{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9e975979-3b3d-4a8d-9db6-b7433cf0d8b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os, random, json\n",
    "import sqlite3\n",
    "\n",
    "import pandas as pd\n",
    "from openai import OpenAI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "98601634-bd9b-4566-b242-2b3c9d04b260",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "63db76a8-31de-4957-b7b9-291c2539f976",
   "metadata": {},
   "source": [
    "### Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ff4d40aa-a42e-4ad7-9ca9-d894653d205e",
   "metadata": {},
   "outputs": [],
   "source": [
    "db_path = \"../database/mock_qna.sqlite\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "98a20c7e-b1dc-42d5-929b-62978959abda",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a11295d9-9bf0-4c9d-b5b2-0feec01bf640",
   "metadata": {},
   "outputs": [],
   "source": [
    "con = sqlite3.connect(db_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a1c1e976-0d75-42e3-8c2e-5045ee0f2c4a",
   "metadata": {},
   "outputs": [],
   "source": [
    "cur = con.cursor()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d78b0cc7-0238-41be-bc9f-688fcac71f73",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = cur.execute(f\"\"\"SELECT COUNT(*)\n",
    "                      FROM qna_tbl\n",
    "                   \"\"\")\n",
    "table_size = res.fetchone()[0]\n",
    "print(f\"table size: {table_size}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "faaacff0-bc67-464d-bd7c-1d51b0901dd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = cur.execute(f\"\"\"SELECT chapter, COUNT(*)\n",
    "                      FROM qna_tbl\n",
    "                      GROUP BY chapter\n",
    "                   \"\"\")\n",
    "chapter_counts = res.fetchall()\n",
    "print(chapter_counts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f83954ba-f92a-42ce-8d1c-758f4054b4c5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "117bbc79-5f58-4b31-9df1-dac75d7ef5a8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8dae73ca-845a-4d1e-8e1f-b1efb36dec8e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c4fddf3-6e7a-40c7-a6c2-2e06f976ec56",
   "metadata": {},
   "outputs": [],
   "source": [
    "id = random.randint(1, table_size)\n",
    "res = cur.execute(f\"\"\"SELECT question, option_1, option_2, option_3, option_4, correct_answer\n",
    "                      FROM qna_tbl\n",
    "                      WHERE id={id}\n",
    "                   \"\"\")\n",
    "result = res.fetchone()\n",
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f55b4a21-45b1-42a6-8ad1-352174b78806",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5ef430b-807c-4090-8ed2-969c43ba228e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_qna_question(chapter_n):\n",
    "    sql_string = f\"\"\"SELECT id, question, option_1, option_2, option_3, option_4, correct_answer\n",
    "                     FROM qna_tbl\n",
    "                     WHERE chapter='{chapter_n}'\n",
    "                  \"\"\"\n",
    "    res = cur.execute(sql_string)\n",
    "    result = res.fetchone()\n",
    "\n",
    "    id       = result[0]\n",
    "    question = result[1]\n",
    "    option_1 = result[2]\n",
    "    option_2 = result[3]\n",
    "    option_3 = result[4]\n",
    "    option_4 = result[5]\n",
    "    c_answer = result[6]\n",
    "\n",
    "    qna_str  = \"Question: \\n\" + \\\n",
    "               \"========= \\n\" + \\\n",
    "                question.replace(\"\\\\n\", \"\\n\") + \"\\n\" + \\\n",
    "               \"A) \" + option_1 + \"\\n\" + \\\n",
    "               \"B) \" + option_2 + \"\\n\" + \\\n",
    "               \"C) \" + option_3 + \"\\n\" + \\\n",
    "               \"D) \" + option_4\n",
    "    \n",
    "    return id, qna_str, c_answer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b61cc8eb-5118-438a-b38f-e01fc92c7387",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "13702036-6457-464d-bd32-0e20dd7050e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "qna_custom_functions = [\n",
    "    {\n",
    "        \"name\": \"get_qna_question\",\n",
    "        \"description\": \"\"\"\n",
    "                        Extract the chapter information from the body of the input text, the format looks as follow:\n",
    "                        The output should be in the format with `Chapter_` as prefix.\n",
    "                        Example 1: `Chapter_1` for first chapter\n",
    "                        Example 2: For chapter 12 of the textbook, you should return `Chapter_12`\n",
    "                        Example 3: `Chapter_5` for fifth chapter\n",
    "                        Thereafter, the chapter_n argument will be passed to the function for Q&A question retrieval.\n",
    "                       \"\"\",\n",
    "        \"parameters\": {\n",
    "            \"type\": \"object\",\n",
    "            \"properties\": {\n",
    "                \"chapter_n\": {\n",
    "                    \"type\": \"string\",\n",
    "                    \"description\": \"\"\"\n",
    "                        which chapter to extract, the format of this function argumet is with `Chapter_` as prefix, \n",
    "                        concatenated with chapter number in integer. For example, `Chapter_2`, `Chapter_10`.\n",
    "                    \"\"\"\n",
    "                }\n",
    "            }\n",
    "        }\n",
    "    }\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1bbb95af-dd82-443f-b23c-97c9a2777e11",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "957fe647-c1f7-4db5-8f31-fb5e1f546c0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "client = OpenAI()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "018fc414-d6df-408f-a14c-0a3857f4c52d",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"I am interested in chapter 13, can you test my understanding of this chapter?\"\n",
    "response = client.chat.completions.create(\n",
    "    model = 'gpt-3.5-turbo',\n",
    "    messages = [{'role': 'user', 'content': prompt}],\n",
    "    functions = qna_custom_functions,\n",
    "    function_call = 'auto'\n",
    ")\n",
    "json_response = json.loads(response.choices[0].message.function_call.arguments)\n",
    "print(json_response)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2408c546-335c-478a-b1ea-9c0921a9b7a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37ec1b9a-2cdd-4838-ab02-8260d392483f",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"I am interested in chapter thirteen, can you test my understanding of this chapter?\"\n",
    "response = client.chat.completions.create(\n",
    "    model = 'gpt-3.5-turbo',\n",
    "    messages = [{'role': 'user', 'content': prompt}],\n",
    "    functions = qna_custom_functions,\n",
    "    function_call = 'auto'\n",
    ")\n",
    "json_response = json.loads(response.choices[0].message.function_call.arguments)\n",
    "print(json_response)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6b8e9f05-bb9a-429b-a1fb-abbaced23230",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "18edebdd-2c7f-4589-8909-f816be5c4d1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"I am interested in 4th chapter, can you test my understanding of this chapter?\"\n",
    "response = client.chat.completions.create(\n",
    "    model = 'gpt-3.5-turbo',\n",
    "    messages = [{'role': 'user', 'content': prompt}],\n",
    "    functions = qna_custom_functions,\n",
    "    function_call = 'auto'\n",
    ")\n",
    "json_response = json.loads(response.choices[0].message.function_call.arguments)\n",
    "print(json_response)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4325b3c-47d6-4d3f-a50a-45914b47a9c0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c558b722-4438-4485-98c0-b4117bc3d46e",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"\"\"There are 15 chapters in the Health Insurance text book, I want to study the last chapter, \n",
    "            can you test my understanding of this chapter?\n",
    "        \"\"\"\n",
    "response = client.chat.completions.create(\n",
    "    model = 'gpt-3.5-turbo',\n",
    "    messages = [{'role': 'user', 'content': prompt}],\n",
    "    functions = qna_custom_functions,\n",
    "    function_call = 'auto'\n",
    ")\n",
    "json_response = json.loads(response.choices[0].message.function_call.arguments)\n",
    "print(json_response)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "049a28bf-abe5-4247-970f-615d1877a2c0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "de49c61a-0b3e-4623-abcb-a7625ac4d0db",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"I am interested in 2nd chapter, can you test my understanding of this chapter?\"\n",
    "response = client.chat.completions.create(\n",
    "    model = 'gpt-3.5-turbo',\n",
    "    messages = [{'role': 'user', 'content': prompt}],\n",
    "    functions = qna_custom_functions,\n",
    "    function_call = 'auto'\n",
    ")\n",
    "json_response = json.loads(response.choices[0].message.function_call.arguments)\n",
    "print(json_response)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "074229dc-82d9-4a2b-9a08-019228da78a1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "289fba25-f547-402a-bd13-0dc4ce7ddf8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "id, qna_str, answer = get_qna_question(chapter_n=json_response[\"chapter_n\"])\n",
    "print(qna_str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "adc9f539-3654-4174-815b-e0939f513a20",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5b6ad929-e6a5-4978-8678-519375ef62eb",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}