchmark Explainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sp = benchmark.perturbation.SequentialPerturbation(explainer.model, explainer.masker, sort_order, perturbation)\n",
    "xs, ys, auc = sp.model_score(shap_values, s)\n",
    "sp.plot(xs, ys, auc)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}