{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Similarity Score\n",
    "This Notebook imports the questions and answers from the QuestionGeneration.ipynb output and scores the similarity\n",
    "\n",
    "Heavy inspiration taken from:\n",
    "\n",
    "https://github.com/karndeepsingh/sentence_similarity/blob/main/Finding_Similar_Sentence.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "import numpy as np\n",
    "from sentence_transformers import SentenceTransformer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the model\n",
    "Uncomment to store the model locally for easy retrieval, but delete the model before uploading to GitHub as too large"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = SentenceTransformer('nli-distilroberta-base-v2')\n",
    "# model.save(\"./Model/model\")\n",
    "# model = SentenceTransformer(\"./Model/model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load in the questions and responses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "Data = pd.read_csv(\"./Results/Compare.csv\", index_col=0)\n",
    "questions = Data.index.values\n",
    "company = Data['Company'].values\n",
    "gold = Data['Meta'].values\n",
    "sentences = np.concatenate((company, gold))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the sentence embeddings and obtain scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "sentence_embeddings = model.encode(sentences)\n",
    "similarity_score = []\n",
    "for i in range(len(company)):\n",
    "    similarity_score.append(cosine_similarity(\n",
    "        [sentence_embeddings[i]],\n",
    "        [sentence_embeddings[len(company) + i]]\n",
    "    ).flatten()[0])\n",
    "Similarity = pd.DataFrame({'Score': similarity_score}, index=questions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Print a few of the scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                               Score\n",
      "What is the file name of the document?      0.858884\n",
      "When was the document last modified?        0.565336\n",
      "What is the file path of the document?      0.755018\n",
      "When was the document last accessed?        0.553015\n",
      "What is the creation date of the document?  0.849555\n"
     ]
    }
   ],
   "source": [
    "print(Similarity.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Save the scores to a .csv file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "Similarity.to_csv('./Results/Similarity_Scores.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "docu_compare",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}