notebooks/qa_system.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary libraries\n",
    "import os\n",
    "import re\n",
    "import json\n",
    "import logging\n",
    "import warnings\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import matplotlib.pyplot as plt\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sentence_transformers import SentenceTransformer, util, losses, InputExample\n",
    "from sentence_transformers.evaluation import InformationRetrievalEvaluator\n",
    "from bayes_opt import BayesianOptimization\n",
    "from scipy.stats import friedmanchisquare, levene\n",
    "import scikit_posthocs as sp\n",
    "from tqdm.autonotebook import tqdm, trange\n",
    "from langchain_community.llms import Ollama\n",
    "from langchain_core.prompts import PromptTemplate\n",
    "\n",
    "# Set plotting style to cyberpunk\n",
    "import mplcyberpunk\n",
    "plt.style.use(\"cyberpunk\")\n",
    "\n",
    "# Suppress the specific FutureWarning from huggingface_hub\n",
    "warnings.filterwarnings(\"ignore\", category=FutureWarning, module=\"huggingface_hub.file_download\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Manufacturer</th>\n",
       "      <th>Probe_Model</th>\n",
       "      <th>Cartridge_Connection</th>\n",
       "      <th>Compatible_Systems</th>\n",
       "      <th>Array_Type</th>\n",
       "      <th>Applications</th>\n",
       "      <th>Min_Frequency</th>\n",
       "      <th>Max_Frequency</th>\n",
       "      <th>Stock</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ATL</td>\n",
       "      <td>C3</td>\n",
       "      <td>0</td>\n",
       "      <td>[HDI 5000]</td>\n",
       "      <td>convex</td>\n",
       "      <td>[abdominal, general]</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ATL</td>\n",
       "      <td>C4-2</td>\n",
       "      <td>0</td>\n",
       "      <td>[HDI 5000]</td>\n",
       "      <td>convex</td>\n",
       "      <td>[abdominal, obstetric, gynecologic, general]</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ATL</td>\n",
       "      <td>C5-2</td>\n",
       "      <td>0</td>\n",
       "      <td>[HDI 1500, HDI 3000, HDI 3500, HDI 5000]</td>\n",
       "      <td>convex</td>\n",
       "      <td>[abdominal, general]</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ATL</td>\n",
       "      <td>C5-IVT</td>\n",
       "      <td>0</td>\n",
       "      <td>[UM9 HDI, HDI 1500, HDI 3000, HDI 3500, HDI 5000]</td>\n",
       "      <td>convex</td>\n",
       "      <td>[intracavitary, endovaginal, gynecological]</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ATL</td>\n",
       "      <td>C7-4</td>\n",
       "      <td>0</td>\n",
       "      <td>[UM9 HDI, HDI 1500, HDI 3000, HDI 5000]</td>\n",
       "      <td>convex</td>\n",
       "      <td>[cardiac, vascular, general]</td>\n",
       "      <td>4.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Manufacturer Probe_Model  Cartridge_Connection  \\\n",
       "0          ATL          C3                     0   \n",
       "1          ATL        C4-2                     0   \n",
       "2          ATL        C5-2                     0   \n",
       "3          ATL      C5-IVT                     0   \n",
       "4          ATL        C7-4                     0   \n",
       "\n",
       "                                  Compatible_Systems Array_Type  \\\n",
       "0                                         [HDI 5000]     convex   \n",
       "1                                         [HDI 5000]     convex   \n",
       "2           [HDI 1500, HDI 3000, HDI 3500, HDI 5000]     convex   \n",
       "3  [UM9 HDI, HDI 1500, HDI 3000, HDI 3500, HDI 5000]     convex   \n",
       "4            [UM9 HDI, HDI 1500, HDI 3000, HDI 5000]     convex   \n",
       "\n",
       "                                   Applications  Min_Frequency  Max_Frequency  \\\n",
       "0                          [abdominal, general]            3.0            3.0   \n",
       "1  [abdominal, obstetric, gynecologic, general]            2.0            4.0   \n",
       "2                          [abdominal, general]            2.0            5.0   \n",
       "3   [intracavitary, endovaginal, gynecological]            4.0            6.0   \n",
       "4                  [cardiac, vascular, general]            4.0            7.0   \n",
       "\n",
       "   Stock  \n",
       "0      0  \n",
       "1      2  \n",
       "2      7  \n",
       "3      0  \n",
       "4      0  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load the probes dataset\n",
    "probes_df = pd.read_csv(\"../data/probes.csv\")\n",
    "probes_df = probes_df.drop(['Description'], axis=1)\n",
    "\n",
    "# Split \"Compatible_Systems\" column\n",
    "probes_df['Compatible_Systems'] = probes_df['Compatible_Systems'].str.split(', ')\n",
    "\n",
    "# Split \"Applications\" column\n",
    "probes_df['Applications'] = probes_df['Applications'].str.split(', ')\n",
    "\n",
    "# Load the systems dataset\n",
    "systems_df = pd.read_csv(\"../data/systems.csv\")\n",
    "\n",
    "# Split \"compatible_probes\" column\n",
    "systems_df['compatible_probes'] = systems_df['compatible_probes'].str.split(', ')\n",
    "\n",
    "# Print the head of the dataframes\n",
    "probes_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ultrasound_system</th>\n",
       "      <th>manufacturer</th>\n",
       "      <th>compatible_probes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HDI 5000</td>\n",
       "      <td>ATL</td>\n",
       "      <td>[C3, C4-2, C5-2, C5-IVT, C7-4, C8-4v, C9-5ICT,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HDI 1500</td>\n",
       "      <td>ATL</td>\n",
       "      <td>[C5-2, C5-IVT, C7-4, C9-5ICT, L7-4]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HDI 3000</td>\n",
       "      <td>ATL</td>\n",
       "      <td>[C5-2, C5-IVT, C7-4, C8-4v, CL 10-5, CT8-4, L7-4]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HDI 3500</td>\n",
       "      <td>ATL</td>\n",
       "      <td>[C5-2, C5-IVT, C8-4v, CT8-4]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>UM9 HDI</td>\n",
       "      <td>ATL</td>\n",
       "      <td>[C5-IVT, C7-4, L7-4]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  ultrasound_system manufacturer  \\\n",
       "0          HDI 5000          ATL   \n",
       "1          HDI 1500          ATL   \n",
       "2          HDI 3000          ATL   \n",
       "3          HDI 3500          ATL   \n",
       "4           UM9 HDI          ATL   \n",
       "\n",
       "                                   compatible_probes  \n",
       "0  [C3, C4-2, C5-2, C5-IVT, C7-4, C8-4v, C9-5ICT,...  \n",
       "1                [C5-2, C5-IVT, C7-4, C9-5ICT, L7-4]  \n",
       "2  [C5-2, C5-IVT, C7-4, C8-4v, CL 10-5, CT8-4, L7-4]  \n",
       "3                       [C5-2, C5-IVT, C8-4v, CT8-4]  \n",
       "4                               [C5-IVT, C7-4, L7-4]  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "systems_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating document corpus\n",
    "\n",
    "Here we convert our tabular data into string sentences and generate tags that correspond to the information it contains."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>tags</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The manufacturer of the C3 probe is ATL.</td>\n",
       "      <td>(probes, ATL, C3, manufacturer)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>The ATL C3 probe is compatible with the follow...</td>\n",
       "      <td>(compatibility, probes, HDI 5000, C3, ATL)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The ATL C3 is a convex array type probe.</td>\n",
       "      <td>(probes, array type, ATL, C3)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The ATL C3 probe is suitable for the following...</td>\n",
       "      <td>(abdominal, general, probes, applications, C3,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>The ATL C3 probe is currently out of stock or ...</td>\n",
       "      <td>(probes, ATL, C3, stock)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             content  \\\n",
       "0           The manufacturer of the C3 probe is ATL.   \n",
       "1  The ATL C3 probe is compatible with the follow...   \n",
       "2           The ATL C3 is a convex array type probe.   \n",
       "3  The ATL C3 probe is suitable for the following...   \n",
       "4  The ATL C3 probe is currently out of stock or ...   \n",
       "\n",
       "                                                tags  \n",
       "0                    (probes, ATL, C3, manufacturer)  \n",
       "1         (compatibility, probes, HDI 5000, C3, ATL)  \n",
       "2                      (probes, array type, ATL, C3)  \n",
       "3  (abdominal, general, probes, applications, C3,...  \n",
       "4                           (probes, ATL, C3, stock)  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Function to generate probe descriptions and tags\n",
    "def generate_probe_info(row):\n",
    "    \"\"\"\n",
    "    This function takes a row from the probes_df DataFrame and generates lists of probe descriptions and tags.\n",
    "    \"\"\"\n",
    "    descriptions = []\n",
    "    tags = []\n",
    "\n",
    "    # Manufacturer\n",
    "    descriptions.append(f\"The manufacturer of the {row['Probe_Model']} probe is {row['Manufacturer']}.\")\n",
    "    tags.append(frozenset(['probes', 'manufacturer', row['Manufacturer'], row['Probe_Model']]))\n",
    "\n",
    "    # Cartridge Connection\n",
    "    if row['Cartridge_Connection'] == 1:\n",
    "        descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} probe has a variant with a cartridge connection.\")\n",
    "        tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], 'connection']))\n",
    "\n",
    "    # Compatible Systems\n",
    "    compatible_systems = ', '.join(row['Compatible_Systems'])\n",
    "    descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} probe is compatible with the following systems: {compatible_systems}.\")\n",
    "    tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], 'compatibility', compatible_systems]))\n",
    "\n",
    "    # Array Type\n",
    "    descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} is a {row['Array_Type']} array type probe.\")\n",
    "    tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], 'array type']))\n",
    "\n",
    "    # Applications\n",
    "    applications = ', '.join(row['Applications'])\n",
    "    descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} probe is suitable for the following applications: {applications}.\")\n",
    "    tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], \"applications\", applications]))\n",
    "\n",
    "    # Stock\n",
    "    if row['Stock'] > 0:\n",
    "        descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} probe is currently in stock and available for sale.\")\n",
    "    else:\n",
    "        descriptions.append(f\"The {row['Manufacturer']} {row['Probe_Model']} probe is currently out of stock or not available for sale.\")\n",
    "    tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], 'stock']))\n",
    "\n",
    "    # Frequency Range\n",
    "    descriptions.append(f\"The frequency range of the {row['Manufacturer']} {row['Probe_Model']} probe is {row['Min_Frequency']} - {row['Max_Frequency']} MHz.\")\n",
    "    tags.append(frozenset(['probes', row['Manufacturer'], row['Probe_Model'], 'frequency']))\n",
    "\n",
    "    return descriptions, tags\n",
    "\n",
    "# Function to generate system descriptions and tags\n",
    "def generate_system_description(row):\n",
    "    \"\"\"\n",
    "    This function takes a row from the systems_df DataFrame and generates lists of system descriptions and tags.\n",
    "    \"\"\"\n",
    "    descriptions = []\n",
    "    tags = []\n",
    "\n",
    "    # System manufacturer\n",
    "    description = f\"The {row['ultrasound_system']} ultrasound system is made by {row['manufacturer']}.\"\n",
    "    descriptions.append(description)\n",
    "    tags.append(frozenset([row['manufacturer'], row['ultrasound_system'], 'manufacturer', 'systems']))\n",
    "\n",
    "    # System compatible probes\n",
    "    compatible_probes = ', '.join(row['compatible_probes'])\n",
    "    description = f\"The {row['manufacturer']} {row['ultrasound_system']} ultrasound system is compatible with the following {row['manufacturer']} probes: {compatible_probes}.\"\n",
    "    descriptions.append(description)\n",
    "    tags.append(frozenset([row['manufacturer'], row['ultrasound_system'], 'compatibility', 'systems', compatible_probes]))\n",
    "\n",
    "    return descriptions, tags\n",
    "\n",
    "# Apply the probe function and directly unpack into separate lists\n",
    "probe_descriptions, probe_tags = zip(*probes_df.apply(generate_probe_info, axis=1))\n",
    "\n",
    "# Apply the system function and directly unpack into separate lists\n",
    "system_descriptions, system_tags = zip(*systems_df.apply(generate_system_description, axis=1))\n",
    "\n",
    "# Flatten the lists\n",
    "all_contents = [item for sublist in (probe_descriptions + system_descriptions) for item in sublist]\n",
    "all_tags = [item for sublist in (probe_tags + system_tags) for item in sublist]\n",
    "\n",
    "# Create a DataFrame from these lists\n",
    "documents = pd.DataFrame({\n",
    "    'content': all_contents,\n",
    "    'tags': all_tags\n",
    "})\n",
    "\n",
    "documents.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating queries for the documents using Ollama (llama3)\n",
    "\n",
    "We define a prompt and include examples of what questions to ask for each document type. We also instruct the prompt to structure its output using numbers to separate questions so each question can be extracted using regex. \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the Ollama LLM\n",
    "llm = Ollama(model=\"llama3\")\n",
    "\n",
    "prompt_template = PromptTemplate(\n",
    "    input_variables=[\"document\"],\n",
    "    template=\"\"\"\n",
    "Generate three concise questions that can be answered using the following information, similar to the example provided. Provide only the questions, numbered as follows:\n",
    "\n",
    "1. [insert question here]\n",
    "2. [insert question here]\n",
    "3. [insert question here]\n",
    "\n",
    "Example 1:\n",
    "Information: The manufacturer of the C3 probe is ATL.\n",
    "Questions:\n",
    "1. Who is the manufacturer of the C3 probe?\n",
    "2. Who makes the C3 transducer?\n",
    "3. Is the C3 probe made by ATL?\n",
    "\n",
    "Example 2:\n",
    "Information: The G.E. RIC5-9D probe is compatible with the following systems: Voluson E6, Voluson E8, Voluson E10, LOGIQ S7, LOGIQ S8, Vivid E95, LOGIQ E9, LOGIQ E10.\n",
    "Questions:\n",
    "1. What systems are compatible with the G.E. RIC5-9D probe?\n",
    "2. Does the G.E. RIC5-9D work with the Voluson E6 system?\n",
    "3. Is the G.E. RIC5-9D transducer compatible with the LOGIQ S7 system?\n",
    "\n",
    "Example 3:\n",
    "Information: The Siemens Acuson 15L8W probe has a variant with a cartridge connection.\n",
    "Questions:\n",
    "1. Does the Siemens Acuson 15L8W probe have a cartridge connection?\n",
    "2. What kind of connector does the Siemens Acuson 15L8W transducer use?\n",
    "3. Does the Siemens Acuson 15L8W probe use a cartridge connector?\n",
    "\n",
    "Example 4:\n",
    "Information: The Philips C8-4v is a convex array type probe.\n",
    "Questions:\n",
    "1. What array type is the Philips C8-4v?\n",
    "2. What type of transducer is the Philips C8-4v?\n",
    "3. Is the Philips C8-4v a convex array probe?\n",
    "\n",
    "Example 5:\n",
    "Information: The ATL P6-3 probe is suitable for the following applications: cardiac, vascular, general.\n",
    "Questions:\n",
    "1. What can the ATL P6-3 probe be used for?\n",
    "2. What are the applications of the ATL P6-3 probe?\n",
    "3. Can the P6-3 probe be used for cardiac applications?\n",
    "\n",
    "Example 6:\n",
    "Information: The G.E. RAB2-5L probe is currently in stock and available for sale.\n",
    "Questions:\n",
    "1. Do you have any RAB2-5L probes by G.E. for sale?\n",
    "2. Is the G.E. RAB2-5L probe in stock?\n",
    "3. Are there any RAB2-5L transducers for sale?\n",
    "\n",
    "Example 7:\n",
    "Information: The frequency range of the ATL L10-5 probe is 5.0 - 10.0 MHz.\n",
    "Questions:\n",
    "1. What is the frequency range of the ATL L10-5 probe?\n",
    "2. What is the maximum frequency range of the ATL L10-5 transducer?\n",
    "3. What is the minimum frequency range of the ATL L10-5 probe?\n",
    "\n",
    "Information:\n",
    "{document}\n",
    "\"\"\"\n",
    ")\n",
    "\n",
    "def extract_questions(response):\n",
    "    pattern = r'\\d+\\.\\s(.+?\\?)'\n",
    "    questions = re.findall(pattern, response)\n",
    "    return questions\n",
    "\n",
    "def generate_questions(text_chunk):\n",
    "    prompt = prompt_template.format(document=text_chunk)\n",
    "    output = llm(prompt)\n",
    "    questions = extract_questions(output)\n",
    "    return questions\n",
    "\n",
    "# Generate questions for each document\n",
    "tqdm.pandas(desc=\"Generating questions\")\n",
    "documents[\"questions\"] = documents[\"content\"].progress_apply(generate_questions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>tags</th>\n",
       "      <th>question</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The manufacturer of the C3 probe is ATL.</td>\n",
       "      <td>[probes, manufacturer, ATL, C3]</td>\n",
       "      <td>Who is the manufacturer of the C3 probe?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>The manufacturer of the C3 probe is ATL.</td>\n",
       "      <td>[probes, manufacturer, ATL, C3]</td>\n",
       "      <td>Is the manufacturer of the C3 probe ATL?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The manufacturer of the C3 probe is ATL.</td>\n",
       "      <td>[probes, manufacturer, ATL, C3]</td>\n",
       "      <td>Who makes the C3 probe?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The ATL C3 probe is compatible with the follow...</td>\n",
       "      <td>[probes, ATL, C3, compatibility, HDI 5000]</td>\n",
       "      <td>What systems is the ATL C3 probe compatible with?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>The ATL C3 probe is compatible with the follow...</td>\n",
       "      <td>[probes, ATL, C3, compatibility, HDI 5000]</td>\n",
       "      <td>Does the ATL C3 probe work with the HDI 5000 s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1003</th>\n",
       "      <td>The Cypress Portable ultrasound system is made...</td>\n",
       "      <td>[Siemens Acuson, Cypress Portable, manufacture...</td>\n",
       "      <td>Is the Cypress Portable system a product of Si...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1004</th>\n",
       "      <td>The Cypress Portable ultrasound system is made...</td>\n",
       "      <td>[Siemens Acuson, Cypress Portable, manufacture...</td>\n",
       "      <td>What company manufactures the Cypress Portable...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1005</th>\n",
       "      <td>The Siemens Acuson Cypress Portable ultrasound...</td>\n",
       "      <td>[Siemens Acuson, Cypress Portable, compatibili...</td>\n",
       "      <td>Which Siemens Acuson probes are compatible wit...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1006</th>\n",
       "      <td>The Siemens Acuson Cypress Portable ultrasound...</td>\n",
       "      <td>[Siemens Acuson, Cypress Portable, compatibili...</td>\n",
       "      <td>Is the 7V3c probe compatible with the Siemens ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1007</th>\n",
       "      <td>The Siemens Acuson Cypress Portable ultrasound...</td>\n",
       "      <td>[Siemens Acuson, Cypress Portable, compatibili...</td>\n",
       "      <td>What probes work with the Siemens Acuson Cypre...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1008 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content  \\\n",
       "0              The manufacturer of the C3 probe is ATL.   \n",
       "1              The manufacturer of the C3 probe is ATL.   \n",
       "2              The manufacturer of the C3 probe is ATL.   \n",
       "3     The ATL C3 probe is compatible with the follow...   \n",
       "4     The ATL C3 probe is compatible with the follow...   \n",
       "...                                                 ...   \n",
       "1003  The Cypress Portable ultrasound system is made...   \n",
       "1004  The Cypress Portable ultrasound system is made...   \n",
       "1005  The Siemens Acuson Cypress Portable ultrasound...   \n",
       "1006  The Siemens Acuson Cypress Portable ultrasound...   \n",
       "1007  The Siemens Acuson Cypress Portable ultrasound...   \n",
       "\n",
       "                                                   tags  \\\n",
       "0                       [probes, manufacturer, ATL, C3]   \n",
       "1                       [probes, manufacturer, ATL, C3]   \n",
       "2                       [probes, manufacturer, ATL, C3]   \n",
       "3            [probes, ATL, C3, compatibility, HDI 5000]   \n",
       "4            [probes, ATL, C3, compatibility, HDI 5000]   \n",
       "...                                                 ...   \n",
       "1003  [Siemens Acuson, Cypress Portable, manufacture...   \n",
       "1004  [Siemens Acuson, Cypress Portable, manufacture...   \n",
       "1005  [Siemens Acuson, Cypress Portable, compatibili...   \n",
       "1006  [Siemens Acuson, Cypress Portable, compatibili...   \n",
       "1007  [Siemens Acuson, Cypress Portable, compatibili...   \n",
       "\n",
       "                                               question  \n",
       "0              Who is the manufacturer of the C3 probe?  \n",
       "1              Is the manufacturer of the C3 probe ATL?  \n",
       "2                               Who makes the C3 probe?  \n",
       "3     What systems is the ATL C3 probe compatible with?  \n",
       "4     Does the ATL C3 probe work with the HDI 5000 s...  \n",
       "...                                                 ...  \n",
       "1003  Is the Cypress Portable system a product of Si...  \n",
       "1004  What company manufactures the Cypress Portable...  \n",
       "1005  Which Siemens Acuson probes are compatible wit...  \n",
       "1006  Is the 7V3c probe compatible with the Siemens ...  \n",
       "1007  What probes work with the Siemens Acuson Cypre...  \n",
       "\n",
       "[1008 rows x 3 columns]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a new DataFrame by exploding the 'questions' column\n",
    "questions_df = documents.explode('questions')\n",
    "questions_df = questions_df.reset_index(drop=True)\n",
    "questions_df = questions_df.rename(columns={'questions': 'question'})\n",
    "questions_df.to_csv(\"../data/questions_docs_tags.csv\")\n",
    "questions_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Filtering out the questions that are not relevant using llama3 again\n",
    "\n",
    "Any document-query pairs that are marked irrelevant are exported, manually reviewed, and re-scored if neccessary before being joined back into the dataset.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7d20f9e842054b55a4cd66b213b51649",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Scoring Questions:   0%|          | 0/1008 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                             content  \\\n",
      "0           The manufacturer of the C3 probe is ATL.   \n",
      "1           The manufacturer of the C3 probe is ATL.   \n",
      "2           The manufacturer of the C3 probe is ATL.   \n",
      "3  The ATL C3 probe is compatible with the follow...   \n",
      "4  The ATL C3 probe is compatible with the follow...   \n",
      "\n",
      "                                         tags  \\\n",
      "0             (ATL, manufacturer, C3, probes)   \n",
      "1             (ATL, manufacturer, C3, probes)   \n",
      "2             (ATL, manufacturer, C3, probes)   \n",
      "3  (HDI 5000, ATL, probes, compatibility, C3)   \n",
      "4  (HDI 5000, ATL, probes, compatibility, C3)   \n",
      "\n",
      "                                            question  score  \n",
      "0           Who is the manufacturer of the C3 probe?      1  \n",
      "1           Is the manufacturer of the C3 probe ATL?      1  \n",
      "2                            Who makes the C3 probe?      1  \n",
      "3  What systems is the ATL C3 probe compatible with?      1  \n",
      "4  Does the ATL C3 probe work with the HDI 5000 s...      1  \n"
     ]
    }
   ],
   "source": [
    "# Define the prompt template for grading questions\n",
    "grading_prompt_template = PromptTemplate(\n",
    "    input_variables=[\"content\", \"question\"],\n",
    "    template=\"\"\"\n",
    "    Given the content: \"{content}\"\n",
    "    Can the following question be sufficiently answered?\n",
    "    Question: \"{question}\"\n",
    "    Your response should only consist of one number, either a 0 (meaning No) or 1 (meaning Yes).\n",
    "    \n",
    "    Example 1:\n",
    "    Content: \"The manufacturer of the C3 probe is ATL.\"\n",
    "    Question: \"Who is the manufacturer of the C3 probe?\"\n",
    "    Response: 1\n",
    "\n",
    "    Example 2:\n",
    "    Content: \"Who manufactures the EPIQ 7 ultrasound system?\"\n",
    "    Question: \"The Siemens Acuson 12L3 probe is compatible with the following systems: Juniper.\"\n",
    "    Response: 0\n",
    "    \"\"\"\n",
    ")\n",
    "\n",
    "# Set up logging\n",
    "logging.basicConfig(level=logging.INFO, format='%(levelname)s: %(message)s')\n",
    "\n",
    "# Precompile the regular expression for efficiency\n",
    "score_pattern = re.compile(r'\\b[01]\\b')\n",
    "\n",
    "def grade_question(content, question):\n",
    "    prompt = grading_prompt_template.format(content=content, question=question)\n",
    "    response = llm(prompt)\n",
    "    # Use the precompiled regular expression to find numbers in the response\n",
    "    matches = score_pattern.findall(response)\n",
    "    if matches:\n",
    "        # Assuming the first match is the relevant score\n",
    "        return int(matches[0])\n",
    "    else:\n",
    "        # Log the issue and return a default value\n",
    "        logging.warning(f\"No valid score found in response: {response}\")\n",
    "        return None  # or set a default value, or handle it in another way\n",
    "\n",
    "\n",
    "# Apply the grade_question function to each row to create a new 'score' column\n",
    "tqdm.pandas(desc=\"Scoring Questions\")\n",
    "questions_df['score'] = questions_df.progress_apply(lambda row: grade_question(row['content'], row['question']), axis=1)\n",
    "print(questions_df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>tags</th>\n",
       "      <th>question</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>The ATL C3 probe is suitable for the following...</td>\n",
       "      <td>(abdominal, general, ATL, probes, applications...</td>\n",
       "      <td>Are there any other uses for the ATL C3 transd...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>The ATL C3 probe is currently out of stock or ...</td>\n",
       "      <td>(ATL, C3, stock, probes)</td>\n",
       "      <td>Is the ATL C3 probe currently in stock?</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>The ATL C3 probe is currently out of stock or ...</td>\n",
       "      <td>(ATL, C3, stock, probes)</td>\n",
       "      <td>Can you still purchase the ATL C3 probe?</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>The ATL C3 probe is currently out of stock or ...</td>\n",
       "      <td>(ATL, C3, stock, probes)</td>\n",
       "      <td>Are there any ATL C3 probes available for sale?</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>The ATL C4-2 probe is compatible with the foll...</td>\n",
       "      <td>(HDI 5000, C4-2, ATL, probes, compatibility)</td>\n",
       "      <td>Is the ATL C4-2 probe only used with HDI 5000 ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>881</th>\n",
       "      <td>The G.E. Vivid E95 ultrasound system is compat...</td>\n",
       "      <td>(systems, G.E., Vivid E95, compatibility, RIC5...</td>\n",
       "      <td>Which G.E. probe is compatible with both the L...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>896</th>\n",
       "      <td>The EPIQ 5 ultrasound system is made by Philips.</td>\n",
       "      <td>(EPIQ 5, systems, Philips, manufacturer)</td>\n",
       "      <td>What company manufactures the EPIQ 5 transducer?</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>928</th>\n",
       "      <td>The Philips iE33 ultrasound system is compatib...</td>\n",
       "      <td>(C5-1, C8-5, L12-3, S4-1, S4-2, systems, compa...</td>\n",
       "      <td>Are the C5-1, C8-5, L12-3, S4-1, and S4-2 prob...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>988</th>\n",
       "      <td>The Siemens Acuson Juniper ultrasound system i...</td>\n",
       "      <td>(systems, Siemens Acuson, compatibility, Junip...</td>\n",
       "      <td>Are all Siemens Acuson probes compatible with ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1001</th>\n",
       "      <td>The Siemens Acuson Sequoia 512 ultrasound syst...</td>\n",
       "      <td>(systems, Siemens Acuson, compatibility, Sequo...</td>\n",
       "      <td>Does the Siemens Acuson Sequoia 512 system sup...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>95 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content  \\\n",
       "11    The ATL C3 probe is suitable for the following...   \n",
       "12    The ATL C3 probe is currently out of stock or ...   \n",
       "13    The ATL C3 probe is currently out of stock or ...   \n",
       "14    The ATL C3 probe is currently out of stock or ...   \n",
       "22    The ATL C4-2 probe is compatible with the foll...   \n",
       "...                                                 ...   \n",
       "881   The G.E. Vivid E95 ultrasound system is compat...   \n",
       "896    The EPIQ 5 ultrasound system is made by Philips.   \n",
       "928   The Philips iE33 ultrasound system is compatib...   \n",
       "988   The Siemens Acuson Juniper ultrasound system i...   \n",
       "1001  The Siemens Acuson Sequoia 512 ultrasound syst...   \n",
       "\n",
       "                                                   tags  \\\n",
       "11    (abdominal, general, ATL, probes, applications...   \n",
       "12                             (ATL, C3, stock, probes)   \n",
       "13                             (ATL, C3, stock, probes)   \n",
       "14                             (ATL, C3, stock, probes)   \n",
       "22         (HDI 5000, C4-2, ATL, probes, compatibility)   \n",
       "...                                                 ...   \n",
       "881   (systems, G.E., Vivid E95, compatibility, RIC5...   \n",
       "896            (EPIQ 5, systems, Philips, manufacturer)   \n",
       "928   (C5-1, C8-5, L12-3, S4-1, S4-2, systems, compa...   \n",
       "988   (systems, Siemens Acuson, compatibility, Junip...   \n",
       "1001  (systems, Siemens Acuson, compatibility, Sequo...   \n",
       "\n",
       "                                               question  score  \n",
       "11    Are there any other uses for the ATL C3 transd...      0  \n",
       "12              Is the ATL C3 probe currently in stock?      0  \n",
       "13             Can you still purchase the ATL C3 probe?      0  \n",
       "14      Are there any ATL C3 probes available for sale?      0  \n",
       "22    Is the ATL C4-2 probe only used with HDI 5000 ...      0  \n",
       "...                                                 ...    ...  \n",
       "881   Which G.E. probe is compatible with both the L...      0  \n",
       "896    What company manufactures the EPIQ 5 transducer?      0  \n",
       "928   Are the C5-1, C8-5, L12-3, S4-1, and S4-2 prob...      0  \n",
       "988   Are all Siemens Acuson probes compatible with ...      0  \n",
       "1001  Does the Siemens Acuson Sequoia 512 system sup...      0  \n",
       "\n",
       "[95 rows x 4 columns]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_df = questions_df[questions_df['score'].isin([0, None])]\n",
    "filtered_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exporting for manual review"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "#filtered_df.to_csv(\"../data/filtered_questions_for_correction.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Re-importing manually reviewed pairs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#corrected_questions_df = pd.read_csv(\"../data/corrected_questions.csv\")\n",
    "\n",
    "# Drop the old rows from questions_df\n",
    "#fine_tuning_dataset = questions_df.drop(filtered_df.index)\n",
    "\n",
    "# Append the corrected rows to the original DataFrame\n",
    "#fine_tuning_dataset = pd.concat([fine_tuning_dataset, corrected_questions_df], ignore_index=True)\n",
    "#fine_tuning_dataset = fine_tuning_dataset[fine_tuning_dataset['score'] == 1].drop(columns=['score'])\n",
    "#fine_tuning_dataset\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate unique IDs for queries and documents based on their content\n",
    "#document_ids = {doc: idx for idx, doc in enumerate(fine_tuning_dataset['content'].unique())}\n",
    "#fine_tuning_dataset['doc_id'] = fine_tuning_dataset['content'].map(document_ids)\n",
    "#fine_tuning_dataset['query_id'] = range(len(fine_tuning_dataset))\n",
    "\n",
    "#fine_tuning_dataset.to_pickle('../data/fine_tuning_dataset.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Start here to import cleaned question-document dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question</th>\n",
       "      <th>content</th>\n",
       "      <th>labels</th>\n",
       "      <th>doc_id</th>\n",
       "      <th>query_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What is the frequency range of the G.E. RIC5-9...</td>\n",
       "      <td>The frequency range of the G.E. RIC5-9D probe ...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>What applications is the Philips C8-4v probe s...</td>\n",
       "      <td>The Philips C8-4v probe is suitable for the fo...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Is the Philips C5-2 probe currently available ...</td>\n",
       "      <td>The Philips C5-2 probe is currently in stock a...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Is the HDI 1500 ultrasound system made by ATL?</td>\n",
       "      <td>The HDI 1500 ultrasound system is made by ATL.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Is the ATL C4-2 probe used in general medicine?</td>\n",
       "      <td>The ATL C4-2 probe is suitable for the followi...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1916</th>\n",
       "      <td>Are there any Philips L17-5 probes available f...</td>\n",
       "      <td>The Philips L17-5 probe is currently in stock ...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>195</td>\n",
       "      <td>1916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1917</th>\n",
       "      <td>Which ultrasound machine is the G.E. 348c tran...</td>\n",
       "      <td>The G.E. 348c probe is compatible with the fol...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>313</td>\n",
       "      <td>1917</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1920</th>\n",
       "      <td>What can the ATL C7-4 probe be used for?</td>\n",
       "      <td>The ATL C7-4 probe is suitable for the followi...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>183</td>\n",
       "      <td>1920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1923</th>\n",
       "      <td>What kind of array is the ATL C8-4v?</td>\n",
       "      <td>The ATL C8-4v is a convex array type probe.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>279</td>\n",
       "      <td>1923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1924</th>\n",
       "      <td>Can you use a RIC5-9D probe on the Voluson E6 ...</td>\n",
       "      <td>The G.E. Voluson E6 ultrasound system is compa...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>303</td>\n",
       "      <td>1924</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>963 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               question  \\\n",
       "0     What is the frequency range of the G.E. RIC5-9...   \n",
       "4     What applications is the Philips C8-4v probe s...   \n",
       "5     Is the Philips C5-2 probe currently available ...   \n",
       "7        Is the HDI 1500 ultrasound system made by ATL?   \n",
       "11      Is the ATL C4-2 probe used in general medicine?   \n",
       "...                                                 ...   \n",
       "1916  Are there any Philips L17-5 probes available f...   \n",
       "1917  Which ultrasound machine is the G.E. 348c tran...   \n",
       "1920           What can the ATL C7-4 probe be used for?   \n",
       "1923               What kind of array is the ATL C8-4v?   \n",
       "1924  Can you use a RIC5-9D probe on the Voluson E6 ...   \n",
       "\n",
       "                                                content  labels  doc_id  \\\n",
       "0     The frequency range of the G.E. RIC5-9D probe ...     1.0       0   \n",
       "4     The Philips C8-4v probe is suitable for the fo...     1.0       2   \n",
       "5     The Philips C5-2 probe is currently in stock a...     1.0       3   \n",
       "7        The HDI 1500 ultrasound system is made by ATL.     1.0       4   \n",
       "11    The ATL C4-2 probe is suitable for the followi...     1.0       5   \n",
       "...                                                 ...     ...     ...   \n",
       "1916  The Philips L17-5 probe is currently in stock ...     1.0     195   \n",
       "1917  The G.E. 348c probe is compatible with the fol...     1.0     313   \n",
       "1920  The ATL C7-4 probe is suitable for the followi...     1.0     183   \n",
       "1923        The ATL C8-4v is a convex array type probe.     1.0     279   \n",
       "1924  The G.E. Voluson E6 ultrasound system is compa...     1.0     303   \n",
       "\n",
       "      query_id  \n",
       "0            0  \n",
       "4            4  \n",
       "5            5  \n",
       "7            7  \n",
       "11          11  \n",
       "...        ...  \n",
       "1916      1916  \n",
       "1917      1917  \n",
       "1920      1920  \n",
       "1923      1923  \n",
       "1924      1924  \n",
       "\n",
       "[963 rows x 5 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Loading from Pickle\n",
    "fine_tuning_dataset = pd.read_pickle('../data/fine_tuning_dataset.pkl')\n",
    "fine_tuning_dataset = fine_tuning_dataset[fine_tuning_dataset['labels'] == 1.0]\n",
    "fine_tuning_dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluating pre-trained embedding models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question</th>\n",
       "      <th>content</th>\n",
       "      <th>labels</th>\n",
       "      <th>doc_id</th>\n",
       "      <th>query_id</th>\n",
       "      <th>tags</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What is the frequency range of the G.E. RIC5-9...</td>\n",
       "      <td>The frequency range of the G.E. RIC5-9D probe ...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>frozenset({'G.E.', 'frequency', 'RIC5-9D', 'pr...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>What applications is the Philips C8-4v probe s...</td>\n",
       "      <td>The Philips C8-4v probe is suitable for the fo...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>frozenset({'transvaginal, transrectal, gynecol...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Is the Philips C5-2 probe currently available ...</td>\n",
       "      <td>The Philips C5-2 probe is currently in stock a...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>frozenset({'stock', 'C5-2', 'Philips', 'probes...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Is the HDI 1500 ultrasound system made by ATL?</td>\n",
       "      <td>The HDI 1500 ultrasound system is made by ATL.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>frozenset({'manufacturer', 'systems', 'HDI 150...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Is the ATL C4-2 probe used in general medicine?</td>\n",
       "      <td>The ATL C4-2 probe is suitable for the followi...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5</td>\n",
       "      <td>11</td>\n",
       "      <td>frozenset({'C4-2', 'ATL', 'abdominal, obstetri...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>958</th>\n",
       "      <td>Are there any Philips L17-5 probes available f...</td>\n",
       "      <td>The Philips L17-5 probe is currently in stock ...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>195</td>\n",
       "      <td>1916</td>\n",
       "      <td>frozenset({'stock', 'L17-5', 'Philips', 'probe...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>959</th>\n",
       "      <td>Which ultrasound machine is the G.E. 348c tran...</td>\n",
       "      <td>The G.E. 348c probe is compatible with the fol...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>313</td>\n",
       "      <td>1917</td>\n",
       "      <td>frozenset({'G.E.', 'LOGIQ 700', 'probes', 'com...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>960</th>\n",
       "      <td>What can the ATL C7-4 probe be used for?</td>\n",
       "      <td>The ATL C7-4 probe is suitable for the followi...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>183</td>\n",
       "      <td>1920</td>\n",
       "      <td>frozenset({'C7-4', 'cardiac, vascular, general...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>961</th>\n",
       "      <td>What kind of array is the ATL C8-4v?</td>\n",
       "      <td>The ATL C8-4v is a convex array type probe.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>279</td>\n",
       "      <td>1923</td>\n",
       "      <td>frozenset({'ATL', 'array type', 'C8-4v', 'prob...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>962</th>\n",
       "      <td>Can you use a RIC5-9D probe on the Voluson E6 ...</td>\n",
       "      <td>The G.E. Voluson E6 ultrasound system is compa...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>303</td>\n",
       "      <td>1924</td>\n",
       "      <td>frozenset({'systems', 'G.E.', 'compatibility',...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>963 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              question  \\\n",
       "0    What is the frequency range of the G.E. RIC5-9...   \n",
       "1    What applications is the Philips C8-4v probe s...   \n",
       "2    Is the Philips C5-2 probe currently available ...   \n",
       "3       Is the HDI 1500 ultrasound system made by ATL?   \n",
       "4      Is the ATL C4-2 probe used in general medicine?   \n",
       "..                                                 ...   \n",
       "958  Are there any Philips L17-5 probes available f...   \n",
       "959  Which ultrasound machine is the G.E. 348c tran...   \n",
       "960           What can the ATL C7-4 probe be used for?   \n",
       "961               What kind of array is the ATL C8-4v?   \n",
       "962  Can you use a RIC5-9D probe on the Voluson E6 ...   \n",
       "\n",
       "                                               content  labels  doc_id  \\\n",
       "0    The frequency range of the G.E. RIC5-9D probe ...     1.0       0   \n",
       "1    The Philips C8-4v probe is suitable for the fo...     1.0       2   \n",
       "2    The Philips C5-2 probe is currently in stock a...     1.0       3   \n",
       "3       The HDI 1500 ultrasound system is made by ATL.     1.0       4   \n",
       "4    The ATL C4-2 probe is suitable for the followi...     1.0       5   \n",
       "..                                                 ...     ...     ...   \n",
       "958  The Philips L17-5 probe is currently in stock ...     1.0     195   \n",
       "959  The G.E. 348c probe is compatible with the fol...     1.0     313   \n",
       "960  The ATL C7-4 probe is suitable for the followi...     1.0     183   \n",
       "961        The ATL C8-4v is a convex array type probe.     1.0     279   \n",
       "962  The G.E. Voluson E6 ultrasound system is compa...     1.0     303   \n",
       "\n",
       "     query_id                                               tags  \n",
       "0           0  frozenset({'G.E.', 'frequency', 'RIC5-9D', 'pr...  \n",
       "1           4  frozenset({'transvaginal, transrectal, gynecol...  \n",
       "2           5  frozenset({'stock', 'C5-2', 'Philips', 'probes...  \n",
       "3           7  frozenset({'manufacturer', 'systems', 'HDI 150...  \n",
       "4          11  frozenset({'C4-2', 'ATL', 'abdominal, obstetri...  \n",
       "..        ...                                                ...  \n",
       "958      1916  frozenset({'stock', 'L17-5', 'Philips', 'probe...  \n",
       "959      1917  frozenset({'G.E.', 'LOGIQ 700', 'probes', 'com...  \n",
       "960      1920  frozenset({'C7-4', 'cardiac, vascular, general...  \n",
       "961      1923  frozenset({'ATL', 'array type', 'C8-4v', 'prob...  \n",
       "962      1924  frozenset({'systems', 'G.E.', 'compatibility',...  \n",
       "\n",
       "[963 rows x 6 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merging back the tags with the fine-tuning dataset\n",
    "tags_df = pd.read_csv(\"../data/questions_docs_tags.csv\")\n",
    "tags_df = tags_df.drop(columns=['Unnamed: 0'])\n",
    "tags_df_unique = tags_df.groupby('content')['tags'].apply(lambda x: ','.join(x)).reset_index()\n",
    "merged_df = pd.merge(fine_tuning_dataset, tags_df_unique[['content', 'tags']], on='content', how='left')\n",
    "merged_df = merged_df[merged_df['labels'] == 1.0]\n",
    "merged_df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "You try to use a model that was created with version 3.0.0.dev0, however, your version is 3.0.0. This might cause unexpected behavior or errors. In that case, try to update to the latest version.\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Load the models\n",
    "models = {\n",
    "    'multi-qa-mpnet-base-dot-v1': SentenceTransformer('multi-qa-mpnet-base-dot-v1'),\n",
    "    'all-mpnet-base-v2': SentenceTransformer('all-mpnet-base-v2'),\n",
    "    'multi-qa-distilbert-cos-v1': SentenceTransformer('multi-qa-distilbert-cos-v1')\n",
    "}\n",
    "\n",
    "# Generate embeddings for the content and questions\n",
    "embeddings = {}\n",
    "for model_name, model in models.items():\n",
    "    embeddings[model_name] = {\n",
    "        'content': model.encode(merged_df['content'].tolist(), convert_to_tensor=True, show_progress_bar=False),\n",
    "        'question': model.encode(merged_df['question'].tolist(), convert_to_tensor=True, show_progress_bar=False)\n",
    "    }\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_rr(embeddings, ranking_method):\n",
    "    results_df = merged_df.copy()\n",
    "    results_df['absolute_rank'] = 0\n",
    "    results_df['rr'] = 0.0\n",
    "\n",
    "    for i, question in enumerate(merged_df['question']):\n",
    "        relevant_doc = merged_df['content'].iloc[i]\n",
    "        query_emb = embeddings['question'][i].unsqueeze(0)  # Add batch dimension\n",
    "\n",
    "        if ranking_method == 'dot_score':\n",
    "            scores = util.dot_score(query_emb, embeddings['content'])[0].cpu().tolist()\n",
    "        elif ranking_method == 'cos_sim':\n",
    "            scores = util.cos_sim(query_emb, embeddings['content'])[0].cpu().tolist()\n",
    "\n",
    "        # Combine docs & scores\n",
    "        doc_score_pairs = list(zip(merged_df['content'], scores))\n",
    "\n",
    "        # Sort by decreasing score\n",
    "        doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)\n",
    "\n",
    "        # Find the rank of the first relevant document\n",
    "        matching_indices = [idx for idx, (doc, score) in enumerate(doc_score_pairs) if doc == relevant_doc]\n",
    "        if len(matching_indices) > 0:\n",
    "            rank = matching_indices[0] + 1\n",
    "        else:\n",
    "            rank = len(doc_score_pairs) + 1  # Relevant document not found\n",
    "\n",
    "        # Calculate reciprocal rank\n",
    "        rr = 1 / rank\n",
    "\n",
    "        # Update the corresponding row in results_df\n",
    "        results_df.at[i, 'absolute_rank'] = rank\n",
    "        results_df.at[i, 'rr'] = rr\n",
    "\n",
    "    return results_df\n",
    "\n",
    "# Compute RR for each model and ranking method\n",
    "results = {\n",
    "    'multi-qa-mpnet-base-dot-v1': compute_rr(embeddings['multi-qa-mpnet-base-dot-v1'], 'dot_score'),\n",
    "    'all-mpnet-base-v2': compute_rr(embeddings['all-mpnet-base-v2'], 'dot_score'),\n",
    "    'multi-qa-distilbert-cos-v1': compute_rr(embeddings['multi-qa-distilbert-cos-v1'], 'cos_sim')\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Define the order of models\n",
    "model_order = ['all-mpnet-base-v2', 'multi-qa-mpnet-base-dot-v1', 'multi-qa-distilbert-cos-v1']\n",
    "\n",
    "# Filter RR values for different 'probes' and 'systems' question types\n",
    "probe_types = ['manufacturer', 'compatibility', 'array type', 'applications', 'stock', 'frequency']\n",
    "system_types = ['manufacturer', 'compatibility']\n",
    "\n",
    "results_rr_by_type = {key: {ptype: df[df['tags'].apply(lambda x: 'probes' in x and ptype in x)]['rr'].dropna() for ptype in probe_types} for key, df in results.items()}\n",
    "\n",
    "# Add 'systems' question types to the results\n",
    "for key, df in results.items():\n",
    "    results_rr_by_type[key].update({f'systems_{stype}': df[df['tags'].apply(lambda x: 'systems' in x and stype in x)]['rr'].dropna() for stype in system_types})\n",
    "\n",
    "# Define the complete list of question types\n",
    "all_types = probe_types + [f'systems_{stype}' for stype in system_types]\n",
    "labels = [ptype.capitalize() for ptype in probe_types] + [f'Sys {stype.capitalize()}' for stype in system_types]\n",
    "\n",
    "# Create the boxplot\n",
    "fig, ax = plt.subplots(len(model_order), 1, figsize=(6, 7))  # Adjust the figure size\n",
    "\n",
    "for i, model_key in enumerate(model_order):\n",
    "    rr_by_type = results_rr_by_type[model_key]\n",
    "    data = [rr_by_type[ptype] for ptype in all_types if ptype in rr_by_type]\n",
    "    box = ax[i].boxplot(data, vert=True, patch_artist=True, medianprops=dict(color='yellow', alpha=0.5), boxprops=dict(facecolor='cyan', alpha=0.3, edgecolor='cyan'), capprops=dict(color='cyan', alpha=0.3), whiskerprops=dict(color='yellow'), flierprops=dict(markeredgecolor='yellow', markerfacecolor='yellow', alpha=0.3))\n",
    "    ax[i].set_xticklabels(labels, rotation=45, fontsize=8)  # Reduced font size for x-tick labels\n",
    "    ax[i].set_ylabel('RR', fontsize=10)  # Reduced font size for y-label\n",
    "    ax[i].set_title(f'{model_key} Reciprocal Rank Distribution by Question Type', fontsize=12)  # Reduced font size for title\n",
    "\n",
    "    mplcyberpunk.make_lines_glow(ax[i])\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Levene's to test variance homogeneity across models and whether to use a parametric or non-parametric test.\n",
    "\n",
    "- Null Hypothesis (H0): The variances of MRR scores are equal across the different models.\n",
    "- Alternative Hypothesis (H1): The variances of MRR scores are not equal across the different models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Levene's test by model p-value = 2.787789801080547e-14\n"
     ]
    }
   ],
   "source": [
    "# Prepare data for Levene's test by model\n",
    "data_by_model = {model_key: [] for model_key in model_order}\n",
    "for model_key in model_order:\n",
    "    for ptype in all_types:\n",
    "        if ptype in results_rr_by_type[model_key]:\n",
    "            data_by_model[model_key].extend(results_rr_by_type[model_key][ptype])\n",
    "\n",
    "# Perform Levene's test by model\n",
    "levene_result_by_model = levene(*data_by_model.values())\n",
    "\n",
    "# Check the result\n",
    "print(f\"Levene's test by model p-value = {levene_result_by_model.pvalue}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The variances of MRR scores are not equal across the different models so we will use Friedman Test (non parametric repeated measure one-way ANOVA) and Nemenyi post hoc test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAADtCAYAAABwHzY2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAABB5ElEQVR4nO3dd1gUVxfA4d8CuywIIkURLNh7b4k1Go0xJkaNLRp7jTF2sSexd2M3JlFjiS0ao9HPbizB3ntBo2IFQUDaVvj+QFdXQHaVJpz3eXx0Z+/ce2bnumfm3plZRcGSDeMQQgghRIZkk94BCCGEECJpkqiFEEKIDEwStRBCCJGBSaIWQgghMjBJ1EIIIUQGJolaCCGEyMAkUQshhBAZmCRqIYQQIgOTRC2EEEJkYHbpHYDIHA7tXQbAF20HERj0xOy9Zp/Vw3dgJ5Ys38TSFZto/HEtRg3tblZGq9MTGBTCug072bRlHwAVy5dg/o/Dzcrp9QaCQ8LYttOPpSs2pdr2vKzxx7Xo2rEZLb8akuj7rq7Z6d29FTXeK49jNgfu3w9k89b9bNi0x1Rmw6oZeOX2SHT9bwdN4cy5qwmWly5ZmO+G9yBXLndmzVvJlm0H32o7Du1dxsRpi/mkYS0eBQYzcdri15b39spJ/ny5OXr8wlu127VjMyqWL0HfwVMSvLdh1Qy27fTDK7cHuT09Ei2TI4czFcuVYN/BE28Vx7vCzs6WTxrWYsu2AwDMmzmcM+eupll/FxmPJGqRYvR6A7WqV+TPzXvNltepVYnY2FizZYFBIXT/ZpzptVM2Bz5rXAffAZ24fecBZ89fM73XpGV/078d1Cpq1azEt73a8OBhEDt2H06lrbHcjEmDePjoMYNGzCQiIoqypYsyuF8H7JR2rF2/w1Ru9oJV7N13PMH6TyMiE623fdtPufcgiAFDZ/D0aeJlrBUXFwfEPfv79UYM6cqZc9feOlEnF8/zP5B4TN/0aI1CocgyifqjD9+n01dNTIl65Jh5GPSGdI5KpCdJ1CLFnD1/jVo1zBO1o6OaMqWKcP1GgFnZ2Ng4noSGm14/CQ1n4S9/UKdmJerUrGSWqF8uB7Buw06qVytLnVqV0z1RFy6UlxLFCtB/yDQio6IBePgoGK/cHnze+AOzRB0VFZNgW17HKZsDZ89f41FgcIrE+jg4lCehTwkOCSM4JMyCNRQp0u7rBIeEERr6FJXKDqVd4l9HCkXqx5GRvLq9ERFR6RSJyCgkUYsU43f4DH16tcHRUU10tAaAGu+V59yF66jV9hbVodcbMBpjky2nS6Zcx3af0aTxB+T0cCU8PIJNW/fz28rNQPxQ4olTl6hQrhgVyhUn8PETZs37neMnLwLg4Z6D4UO6UqFscQLuPeTI0fNJthMbG38WWLVyabMzvj837WX3P0ct2ubEzJs5nEoVSlCpQkm6dmxGzfqdcXZypHeP1tSuWRGVUonfkbPMnv87EZHRVCxfglFDu3P0+Hk++vB9VqzZyqq128zqvH3nAfcfBBFw95Ep+VeqUJJ+vduSP78XwcGhrFq3jc1b9zNqaPdn7ZcwDVvn9HClb++2VK1Uiti4OHbvPcqCX9ahf3a2917VsvTq1gKffF7cvR/IvJ/WcOrMFbMYVEols6f7otXq8B01izsBD7j/MAitTodOl/CssWvHZjT+uBYQPxXy9/8OUL9eNTr1+M5U5stWjahTsyLfDJjMob3LmDxjKR3bfYZrDmf8Dp9l2qxlxGi0AJQvW4x+vdtSsEAe7t0PZOmKzez/92Si+8DOzpZve31Jo4Y1iY6OYd5Pa5nwQx9atBvCo8BgDu1dZjZt8eoUScECeRjUtz2lSxYmMCiEPzbu5q+//wHAKZsjI4Z0pXKlUsTFxXHk2DlmzFlB8aIFTNNCh/Yuo0W7IYwa2t1s6Lvxx7X4qk1jvHJ7cOv2feb+tIZzF64D8VMJq9Zt45OPalKkSH4CAh4yecYSrvnfAaBl8wa0bdUIV1cXbt26x5yFqzl/0T/Z/ijSl1xMJlLMzVv3eBwcyvtVy5qW1alVmYOHTie7rlJpR/PPP6SAjzf7/RL/4gSwsVFQt3YVqlUuw/4khkI/aViT1l80ZMrMpXzZaTi/rfyb7p2bU6yoj6lMp68+Y8++Y3ToNgr/GwEMG9TFdCYz4Yc+2NrY0KPPWFat3UbrFh8lGc+t2/c5efoyE37ow9JFY+jVrQUVy5cgOiaGBw8fJ7vdSRk5Zh4XLvmz+o/tpqH/SWP7UbRIfnxHzWbA0OkUyO9lNtfvldsDlUpJ195j2JPIQcKAodO5dz+Q31ZuZvuuQ9jYKJjwQx/+OXCCdp1HsHjZXwzu15ECPt7MXrDK1P7IMfOws7Nl7sxhOKjt6TNoCt+NW0iN98vzTc/WABT08WbahP4c+PcUnXp+z55/jjFlfH/cXF1M7SsUCsaM/hpbWxtG/DAXg8HIlJm/ceLUJbbvOmQ6kHrZmj+2s3ffMfbuO0b3b8ayd99RihTKR768nqYyH35QlT0vTSn06PIFs+avou/gqRQulJehAzsD4ObqwrQJA9i2048O3Ufz+7ptjBrajfJliyW6D3p2bUGtGhUZNno2YyYuolvnZhbvP5VKyczJgzh/wZ+OPb5j/s/r6NKhKR83qAFAt87NcHNzoXe/CfQbPJWihfPTuf3nXLjkz+wFqwgMCqFJy/4EPQ4xq7fxx7UY2Lc9K9dspXPP7zlx+jIzJg/CwyOHqUy3Ts1YufZ/dOoxmsioaAZ82x6AokXy06dnG2bMWUm7ziM4d+E647/vk+VGLN5FckYtUpTf4TPUqlGRfw6cQKm0o1qVMvw4byUN61c3K+eZy43dWxeZXturVATcfcj343/i0uWbZmVfLqdSKQkMDGHuT2vYuz/hfC/Ez39PnL7EdDa3aes+unRsSkGfPFx/dmZx+Nh5tu30A2D5qr9Z8esE3N1ccHZypGzpoqaL4m7deUDxYgX48INqSW6z78hZtG3diEYf1aBjuyZ0bNeE+w+CGDNxEZev/mcqN2RAJwb2bW8ea2AI7buNSlBnREQUer2RmBgtT0LDKVwoL5UqlODLTsO4ey8QgLGTf2HNssnkz5vbtN7va7dx/0FQkrG+LFs2R1yyOxEaFs6jwGAeBQYTHBJKSEgYUVExpvYjIqKoVb0COd1d6dlnHBGR8UP8M+euZNqE/vyy5E8+a1yH8xdvsHzVlmdx/A8HtT1OTo6m9gb2bU++PJ58M3AyGo3OohhjNFq0Oj0AYeERhIVHcOnKTerVqcaK1VvwzOVO8aI+DB09+6XP4H8cOXYOiL8uYNY0X2bMXUGLpvU5efqyaWrm/oMgihXxofUXDU1npM8pFAo+bVSbuQtfnK0u/OUPpk8caFHcDetXJzQsgl+XbQTg3v1AVqzaQusWDdm55zBenh7ExGh48CgYrVbHqLELUCjAYDASFRWTYGrouZbNG7Dhr92mKZ9Fi9dTsXxxWjZtwKIlGwDYvvMQ/z47OF67fgcTfvgWAC9PD+LAtK9/+e1PDh09i41CgdGCaxZE+pFELVLUv4fPMHHMt9ja2FClYin+u3WXsLCIBOWCQ8L4dtAUFAoFpUoWYkCfr9i6499ELxjq3PN7AAr4eOM7oBP/Hj7NxlcuWHvZ6bNXKVWiEF93a4mPjzfFiuTHwz0HtrYvBpDu3X9k+ndUVPwwva2tLQV88hD+NNLsyvUrV2+9NlHr9HqWr9rC8lVbyOOVkxrVK9C2VSOmTuhPi7ZD0OnjE83iZRs58O8ps3UNBmOS9b6sQH5vnkZEmZI0QMDdhzx9GomPjzeRz5Lno0eWz2dHRESxcfNehg/uSuf2n3PoyFm27vjXlIhf5uPjzd17j8zeu3DJHzs7O/Lm8SR/Xi+u+d82W+d5kgIoU6owFcoV4/LV/956znXPP8do/HEtVqzeQv261Thz7qpZH7vw0lDu1Wu3sbO1JV/e3Pj4eFGzegWzAz87W1vu3nvEq1xcnMjh4oz/zRfXVly4dMPiGH3ye1GkcD6ztmxtbDAa4/f3Hxt3M3V8P/735zxOnr7EvoMn2f3PkWTrLZDfm6UrzEceLl6+gY+Pt+n13Zf7drQGpTL+a/7YyYv8999dfl8ykWv+t/n30Bm2bDuAMTb5qSaRviRRixR1/tnZR7myxahdsxIH/BIf9jYaY01nfvfuB2I0xjJudG8ePnzMAT/zZPa83P0HQYSHR7Jw9kiCHoeybsPOROtu0rgO/Xq3Zcu2g+w/eJL5i9Yyb+YwszJ6fcIE+XwI8NWBQL0h6Stu69auQo4czqZbyu4/fMz6jbs5duIia5ZNplChvFy9dguAsLAIi892X/X8rPJVNjY22Nq8OAB5flBgqZlzV7Jx815q16xMnZoVafpZXYZ9NyfBld66RNp/3q6NjQKD8fUHHNExGkb+MI/pEwfSpHGdt7rVbO/+Y3z79Zfk8c5F3TpV+Pt/B8zef/ngx+ZZjHGxsdja2rJzz2FWrN6aZPnntM/O+F8eFdYn89m+fCBoZ2vLqdOXmTl3ZaJlT5+9QvMvB1O7ZkWqv1eeYYM6817VMoyb/Mtr20isH9jY2Ji2M6ntAdBqdfT4djwVyxenZvUKfNqoNs0//5CuvX8gODjste2K9CVz1CJFGWNjOXz0HLVqVKBm9QocfCXpJmXvvmP4HTnL4P4dcXRUJ1nu4uUb/PX3P/Ts8gWeudwTLdPss3r8tvJv5v60hp17DhMeHmk2V/o6/92+R/bsTuTxzmVaVqyIT5LlPT3d6dL+c1Qqpdny52e4iY0mvImAu4/I7pzNbJi7gI83Tk6O3Ln78I3qdHN1YVC/Dtx7EMSK1Vvo3mccJ09foVb1is9KvBgODbj7iHx5c+PsnM20rEypIhgMBu4/eMy9e4EUKZTPrP5Fc0dRv957APx36z7nLlxn2aq/6dWtpVk9yXn1VrKQJ+GcOXeVzz6pTZHC+TnwysVgRYvkN/27RPEC6HR6Au4+IuDuQ/Ll9eT+gyDTn9o1K/FxA/NpGYgfcn8UGELJ4oVMy17tBzqd3qyvenvlNP37zt2H5Mubm4ePHpvaKl2qMC2bNwCgTYuGFC/mw/Zdh/h+/EImTl9C3dpVEt3el929+4gyJQubLStdsjABFvSB0qUK06HtZ5w+e5V5P62lbefhqFRKypdJfI5eZBySqEWK8zt8hiaffEBo6FMeWjEUO2fBKpycHOnasdlry/3620ZiYrT069020ffDn0ZSpVIp8uX1pHhRH8Z91xul0g6VUplo+ZfdCXjIiVOXGOnbjcKF8lKrRkXTl2titu30w2CMZfa0IVSuWJLcnh5UqVSKcd/1Zv/Bk2a3VmXL5oCbq0uCP2q1Ktm4Au4+5Mixc4we3oMSxQtSsnhBRj+7GvjW7fvJrp+YpxGRfFCrMv16tyWPV07Kly1G0SL5uX4jfh4/RqMlXx5PcuRw5sSpizx4GMT3w3tSqGD8fPnAvu3Z/c9RIqOi2bTlH8qXLUablh+TxzsXHdp+SkGfPJx76TY7gD/+3EVEZDRfd2tpcZwajZbcnh5mF0zt2XeUNi0+5sSpSwmG6rt3bk6FcsUpVaIQA779iu27DhGj0bJx8z+UKFaQHl2+IG8eTz768H16dW3Bo8AQErPmj+306NycqpVLU7yoDwO/Nb++4Mq1W7Rs1oC8eTypVb0CjT+ubXpv154jqNUqhg7sTP58XlSvVo6Bfb4iNDT+wC1nTjcG9e1A6ZKFyZvHk3p1qph97s5OjuTN42k2WgKwdsMOWjRvwMcNapAvrydfd29FkcL5LRqh0Gn1dO3YlCaN65Db04MG9d7DwcGeG//dTXZdkb5k6FukuGMnLmBnZ2vR1d4ve/DwMWvWbeerLxubHvaQmMioaBYt2cCIIV2pUqkUJ09fNnt/9oLVjBrajeW/jCc07Cl79x9Ho9FStGj+JGo09/2EhQwf1IWf537Ho6Bg1m/cxaeN6iRaNiIiim8GTKRHlxZ8P7wnLi7OhDwJZ/feIyx95SrmAX2+YkCfrxLU8cvSP00XYb3O+Cm/MrBve+ZOH4oxNha/w2eYu3C1RduUGIPByLDv5jCgTzuW/zqB6BgNW7cfNH3pb9l2kJFDuuLjM5iuX49h2HdzGNS3A7/O/47oGA279h7h58V/AvFD/qPGzqd391b06taSW7fvM3T07AT3axsMRuYuXMPU8f3Zsu0AV6/fTjbOHbsPM3lcP5b/Mp5Pv+gLwP6DJxnSvyN79x1LUH77rkOMHtYDJydH9vxzlDnPPqPAoBCGjp5N7x6tadf6Ex4HhzJ/0Vp27U18bvjPzXtxdc3O2NG9MRiMrN2ww+zOgVnzf2f44K6sXDyBK9dusXjZX3T6qgkQP9Q/eMSP9P+mHct+GcfTp5H8uXkvK9fED7v/+ttGnLI5MHV8fxwc7Dlz/hpjJ/0MwKkzV+JHORZP4Jv+E81i+ufACdzcXOjR5Qvc3Fy4cSOAgUOnW3RG7X8zgEkzltCl/ecM7NuBwMBgxk3+hTsBbzYiI9KOomDJhnK5nxDinZI3jyfLfh5Hk5b9TPdIAwnubU5JuT09+HP1DNN91EKkFTmjFkK8Mxwd1FSrWoZmn9Vlz76jZklaiMxK5qiFEO+U4YO74uzsxM9L/0zvUIRIE+k69K1U2rH0pzH8OO/3JIeqihbJj++AThQumJdbt+8zffZy0+PwhBBCiMwu3c6oVUolY0d9TaGCeZMso1armDFpEOcuXKdr7zFcuHyD6ZMGWXSVrBBCCJEZpEuiLuDjzS/zv8P7pXtVE1O/7ntotToW/LyOOwEPmbNgNdHRMa99SpQQQgiRmaRLoq5Qrjinz16hV98Jry1XumThBL/scuHSDUqXKpzEGkIIIUTmki5XfT9/3GJy3N1zJHiYw5PQcAoVSHq4XAghhMhMMvTtWWp7VYLn6+r1BtND5hNdx8EVhY1czC6EECJji4lK/Kl4r8rQiVqn06N85bGPSqUdWm3SP5GniQkl4c8qiMQ4ZHMjJupJ8gVFlid9RVhK+krKy9CJ+nFwKO5u5j+m4O7mQvCTsGTWlIetJe/lgxn5vMTrSF8RkNPDjaKvuUsnngJb++wYtU9Jrq/437rH42BJ6JbI0In60pWbtG/7qdmysqWLWvRcZCGEEClnSP+OtGrxSYrVt/7P7Qz7bnaK1ZeZZbhE7ebqQmRUNDqdnn0HT9C7eyv692nH5q37afpZXdRqe/45cDy9wxRCiCxlxpwVbP77nyTft1UpWf7rJAA69RiFUZf0FCXEn1ELy2S4RL1lwxwmTlvMtp1+REdr8B01C9+BnWj6aV1u/HeXISN/RKN5fQcQQoiMqGrlJ9ir3tXpgxDAP8l3lXbql/59gLhYzWtrK140/s+7SKtTcOKUW5q1l+6Jumb9zq99feXaLbp+PSbN4hFCiNRir4rDzU2LUzZjeoeS4pTKFwcg+fLGoNfHpGM0qScyypYnT+zTtM10T9RCCJGVOGUzkiuXBoMhc91Gamdna/q3q6sOgyHzjXza2cVCkJonaXwNnCRqIYRIYwaDDZevOqd3GFZxye6Bl1f+JN+PQ/XSv2sTG/f6nyB9+DCA8Kfv1u96lyoRkS7tSqIWQgiRrOZNvqFWzeYWle3/zcJky/gd+otlq8a9bVhZgiRqIYQQyfpry0KOntyaTCkFKpUjOl00yd1H/fBhQIrFltlJohZCCJGs8KfBFgxVK7BXO6PVRCAPx0k5metqBiGEECKTkUQthBBCZGCSqIUQQogMTBK1EEIIkYFJohZCCCEyMEnUQgghRAYmiVoIIYTIwCRRCyGEEBmYJGohhBAiA5NELYQQQmRgkqiFEEKIDEwStRBCCJGBSaIWQgghMjBJ1EIIIUQGJolaCCGEyMDk96gzoZwebhQtmDeZUgps7bNj1D4lud+N9b91j8fBT1IsPiGEEJaTRJ0JDenfkVYtPkmx+tb/uZ1h381OsfqEEEJYzqpEXapEIS5f/c9smaOjmt7dWzFz7soUDUy8uRlzVrD573+SfN9WpWT5r5MA6NRjFEad7rX1+d+6l6LxCSGEsJxViXrahAGMGjufcxeuA1C/3nv0692WOwEPUiU48WYeBz957VC1g4O96d8nT18kJkaTFmGJDEimSYTI+KxK1N+NX8i4777hl6V/Ur9uNQoVzMu8n9awZ98xqxpVKZUM6t+BurWroNXqWLN+B2vX70i0bJ2alejVrSW5crnhfyOA2QtWcd3/jlXtCSESJ9MkQmR8ViXqM+euMmTkj8ycPJhbt+/zZadhaDSvHzZNTJ9ebShRrAD9hkwldy53Rg/rwaPAYPYfPGlWrqCPN2NGfc20Wcu4cNGfNi0/ZvrEgbTuMBSt1vp2hRDmZJpEiIzPokQ9YkhXs9cBdx9SvmwxfAd0xmAwADB5xlKLGlSrVTRpXIfBI37kuv8drvvfYdW6bbRo2iBBoq5WpQy3bt9nx+7DACxavIEWzRpQ0Mebq9dvW9Tem4qpXAZUqlRtI92olKZ/xlQrT4xOn47BpCKdDodTF9M7igxNpkmEyPgsStSPAkMSvD515sobNVikUH5s7Wy5cMnftOz8RX86fdUEhUJBXNyLObDwp5EULJCHsqWLcPHyTRo3qkVkZDT3HwS9UdtWUakwuGUnNlu21G8rjdkpX+x2fV4vdPrMl6htoqKwe/I0vcMQQoi3ZlGi/m3l5hRr0MM9B+HhkRgMRtOyJ6Hh2NurcMnuRFh4hGn53v3HqVWjIovmjsZgNBIXG4fvqFlEREYn04oiRWKNzeZEbC43eDZqkFnE2r3Y7bGu2YnNZNuHnR0EKeDJU1KqL7xO1egy2Mcqky/4DlIZX2xX9cjy6GIy30Hdc1obPScc03IEJvX7ZvqS7Uve6y/OfM6qOWqv3B60//JT8njnwsbGPMh+Q6ZZVIe9vQr9K2dwen18olAqzcNxye6Em6sLM+eu5NLlGzT//ENG+najy9c/EBYWQWLUDq4obN7+gWtGtTPYO2K0UaG+/fCt68tIHF4a+nYKCMY2kw19a4r5YGvviErtjEM2t1Rvz9nghqshO05Gh1RvK60pFS/+TxZSFEavyGQHdc9E2sYQavc01fuLSq1Faa9AqbLDXu2cqm2lp8y6bUqVAaW9PSq1U4r0lZiokOQLYe1V38N7EhEZzZ59xzAajcmvkAidTo9SaX728TxBa165QKx3j9b8d+seGzfvBWDqj8tY/dskPm1Um1VrtyVavyYmlJQ40tFpItBpnYnV2WGjSfyg4F1l99LZn14TgS6TJWq9LhqjNho0EcREpf6tQjpNBPY6Z1wNagyZLJHZvfQV4RRjhyG5wax3kF2cHXq7aHSqCGLsUre/6DQR6LXR6HU6tJrM+bwpe7Uz2kz2nfmcXheNXmtAp4kjJso++RVSiFU9pUihfHzWot9bzWk+Dg7FxcUJWxsbjLGxALi7uaDRaIl8ZUi7eDEfNmzcY3odFxeH/8275M7lnkwrlg0nJC/ulb/fDbmyO1HKyzPJ91+eo65RrBCGZPbn5YeBBD2NTLH40kZa77s4DAoDl9U306i9tKGyf3FQd9X+FrrYzHe3RSlNYdLn//q79b1imZdPkjLj9r0s7bbPqkR99fotChXM81ZXXPvfDMBoMFK6VGHOX4y/oKxcmWJcuXbL7EIygODgMAr4eJsty58vN7uu3Xrj9rOCEU3q0rZmNYvK/v5Nu2TLrDl0nIGrtr5tWEIIId6AVYn6waNg5v04nNNnrxIWZn5FraW3Z2m1OrbvOoTvgE5MnL6EnB6utG3ViEnTlwDg5upCZFQ0Op2eLdsOMGpod65cu8XFyzdo0vgDcnt6sG2XnzVhZzmTt+xn88nkrspXYKNyIFYXQ3JHhpcfBqZUaEIIIaxkVaIODAxh9brtb93o3J/W4DugI/NmDiMqKoYlyzdxwO8UAFs2zGHitMVs2+nH3v3HcXBQ07HdZ+TK6Yr/zQD6DZma5IVkIl7Q00gLhqoVqNTO6DQRZP4hKpEUl1xOeBVPeppEafviK6L4e4XQG18/TfLwWiDhQe/aNIkQGZtViTqp27QUCusu3tJqdUyYupgJUxcneK9m/c5mr7duP8jW7Qetql8IYZnmQ+tSq41l0yT9VyY/TeK37jjLhsg0iRApyapEXdDHm47tmuDm5sLz3GxnZ0vePLn5vFX/1IhPCJGK/pq2n6Obk58mUdk6oDMmP03y8JpMkwiR0qxK1COHdufuvUfcCXhAHm9Pjp44T4um9flz057kVxZCZDjhQZEWDFUrXrrlRqZJhEhrVj0ZpFCBPEyavoT1f+1GrVaxfuNuhn8/l/p1LRs6E0IIIYR1rErUIU/CUSgUPHjwGJ/8XgDcvvMAz2TvaxZCCCHEm7AqUZ8+d5Xx332Dg4OaGzcD6NapGV992ZjgkLBUCk8IIYTI2qxK1LPm/c41/zsoFDB91nJKlyzMB7UqM/XH31IrPiGEECJLs+piMq1WZ7pFKyIymkHDZwJQsECelI9MCCGEEJadUavVKr7u3oqpEwbQ+ouGpuV2drb06PIFS38ak1rxCSGEEFmaRWfUwwd3pWjh/Bw9cZ7WLRoSRxz/HjrD1An9yZHdifFTfkntOIUQQogsyaJEXbVyabp+PYbAoBD+t8OPMSN70abFx5w6c5k5C1cTHa1J7TiFEEKILMmiRK1S2hEYFP8D1//dukf+fF4sWrKBtet3pGpwQgghRFZn0Rz1K78+iV6vlyQthBBCpAGrbs96zmAwpnQcQgghhEiERUPf9moVc2cMNb12dFSbvQboN2RaykYmhBBCCMsS9ZQZS81eb991KFWCEUIIIYQ5ixK1JGYhhBAifbzRHLUQQggh0oYkaiGEECIDk0QthBBCZGAWzVGXL1ss2TLnLlx/62CEEEIIYc6iRD16WI/Xvh8XF0frDkNfW0YIIYQQ1rMoUbdq75vacQghhBAiEVb9HrVareKjD6vj7uaCQvGsAls7fHy8GD12QWrEJ4QQQmRpViXq70f0In/e3IQ/jSSbowMPHj3m/apl+d+Of1MrPiGEECJLsypRV6lYilbtfcmVy40eXb5g5A/zqFm9Am1aNEyt+IQQQogszapEHROjIfxpJBqtjqKF8wNw6MhZRvp2s6pRlVLJoP4dqFu7ClqtjjXrdyT5a1yFCuZlSP+OlChWgHv3A5m9YBWnz161qj0hhBDiXWXVfdT+NwNo/+WnGAxGIiKiqFyxJKVLFsZojLWq0T692lCiWAH6DZnKzDkr6NqhKXXrVElQLls2B2ZP8+X2nQd06D6aA36nmDS2HzlyOFvVnhBCCPGusipRz16wmveqlsHNNTsLf/2DKeP6sWDWCJau2GRxHWq1iiaN6zBnwWqu+9/h4KHTrFq3jRZNGyQo+0nDmsTEaJgxZzn3HwSxZPkm7t17RMliBa0JWwghhHhnWTX0fe9+IAN8pxNHHI+DQ/mq6ygio6KJjtZYXEeRQvmxtbPlwiV/07LzF/3p9FUTFAoFcXFxpuWVypfg38NniI19sax7n3HWhCyEEEK806w6oy5Xpiib/phF8aIFAGjSuA6rlk6iWFEfi+vwcM9BeHgkBoPRtOxJaDj29ipcsjuZlfX2ykVYWARDB3bm7/Vz+GXed5QtXcSCVhQp8IdX/s5sf1L688pIf9J636V1e+nxeT6X3rG86/svK3yWsn1vVl/SrDqjHtS3Az/O/Z0r124BsGT5Jm7eusfQAZ0sPtO1t1eh1+vNlun1BgCUSvNwHBzsad/2U9Zv3MXgETNpUO89Zk31pV2XEQQ9fpJo/WoHVxQ2b/8Ic6PaGewdMaocUakz75x4Zty2WJUjtvYGVGpnHLK5pXp7KoMzSoUjSoMj9pnw83wus26bMtYRpZ0BlX3q9xeVWovSXoFSZZdpP0/IxH1FZUBpb49K7ZQifSUmKsSiclYl6jzeOdn/70mzZQf+PcXwwV0trkOn06NUKs2WPU/QGq3ObLnRGIv/jTssWb4JAP8bAVSrXIZGH9VgxeqtidaviQnFmiOVJOPURKDTOhOrs8NGE/HW9WVEKrUzuky4bXpdNEZtNGgiiIlK/IAuJek0Eeh1zuiNdmhtMt/nCfFfvNpM2Fcgvr/ojdHo4iKIsUvd/qLTRKDXRqPX6dBqrPr6fWdk+r6iNaDTxBETZZ9m7VrVU27eukfzzz9k4+a9pmWff/oBN/+7a3Edj4NDcXFxwtbGBmNs/NXi7m4uaDRaIiOjzcqGPAnjTsBDs2V37z0iV87kjmTiknnfUnGv/J2ZvHwwkxm3L633nfSVd1t67L/M+Flmhb7yXNptn1WJeubclUwd35+2rRoRHBJGTg9XYmNjGTZ6tsV1+N8MwGgwUrpUYc5fjL+grFyZYly5dsvsQjKAS5dvUqF8cbNl+fN7sfufo9aELYQQQryzrErU/jcCaN1hKOXKFMXV1YXg4FAuXblpdmFYcrRaHdt3HcJ3QCcmTl9CTg9X2rZqxKTpSwBwc3UhMioanU7PX1v20aJ5A7p2bMbOPYf5pGFN8njlZOeew9ZtpRBCCPGOsvqqK8+cbpQpVYQKZYtRuWIp8uXNbXWjc39awzX/28ybOYzB/TqwZPkmDvidAmDLhjk0qPceAIFBIQwaNoOa1SuwcslEar5fgSEjZxEcHGZ1m0IIIcS7yKoz6qqVSzPxh285dOQsgUEh+OT34pf53/HDhJ84fPScxfVotTomTF3MhKmLE7xXs35ns9cXLt2gW+8x1oQphBBCZBpWJerePVoxetwCjp+8aFr2XtWy9OnZ2qpELYQQQgjLWDX07e2VixOnLpktO3HqIp653FM0KCGEEELEsypR37p9n4/qv2+2rMGH73M74EGKBiWEEEKIeFYNfc9ftJYfpwym2Wf1CAwMIXduD/Lny43vyFmpFZ8QQgiRpVmVqINDQmnTaRi1qlckh4szR06c58ix80RERKVWfEIIIUSWZlWiXrJwDK3a+7J1+8HUikcIIYQQL7Fqjvp2wAMqVyyZWrEIIYQQ4hVWnVFnc3Rg8rh+aLV6QsOemj3ys3WHoSkenBBCCJHVWZWo5yxcnVpxCCGEECIRViXqs+evkSOHMzXeK4+ra3YeB4dy5Og5Il751SshhBBCpAyr5qjfr1aW9Sun0fjjWhQplI9mn9Vj3cpplC1dJLXiE0IIIbI0q86o+3/TjrGTfsbvyFnTsjo1KzG4f0c69/w+pWMTQgghsjyrzqjdXF04euKC2bJDR86S08M1RYMSQgghRDyrEvVfW/YxqG8HsmVzAEClVNKre0u27fBLleCEEEKIrM6qoe9PGtbEzTU7n31Sh8ioaLI5qrGxic/1bVp+jEIBcXFQp2HXVAlWCCGEyGqsStQ9vx2fWnEIIYQQIhEWJeqcHq48Dg4F4pIsExj0JKViEkIIIcQzFiXqVUsn0fDz3mxYNSPR92W4WwghhEgdFiXqhp/3BqD2R5KMhRBCiLRk1VXf7m4uTBrbl4IF8gDQvXNzxn/fB2fnbKkSnBBCCJHVWZWoR/p2JzAohMDAEAA2bNpDcHAoI327pUpwQgghRFZnVaIuXaowcxasJjpGA0BYWATzF62lQrniqRKcEEIIkdVZlahDQ8MpXaqw2bKiRX0IeRKWkjEJIYQQ4hmr7qNesnwTP04ezMFDp3kcHIqHew5q16jI5BlLUys+IYQQIkuzKlHv2XeM23ce8EHtyri5uvDg4WO+7j+ROwEPUys+IYQQIkuzaugb4MZ/dzl09BwnTl1i1bptREXFWN2oSqlk+JCu7Ni8kM1/zObLVo2SXSe3pwe7ty6iYvkSVrcnhBBCvKusOqP29srJ5HH9yO6cDadsDnT5egzLfxnHiB/mcfzkRYvr6dOrDSWKFaDfkKnkzuXO6GE9eBQYzP6DJ5NcZ8iAjjg6qK0JVwghhHjnWXVG7TugE//b/i/NvxyEwRjLvfuBjJv8C9/0bG1xHWq1iiaN6zBnwWqu+9/h4KHTrFq3jRZNGyS5TsP61SVJCyGEyJKsStQlihVgw6bd8S/i4p/7fcDvFLk9PSyuo0ih/Nja2XLhkr9p2fmL/pQuWQiFQpGgfPbs2fimZ2umz1pmTahCCCFEpmDV0PejwBDKly3OmXNXTctKlSjEo0fBFtfh4Z6D8PBIDAajadmT0HDs7VW4ZHciLDzCrHy/3m3ZvsuPW3ceWBFpwoT/ZhSv/J1ZZcbtS+t9J33l3ZYe+y+zfpbPyfYlL+kfunqZVYl64a9/MHlcPw4dOYu9vYpBfdtT74OqTJy22OI67O1V6PV6s2V6vQEApdI8nCqVSlGuTDHadxtlcf1qB1cUNlZfI5eAUe0M9o4YVY6o1M5vXV9GlRm3LVbliK29AZXaGYdsbqnensrgjFLhiNLgiH0m/Dyfy6zbpox1RGlnQGWf+v1FpdaitFegVNll2s8TMnFfURlQ2tujUjulSF+JiQqxqJxVifrEqUt06fU99eu+R1R0DE9Cn9JnwGQC7j2yuA6dTo9SqTRb9jxBa7Q60zKVSsnQgZ2ZMWcFOp15Yn8dTUwoKXGko9NEoNM6E6uzw0YTkfwK7yCV2hldJtw2vS4aozYaNBHERKX+z6/qNBHodc7ojXZobTLf5wnxX7zaTNhXIL6/6I3R6OIiiLFL3f6i00Sg10aj1+nQaqz6+n1nZPq+ojWg08QRE2WfZu1a3VMePgrm97X/M1vW7LN6bNq6z6L1HweH4uLihK2NDcbYWCD+xz40Gi2RkdGmcqVKFCKPdy4mjvnWbP2Zkwexfdchps9e/ppWLBtOSF7cK39nJi8fzGTG7UvrfSd95d2WHvsvM36WWaGvPJd222dRoi7g483IId0oWMCbk2euMH7KL0RHa8iX15MRQ7pStIiPxYna/2YARoOR0qUKc/5i/AVl5coU48q1W8TFvdjwy1f/o3WHoWbr/rFyGlNm/sbxU5bfCiaEEEK8yyyazB3SvyOPAoMZO+ln7FVKundqTo33y7Nk4RiiojR06Dba4ga1Wh3bdx3Cd0AnShQvSO2alWjbqhHrN8ZfTe7m6oJKpUSn03P/QZDZH4g/Iw8Ly5zDKkIIIcSrLDqjLlbEh6atBxCj0XLx8k2W/zKOBh++z4/zVrJj92GrG5370xp8B3Rk3sxhREXFsGT5Jg74nQJgy4Y5TJy2mG07/ayuVwghhMhsLJ6jjtFoAQgLj8A5uxNDRvzI6bNX3qhRrVbHhKmLmTA14dXiNet3TnK9170nhBBCZEZvdB+TXqd/4yQthBBCCMu9UaKOfXa1thBCCCFSl0VD3w4O9hzc9eI3pxUKTK8VivinidZp2DV1IhRCCCGyMIsSdcuvfFM7DiGEEEIkwqJEHRhk2WPOhBBCCJGy3v6h2EIIIYRINZKohRBCiAxMErUQQgiRgVn1oxy2trbUrVOFvN65sLEx/4Wq31b+naKBCSGEEMLKRP3DiJ6ULlWEs+evYTQaTctf/jENIYQQQqQcqxJ11Spl6NBtFMEhYakUjhBCCCFeZtUc9ZMn4RheOpMWQgghROqy6ox67/7jzJk+lO27DhEW9tTsvTf5FS0hhBBCvJ5Vibpi+eI8fRpJzffLmy2Pi4uTRC2EEEKkAqsSdd/BU1MrDiGEEEIkwqpErVar+OjD6ri7uaB4dneWna0dPj5ejB67IDXiE0IIIbI0qxL19yN6kT9vbsKfRpLN0YEHjx7zftWy/G/Hv6kVnxBCCJGlWZWoq1QsRav2vuTK5UaPLl8w8od51KxegTYtGqZWfEIIIUSWZtXtWTExGsKfRhJw9xFFC+cH4NCRsxQulC9VghNCCCGyOqsStf/NANp/+SkGg5GIiCgqVyxJ6ZKFMRpjUys+IYQQIkuzKlHPXrCa96qWwc01Owt//YMp4/qxYNYIlq7YlErhCSGEEFmbVXPU9+4Hmm7RehwcSqNm36JS2hGj0aZKcEIIIURWZ/XPXNarU5UZkwaydNEYcrg48XX3VqiUytSITQghhMjyrErUbVs1olunZhzwO423Vy50Oj2FCuZhUP8OqRWfEEIIkaVZNfT9RdP6fDtoMoFBT/imZ2siIqMZ+cM81iyfYlWjKqWSQf07ULd2FbRaHWvW72Dt+h2Jlq3+Xnl6dW1Bnjy5ePDwMb8u/RO/I2etak8IIYR4V1l1Rq1SKQl/Ghn/4tlvUOv0Bquv+u7Tqw0lihWg35CpzJyzgq4dmlK3TpUE5QoXysukMd+ydcdBOvf8ns1b9zPhh28pIreDCSGEyCKsStSHjpzlu2E9yeOVE4Ds2bMxsG97jhw/b3EdarWKJo3rMGfBaq773+HgodOsWreNFk0bJCj70YfVOXX2Chv+2sP9B0Fs3LyX02ev8GHdataELYQQQryzrErUc39aTfjTSFYsnoCTkyOb1s3GztaGuQtXW1xHkUL5sbWz5cIlf9Oy8xf9KV2yEIrnDxB/ZvsuPxb9uj5BHU7ZHKwJWwghhHhnWTVHrdHomDZrGdNmLSOHizPhTyOJezYEbikP9xyEh0diMBhNy56EhmNvr8IluxNh4RGm5XcCHpqtW9DHm8qVSrFpy75kWlEk876lFK/8nVllxu1L630nfeXdlh77L7N+ls/J9iXPsvxpUaJu9FGNZMtY+nvU9vYq9Hq92TK93gCAUpl0OC7ZnZg4pi8XLvrz7+EzSZZTO7iisLH6rrMEjGpnsHfEqHJEpXZ+6/oyqsy4bbEqR2ztDajUzjhkc0v19lQGZ5QKR5QGR+wz4ef5XGbdNmWsI0o7Ayr71O8vKrUWpb0Cpcou036ekIn7isqA0t4eldopRfpKTFSIReUsStSjhnYnOlrD9Rt3iIuLSzBEHRcXZ3Gi1un0KF+57/p5gtZodYmu4+qandnTfFHYKBg9dsFrz+I1MaGkxJGOThOBTutMrM4OG01E8iu8g1RqZ3SZcNv0umiM2mjQRBAT9STV29NpItDrnNEb7dDaZL7PE+K/eLWZsK9AfH/RG6PRxUUQY5e6/UWniUCvjUav06HVWDWg+c7I9H1Fa0CniSMmyj7N2rWop0yesZQPP6hK/nxe7P/3JHv2HeO6/503avBxcCguLk7Y2thgjI2/WtzdzQWNRktkZHSC8h4eOZg3YxgAfQdNMRsaT5p1w/HJ15NS9WUkLx/MZMbtS+t9J33l3ZYe+y8zfpZZoa88l3bbZ1Gi3rbTj207/XB2zkbd2lXo07MN7u4u7D8Yn7Rv33lgcYP+NwMwGoyULlWY8xfjLygrV6YYV67dSnCmrFar+HHyYGJj4+g7eCpPQsOt2DQhhBDi3WfV2EtERBRbth1gy7YDuLpmp27tKkz4oQ8Gg5HOPb+3qA6tVsf2XYfwHdCJidOXkNPDlbatGjFp+hIA3FxdiIyKRqfT07FdE/J45+LbQVNM7wFodTqiomKsCV0IIYR4J73RJIm3V04+/KAa9T6oinM2R/YdPGnV+nN/WoPvgI7MmzmMqKgYlizfxAG/UwBs2TCHidMWs22nH3VrV0Gttmfxwh/M1t+204+J0xa/SehCCCHEO8XiRP08OX9Ytyo5PVw58O8p5i9ay5lzV61uVKvVMWHqYiZMTZhsa9bvbPp3uy4jrK5bCCGEyEwsStRLFv6Ap6c7B/1Os+DndZw6cyW14xJCCCEEFibqYkV9AGjSuA6ffVLH7D2FIv6x33Uadk356IQQQogszqJE3fIr39SOQwghhBCJsChRBwZZ9vQUIYQQQqSst3/WphBCCCFSjSRqIYQQIgOTRC2EEEJkYJKohRBCiAxMErUQQgiRgUmiFkIIITIwSdRCCCFEBiaJWgghhMjAJFELIYQQGZgkaiGEECIDk0QthBBCZGCSqIUQQogMTBK1EEIIkYFJohZCCCEyMEnUQgghRAYmiVoIIYTIwCRRCyGEEBmYJGohhBAiA5NELYQQQmRgkqiFEEKIDEwStRBCCJGB2aVHoyqlkkH9O1C3dhW0Wh1r1u9g7fodiZYtWiQ/vgM6UbhgXm7dvs/02cu55n8njSMWQggh0ke6nFH36dWGEsUK0G/IVGbOWUHXDk2pW6dKgnJqtYoZkwZx7sJ1uvYew4XLN5g+aRBqtSodohZCCCHSXponarVaRZPGdZizYDXX/e9w8NBpVq3bRoumDRKUrV/3PbRaHQt+XsedgIfMWbCa6OgYPvygWlqHLYQQQqSLNE/URQrlx9bOlguX/E3Lzl/0p3TJQigUCrOypUsW5vxFf7NlFy7doHSpwmkSqxBCCJHe0nyO2sM9B+HhkRgMRtOyJ6Hh2NurcMnuRFh4hGm5u3sObt2+b7b+k9BwChXIm0wrimTet5QC7OzQlsicBwaxKkf0uuj0DiPl2dnxog+kVF9IjgK7ODtKaTJnX1HGZtK+AtjFpX1/sbOLo1SJiOQLvoOUKkPm7St2cS+9Som+Epd8EdIhUdvbq9Dr9WbL9HoDAEqleTjqJMq+Wi4hyzb+dRyOnMbhyOm3rkdkfn5OpwHpKyJ5fkfc4Yh7eochUsTb5xlLpfnQt06nR6lUmi17nng1Wp1FZbWvlBNCCCEyqzRP1I+DQ3FxccLW5kXT7m4uaDRaIiOjE5R1d3MxW+bu5kLwk7C0CFUIIYRId2meqP1vBmA0GM0uCCtXphhXrt0iLs58KOHSlZuUKV3EbFnZ0kW5dPlmmsQqhBBCpLc0T9RarY7tuw7hO6ATJYoXpHbNSrRt1Yj1G3cD4ObqgkoVP9y97+AJnLM50r9POwr4eNO/TzvUanv+OXA8rcMWQggh0oWiYMmGaTcj/oy9vQrfAR35oHYVoqJiWL1uO39s3AXAob3LmDhtMdt2+gFQsnhBfAd2okB+b278d5fps5fjfyMgrUMWQggh0kW6JGoRL7enB/N/HE7Lr4Ywamh3zpy7ajpASS8lixekbp0q/PTr+gTvZZQYRUL29iq6d25OreoV0Or0aLU6fl+7jX8PvbgavWfXFrRs1oBmbQYSHaNJx2hfqFm9AnnzeLJuw870DuWdllL7v/HHtej79ZcEBj0BBdirlJw6c4V5i9ai1eqoVb0CJYoXZPGyvxJd3yu3B53af86UGUspUawAzZrUY8rM35g3czhLV2zi4aNg1i6fwu07D57d/WrL+Qv+zJ6/Cp1ez6ih3alUoSQREVFm9fbznUrhgvmYNnEA9+8HAWBjo8DBQc2qtdvwO3KGGZMGAeD27LqmJ0/CARg6ejZBj5+83QecztLlWd8i4ypYIA9uri7JFxQZyuSxfXnw8DEduo/GYDCSN48n0ycNRKVSsnffMRQKBR83qM6JU5doWL86m7buS++QAShetEB6h5AppOT+9ztylonTFgPxiXT0sB74DujIhKmL8TtyFr8jZ5NcN7enB3m8cgFw9fptpsz8LUGZ4JAwOvf63vR6yrh+fPpJbf76+x8Aliz/K8mTgavXbtN38BTT66KF87N44ffs+ueIqc6uHZsBsHTFpiTjfNdIok4jtjY2DHn24yI5cjhz734gc39ak+x6jT+uRc33K+Dm5kKunG5s3X6A7M5OVKoYf9Q5aPhM3FxdmDF5IAEBD8mfz4ugx08YM2kRT59GsWHVDHbsPkS1ymXInj0bP877neMnL5LDxZmhAzvh6ekBxHfqC5f86dapOY4O9nTp8Dm/rfw7QTw13i9Pi6b1UamULPv9b/buP46jo5oRQ7qRK6crHu45OH/Rn7GTfsYluxNjR3+Ns7MTxMXx2+9/43f4DHm8cjJkQCdcXJzQ6w3M+2ktFy/fSPHPPKsoV6YoBXy8GTp6tulBQvfuBzJ3wWr69WnH3n3HeK9qGUJCwtn8v/1806N1kl/UXTs2wyu3BwV9vHFzc+H3tdsoWbwgpUsW5uate3w3bgEVy5egR5fmaLV6PHO5c83/NpOnL0Wn17N1w1x27T1C+bLFUCrtmDhtMdf87yS6z6NjNDRrUg+Ax8FP2LLtYJp9ZplJSu7/VxkMRub9tIaNa2Yye/5q6tSqRMXyJZg4bTFdOnxO3TpVMRqNXLl2i+mzljO4Xwc8Pd3xHdCJPfuO0bVjM7PE+iqVUom9vYrQ0KdvtO25c3sQo9GZnsWRWUmiTiNlShchNjaWnn3HAzBv5jDer1bOonVLlihI+26jcHbKxsY1Mxk4bAZzFq5m3sxhVKtShhs371LQJw+z5v3OqTNX+PbrL+ne+Qt+nLsSgMioGHr2HU+dmpXo1a0lx09epH+fduzcc4QDfqdwye7EL/O/o1e/CSxZ/hcVy5dINEkDqNX29Px2PK6u2Vny0xjOnLtGpQol+O/WXb4btwBbW1tWLZ1I8aI+lCtbjBv/3WP+orX45Pfi808/wO/wGUYN68Hchau5ev02ebxyMnv6UL7sOAxjbGzKfNhZTMkShbjuf8fsaX8AZ85fJX/e3Dg7Z+PTRnXYe+A4p05fxsM9B6VLFU7y7olCBfPSs884ypYpytwZw+jYYzR37z5i1W+TKFIoHwAliheiS6/vuRPwkPHf96Fl8was/mM7rq7ZOX/Rn7k/raFNi4Z0bNeEUWPnJ7nPN22JTxiSpN9cSu//V4U8CSciIpp8eT1Ny7Jlc+CLpvVp2noAcXEwuF8HcuV0Y+bclXTt2Izps5dTsXyJROvzcM/Bsp/HgQI8c7nz6FEwR0+cN73frVNzWn/R0PT6xn93mTD1VwBKFC/Asp/HYa9Wkd05G2fPXWPA0OmSqEXKOHfhOk+fRvJF0/r45PMiX97cOKivWrxudLSG6Oj4eaVTpy8D8CgwhOzO2QC4/yCIU2euALB9lx9jRn5tWv/w0XNA/K1xz8tXq1KGggXy0KVDUyB+iKtAfu9kY9m5+zDG2FiCQ8K4dPkGpUoWYs++Y5QsXpDWXzTEx8cLFxdnHBzUnDl7lR+nDsE7twdHjl9gyfJNOKjtKVWiEMMHdzXVaWdni6enOw8ePrbo8xCWUdvbA/Gfb833yzNn4SqMsbHsO3iC5k0+TPKL+vjJCxhjY3kUGELIk7D4+UTgcXAYzs/6z4WL/twJeAjAzt2H+Pyzuqz+Yztg3t9q1aj42n0uUs+b7v+kaHUvnhIZFRXD7TsPWLzwBw4dOcufm/YQ9PgJebxzJVvPy0Pfdna2jPLtzqC+HZg0fQlg2dC3nZ0t3w/vid5g5Oq1W1Ztx7tIEnUaqVW9Aj26tmDdhp38b+e/uLg4ERgUkqBcs8/qmYYDN23Zh06vT3CknNiZp9H4ooyNwsasjO7Zf7C4OEw/fGJjY0PfwVNNF224u7kQFhaBt1dO03rDB3ehRLGCAEyZuRSA2JfbVigwGoy0bNaAD+tWZdOW/Zz46xKFCuRFoVBw47+7fNlpGNWrlaPG+xVo26oRPb4dh06nN5uj8vDIQUhIePIfokjU1Wu3aNOiIXZ2thgMRlyyO/E0IooypYvwKDCYBvXeA4WCRXNGA/FP93N2cmTOwtW0adGQWtUrArB42UYA9PoXfcloTHyU4+X+prCxIfalcjr9i/4GCmxsbWSfp6KU3v+v8vDIgaOjmvsPgihRrIBped/BUylbugjvVy3Hj1OHMHbSz1bHbjAY2XvgOF93a2n1elNnLWPt8il83KAGO/cctrrtd4kk6jRSpVJp9u47xradfni456BCueKcfHZm/LJNW/eZzR81/riWRfXn8fakeFEfrvnfoXGjWpw4eem15U+fucIXn3/I8lVbyJvHk1/nf0fbLiMwGI3Y2sbfXp/YhSD1673H3v3H8czlTsniBZk68zeaNanHX3/vY/c/Ryno403RwvmxsVHQ/stPcVDb8+uyjRw+eo6Na2aiQMHd+4E0bFCdXXuOUK5MUSb88C1ftB2U4IBEWObchevcuv2A/t+0Y87C1XxQuzJNP6tHduds/PrbRr5s2YhpP/7Gjt0vvsx+WzSWTxvVZvGyv8yu4C1m4cVdZUsXIVdONx4Hh/LJRzU4firp/hYVFZPkPjcajabnJog3k5L7/9XvG5VSSZ8ebdi641+zRzfn9vRg6oT+9OgzjguXbuDhkYMihfJx/cYd0/eHpSpXLMnV67et3u6oqBiWLNtE7x6t2P/vyUz9aGlJ1Gnk720HGDOyF/XqVEWvN3Dx8g0qVyiZYvU/fRpJ5w5NyZfHk/9u32fKjKWvLT9r/u/4DujE8l/HY6NQMHnGUsLCIrh0+SZdOzSlT882LPhlXYL1oqNjWLpoDHa2tkybtZzwp5Gs+3MXvgM68mWrRsTEaLhwyR9vr1xs2XaAMaO+Zvmv4zEajSxZvonIqGjGTvoZ3wEd+apNY2KNsYweO1+S9Fsa8cNcenT5gpVLJmI0GNFotDx4+JhGH9WgYAFv9u43f0jQ2g076dzhc9au3/FG7QU9DmWkbzc8c7lx8vRlNm95/cVJSe3z02ev8P2IXjx91o/Em0nJ/V+regWW/TyOuLg4bG1tOH7qEosWm9+u+SgwmL37jvPborFotDoCA4PZttMPWztbnLI5MGbU12zeuj/RWE1z1IBKpeTu/UDTiB0knKMGmDZrWaJ1bdl2gFZffMRXbRpnqqu8XyX3UWcCL9+PLcTLqlUuzdkL103THymhYvkSyV7NKzKG1Nj/Iu3JGbUQmdjrhqRF5if7P3OQM2ohhBAiA0vzH+UQQgghhOUkUQshhBAZmCRqIYQQIgOTRC2EEEJkYJKohRBCiAxMErUQQgiRgUmiFkIIITIwSdRCCCFEBvZ/3IBHha7Hhg0AAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x275 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate means and standard errors\n",
    "mean_rr_by_type = {}\n",
    "se_rr_by_type = {}\n",
    "friedman_results = {}\n",
    "\n",
    "for ptype in all_types:\n",
    "    mean_rr_by_type[ptype] = []\n",
    "    se_rr_by_type[ptype] = []\n",
    "    friedman_data = []\n",
    "    for model_name in model_order:  # Use the defined model order\n",
    "        rr_by_type = results_rr_by_type[model_name][ptype]\n",
    "        mean_rr = np.mean(rr_by_type)\n",
    "        se_rr = np.std(rr_by_type) / np.sqrt(len(rr_by_type))  # Standard Error of the mean\n",
    "        mean_rr_by_type[ptype].append(mean_rr)\n",
    "        se_rr_by_type[ptype].append(se_rr)\n",
    "        friedman_data.append(rr_by_type)\n",
    "\n",
    "    # Perform Friedman Test\n",
    "    friedman_result = friedmanchisquare(*friedman_data)\n",
    "    friedman_results[ptype] = friedman_result\n",
    "\n",
    "    # If Friedman Test is significant, perform Nemenyi post hoc test\n",
    "    if friedman_result.pvalue < 0.05:\n",
    "        # Combine data into a DataFrame for the Nemenyi test\n",
    "        data = {model_name: rr_by_type for model_name, rr_by_type in zip(model_order, friedman_data)}\n",
    "        data_df = pd.DataFrame(data)\n",
    "        nemenyi_result = sp.posthoc_nemenyi_friedman(data_df)\n",
    "\n",
    "# Colors for different models\n",
    "colors = ['cyan', 'magenta', 'yellow']\n",
    "\n",
    "# New labels for the models\n",
    "new_model_order = ['all-mpnet-base', 'QA-mpnet', 'QA-DistilBERT']\n",
    "\n",
    "for ptype in all_types:\n",
    "    # Show plot only if Friedman test indicates significant difference\n",
    "    if friedman_results[ptype].pvalue < 0.05:\n",
    "        # Create a new figure for each question type\n",
    "        fig, ax = plt.subplots(figsize=(5, 2.75)) \n",
    "\n",
    "        # Data for plotting\n",
    "        means = mean_rr_by_type[ptype]\n",
    "        errors = se_rr_by_type[ptype]\n",
    "        x = np.arange(len(means))\n",
    "\n",
    "        # Creating bar chart\n",
    "        bars = ax.bar(x, means, yerr=errors, color=colors, capsize=7, alpha=0.3, linewidth=2, ecolor='white')\n",
    "        for bar, color in zip(bars, colors):\n",
    "            bar.set_edgecolor(color)\n",
    "        ax.set_xticks(x)\n",
    "        ax.set_xticklabels(new_model_order, fontsize=8)  # Set new model order labels\n",
    "        ax.set_title(f'MRR and SE for \"{ptype}\" type questions', fontsize=10)\n",
    "        ax.set_ylim(0, 1)\n",
    "        ax.set_ylabel('Mean Reciprocal Rank', fontsize=9)\n",
    "        ax.set_yticks(np.linspace(0, 1, 6))\n",
    "\n",
    "        mplcyberpunk.make_lines_glow(ax, alpha_line=0.6)\n",
    "\n",
    "        plt.tight_layout(rect=[0, 0.1, 1, 1])  # Adjust the bottom margin\n",
    "        plt.savefig(f'../images/bar_chart_{ptype}.png', dpi=90, bbox_inches='tight')\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fine-tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remove duplicates and reset index\n",
    "unique_corpus = fine_tuning_dataset.drop_duplicates(subset=['content']).reset_index(drop=True)\n",
    "\n",
    "# Create document IDs\n",
    "document_ids = {doc: idx for idx, doc in enumerate(unique_corpus['content'])}\n",
    "\n",
    "# Create queries and query-document IDs\n",
    "queries = {idx: row['question'] for idx, row in fine_tuning_dataset.iterrows()}\n",
    "query_doc_ids = {idx: document_ids[row['content']] for idx, row in fine_tuning_dataset.iterrows() if row['labels'] == 1.0}\n",
    "\n",
    "# Split the data into training, validation, and test sets\n",
    "train_queries, temp_queries = train_test_split(list(queries.keys()), test_size=0.3, random_state=42)\n",
    "val_queries, test_queries = train_test_split(temp_queries, test_size=0.5, random_state=42)\n",
    "\n",
    "# Create relevant documents for validation and test sets\n",
    "relevant_docs_val = {q_id: {query_doc_ids[q_id]} for q_id in val_queries if q_id in query_doc_ids}\n",
    "relevant_docs_test = {q_id: {query_doc_ids[q_id]} for q_id in test_queries if q_id in query_doc_ids}\n",
    "\n",
    "# Create corpus dictionary\n",
    "corpus = {doc_id: doc for doc, doc_id in document_ids.items()}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class QADataset(Dataset):\n",
    "    def __init__(self, examples):\n",
    "        self.examples = examples\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.examples)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        return self.examples[idx]\n",
    "\n",
    "def get_content(query_id):\n",
    "    doc_id = query_doc_ids.get(query_id, None)\n",
    "    if doc_id is not None:\n",
    "        return corpus[doc_id]\n",
    "    else:\n",
    "        return \"\"\n",
    "\n",
    "train_examples = [InputExample(texts=[queries[q_id], get_content(q_id)]) for q_id in train_queries]\n",
    "val_examples = [InputExample(texts=[queries[q_id], get_content(q_id)]) for q_id in val_queries]\n",
    "test_examples = [InputExample(texts=[queries[q_id], get_content(q_id)]) for q_id in test_queries]\n",
    "\n",
    "train_dataset = QADataset(train_examples)\n",
    "val_dataset = QADataset(val_examples)\n",
    "test_dataset = QADataset(test_examples)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LoggingInformationRetrievalEvaluator(InformationRetrievalEvaluator):\n",
    "    def __init__(self, *args, **kwargs):\n",
    "        super().__init__(*args, **kwargs)\n",
    "        self.logger = logging.getLogger('training_logger')\n",
    "        self.last_logged_epoch = -1\n",
    "\n",
    "    def __call__(self, model, output_path=None, epoch=-1, steps=-1):\n",
    "        if epoch != self.last_logged_epoch:\n",
    "            results = super().__call__(model, output_path, epoch, steps)\n",
    "            log_message = f\"Epoch: {epoch}, Steps: {steps}\\n\"\n",
    "            log_message += f\"Information Retrieval Evaluation of the model on the {self.name} dataset:\\n\"\n",
    "            log_message += f\"Queries: {len(self.queries)}\\nCorpus: {len(self.corpus)}\\n\\n\"\n",
    "            \n",
    "            # Log only the key metrics\n",
    "            key_metrics = [\n",
    "                'validation-ir-eval_cosine_accuracy@1',\n",
    "                'validation-ir-eval_cosine_accuracy@5',\n",
    "                'validation-ir-eval_cosine_recall@1',\n",
    "                'validation-ir-eval_cosine_recall@5',\n",
    "                'validation-ir-eval_cosine_ndcg@10',\n",
    "                'validation-ir-eval_cosine_mrr@10',\n",
    "                'validation-ir-eval_cosine_map@100'\n",
    "            ]\n",
    "            for metric in key_metrics:\n",
    "                if metric in results:\n",
    "                    log_message += f\"{metric}: {results[metric]}\\n\"\n",
    "            \n",
    "            self.logger.info(log_message)\n",
    "            self.last_logged_epoch = epoch\n",
    "            return results\n",
    "        else:\n",
    "            return super().__call__(model, output_path, epoch, steps)\n",
    "\n",
    "# Setup logging configuration\n",
    "def setup_logging(log_file_path):\n",
    "    logger = logging.getLogger('training_logger')\n",
    "    logger.setLevel(logging.INFO)\n",
    "    fh = logging.FileHandler(log_file_path)\n",
    "    fh.setLevel(logging.INFO)\n",
    "    formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')\n",
    "    fh.setFormatter(formatter)\n",
    "    logger.addHandler(fh)\n",
    "    return logger, fh"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bayesian Optimization for Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "|   iter    |  target   | learni... | num_ep... | per_gp... | warmup... | weight... |\n",
      "-------------------------------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "158ab83e2f654a65afed81c2e4dbbbfb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/56 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 181.3778, 'train_samples_per_second': 14.864, 'train_steps_per_second': 0.309, 'train_loss': 0.3011243002755301, 'epoch': 4.0}\n",
      "| \u001b[0m1        \u001b[0m | \u001b[0m0.9345   \u001b[0m | \u001b[0m2.498e-05\u001b[0m | \u001b[0m4.852    \u001b[0m | \u001b[0m51.14    \u001b[0m | \u001b[0m299.3    \u001b[0m | \u001b[0m0.04681  \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ceefda625dd64e5bbe1940c1b30d6d44",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/24 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 90.2361, 'train_samples_per_second': 14.939, 'train_steps_per_second': 0.266, 'train_loss': 0.572495182355245, 'epoch': 2.0}\n",
      "| \u001b[0m2        \u001b[0m | \u001b[0m0.9076   \u001b[0m | \u001b[0m1.624e-05\u001b[0m | \u001b[0m2.174    \u001b[0m | \u001b[0m57.58    \u001b[0m | \u001b[0m300.6    \u001b[0m | \u001b[0m0.2124   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1bc39bab0bc24178a406df433642feeb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/52 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 182.8944, 'train_samples_per_second': 14.741, 'train_steps_per_second': 0.284, 'train_loss': 0.36142991139338565, 'epoch': 4.0}\n",
      "| \u001b[95m3        \u001b[0m | \u001b[95m0.9347   \u001b[0m | \u001b[95m1.082e-05\u001b[0m | \u001b[95m4.91     \u001b[0m | \u001b[95m55.96    \u001b[0m | \u001b[95m106.2    \u001b[0m | \u001b[95m0.05455  \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f7e8d5efeed644eaaf6c383aca22c04f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/34 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 83.3865, 'train_samples_per_second': 16.166, 'train_steps_per_second': 0.408, 'train_loss': 0.3781729305491728, 'epoch': 2.0}\n",
      "| \u001b[0m4        \u001b[0m | \u001b[0m0.9261   \u001b[0m | \u001b[0m1.734e-05\u001b[0m | \u001b[0m2.913    \u001b[0m | \u001b[0m41.19    \u001b[0m | \u001b[0m216.0    \u001b[0m | \u001b[0m0.08737  \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e56ba28e0df34702b814c1406c1906b8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/46 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 84.4542, 'train_samples_per_second': 15.961, 'train_steps_per_second': 0.545, 'train_loss': 0.21850511302118716, 'epoch': 2.0}\n",
      "| \u001b[0m5        \u001b[0m | \u001b[0m0.9329   \u001b[0m | \u001b[0m3.447e-05\u001b[0m | \u001b[0m2.418    \u001b[0m | \u001b[0m30.02    \u001b[0m | \u001b[0m183.2    \u001b[0m | \u001b[0m0.1368   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4b46d8bd04314c1789432143f629f335",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/46 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 81.9456, 'train_samples_per_second': 16.45, 'train_steps_per_second': 0.561, 'train_loss': 0.18444038474041483, 'epoch': 2.0}\n",
      "| \u001b[95m6        \u001b[0m | \u001b[95m0.9372   \u001b[0m | \u001b[95m3.635e-05\u001b[0m | \u001b[95m2.227    \u001b[0m | \u001b[95m30.1     \u001b[0m | \u001b[95m183.1    \u001b[0m | \u001b[95m0.1556   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "da123c3eba5f4cfda4f539bd58d13897",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/52 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 176.5368, 'train_samples_per_second': 15.272, 'train_steps_per_second': 0.295, 'train_loss': 0.220610471872183, 'epoch': 4.0}\n",
      "| \u001b[95m7        \u001b[0m | \u001b[95m0.9453   \u001b[0m | \u001b[95m4.363e-05\u001b[0m | \u001b[95m4.582    \u001b[0m | \u001b[95m56.09    \u001b[0m | \u001b[95m107.4    \u001b[0m | \u001b[95m0.1278   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6dcf359c46014e57b829db15988b9358",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/52 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 171.5594, 'train_samples_per_second': 15.715, 'train_steps_per_second': 0.303, 'train_loss': 0.2598581680884728, 'epoch': 4.0}\n",
      "| \u001b[95m8        \u001b[0m | \u001b[95m0.9551   \u001b[0m | \u001b[95m3.632e-05\u001b[0m | \u001b[95m4.612    \u001b[0m | \u001b[95m56.87    \u001b[0m | \u001b[95m106.4    \u001b[0m | \u001b[95m0.2091   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "23c26523582e4b5eb37fe7b176f12f07",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/48 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 174.0721, 'train_samples_per_second': 15.488, 'train_steps_per_second': 0.276, 'train_loss': 0.2994123299916585, 'epoch': 4.0}\n",
      "| \u001b[0m9        \u001b[0m | \u001b[0m0.9491   \u001b[0m | \u001b[0m3.373e-05\u001b[0m | \u001b[0m4.506    \u001b[0m | \u001b[0m57.85    \u001b[0m | \u001b[0m106.2    \u001b[0m | \u001b[0m0.2578   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9269f5425d7945d281ee418faff78d78",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/39 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 130.2781, 'train_samples_per_second': 15.521, 'train_steps_per_second': 0.299, 'train_loss': 0.32474605853740984, 'epoch': 3.0}\n",
      "| \u001b[0m10       \u001b[0m | \u001b[0m0.949    \u001b[0m | \u001b[0m2.979e-05\u001b[0m | \u001b[0m3.522    \u001b[0m | \u001b[0m56.55    \u001b[0m | \u001b[0m106.4    \u001b[0m | \u001b[0m0.179    \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e1009bd5e39b4038ae0ac3600a30716e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/48 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 181.3665, 'train_samples_per_second': 14.865, 'train_steps_per_second': 0.265, 'train_loss': 0.25574855009714764, 'epoch': 4.0}\n",
      "| \u001b[0m11       \u001b[0m | \u001b[0m0.9546   \u001b[0m | \u001b[0m4.823e-05\u001b[0m | \u001b[0m4.612    \u001b[0m | \u001b[0m57.65    \u001b[0m | \u001b[0m108.5    \u001b[0m | \u001b[0m0.1361   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "af2e13b47f9b4eb28adf2e46491762b0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/39 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 130.5205, 'train_samples_per_second': 15.492, 'train_steps_per_second': 0.299, 'train_loss': 0.31038594857240337, 'epoch': 3.0}\n",
      "| \u001b[0m12       \u001b[0m | \u001b[0m0.9486   \u001b[0m | \u001b[0m3.723e-05\u001b[0m | \u001b[0m3.901    \u001b[0m | \u001b[0m56.84    \u001b[0m | \u001b[0m108.0    \u001b[0m | \u001b[0m0.1621   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fa06a1eeca844480b1b1947ae9c2484a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/36 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 134.8913, 'train_samples_per_second': 14.99, 'train_steps_per_second': 0.267, 'train_loss': 0.3753413094414605, 'epoch': 3.0}\n",
      "| \u001b[0m13       \u001b[0m | \u001b[0m0.9352   \u001b[0m | \u001b[0m2.011e-05\u001b[0m | \u001b[0m3.35     \u001b[0m | \u001b[0m59.15    \u001b[0m | \u001b[0m107.8    \u001b[0m | \u001b[0m0.1827   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bce4edf62320413da5e231ae5396caff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/65 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 218.8083, 'train_samples_per_second': 15.402, 'train_steps_per_second': 0.297, 'train_loss': 0.21904801588792067, 'epoch': 5.0}\n",
      "| \u001b[0m14       \u001b[0m | \u001b[0m0.9485   \u001b[0m | \u001b[0m5e-05    \u001b[0m | \u001b[0m5.0      \u001b[0m | \u001b[0m56.93    \u001b[0m | \u001b[0m109.6    \u001b[0m | \u001b[0m0.3      \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5cdd6cc4eda0420fb9cdb1f1ea9c40f5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/24 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 91.5147, 'train_samples_per_second': 14.73, 'train_steps_per_second': 0.262, 'train_loss': 0.4352372884750366, 'epoch': 2.0}\n",
      "| \u001b[0m15       \u001b[0m | \u001b[0m0.9209   \u001b[0m | \u001b[0m2.323e-05\u001b[0m | \u001b[0m2.93     \u001b[0m | \u001b[0m58.5     \u001b[0m | \u001b[0m111.1    \u001b[0m | \u001b[0m0.2766   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6c0ec301ea5d4c67be6d84b41ffb1e4f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/44 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 83.8528, 'train_samples_per_second': 16.076, 'train_steps_per_second': 0.525, 'train_loss': 0.23672565546902744, 'epoch': 2.0}\n",
      "| \u001b[0m16       \u001b[0m | \u001b[0m0.922    \u001b[0m | \u001b[0m2.153e-05\u001b[0m | \u001b[0m2.04     \u001b[0m | \u001b[0m32.08    \u001b[0m | \u001b[0m181.8    \u001b[0m | \u001b[0m0.253    \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e7c198cf0b364d47bbd5814cae06cb57",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/24 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 89.8087, 'train_samples_per_second': 15.01, 'train_steps_per_second': 0.267, 'train_loss': 0.4640875260035197, 'epoch': 2.0}\n",
      "| \u001b[0m17       \u001b[0m | \u001b[0m0.9222   \u001b[0m | \u001b[0m2.291e-05\u001b[0m | \u001b[0m2.033    \u001b[0m | \u001b[0m57.93    \u001b[0m | \u001b[0m105.0    \u001b[0m | \u001b[0m0.1316   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fda268e1e3e746f69601a6aec0dee3c9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/48 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 176.9714, 'train_samples_per_second': 15.234, 'train_steps_per_second': 0.271, 'train_loss': 0.3607181708017985, 'epoch': 4.0}\n",
      "| \u001b[0m18       \u001b[0m | \u001b[0m0.9403   \u001b[0m | \u001b[0m1.737e-05\u001b[0m | \u001b[0m4.675    \u001b[0m | \u001b[0m57.35    \u001b[0m | \u001b[0m107.0    \u001b[0m | \u001b[0m0.02736  \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "267984da7a7d4a75bb797e01fc7c28c2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/52 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 183.6405, 'train_samples_per_second': 14.681, 'train_steps_per_second': 0.283, 'train_loss': 0.30127299748934233, 'epoch': 4.0}\n",
      "| \u001b[0m19       \u001b[0m | \u001b[0m0.9399   \u001b[0m | \u001b[0m1.716e-05\u001b[0m | \u001b[0m4.825    \u001b[0m | \u001b[0m55.6     \u001b[0m | \u001b[0m108.7    \u001b[0m | \u001b[0m0.2361   \u001b[0m |\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "50924fb5cf1048649a6dea91f66ab91e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/36 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 132.111, 'train_samples_per_second': 15.305, 'train_steps_per_second': 0.272, 'train_loss': 0.360516177283393, 'epoch': 3.0}\n",
      "| \u001b[0m20       \u001b[0m | \u001b[0m0.941    \u001b[0m | \u001b[0m2.8e-05  \u001b[0m | \u001b[0m3.558    \u001b[0m | \u001b[0m57.23    \u001b[0m | \u001b[0m105.8    \u001b[0m | \u001b[0m0.05252  \u001b[0m |\n",
      "=====================================================================================\n",
      "Best hyperparameters:  {'learning_rate': 3.631692566303126e-05, 'num_epochs': 4.61248579441536, 'per_gpu_batch_size': 56.871775650047006, 'warmup_steps': 106.44852169800234, 'weight_decay': 0.20906624405054877}\n"
     ]
    }
   ],
   "source": [
    "# Function to remove and close all handlers from the logger\n",
    "def close_logger(logger):\n",
    "    handlers = logger.handlers[:]\n",
    "    for handler in handlers:\n",
    "        handler.close()\n",
    "        logger.removeHandler(handler)\n",
    "\n",
    "# Delete the existing training log file if it exists\n",
    "log_file_path = '../models/bayes_opt/training.log'\n",
    "if os.path.exists(log_file_path):\n",
    "    close_logger(logging.getLogger('training_logger'))\n",
    "    os.remove(log_file_path)\n",
    "\n",
    "# Define the objective function\n",
    "def objective(per_gpu_batch_size, weight_decay, learning_rate, warmup_steps, num_epochs):\n",
    "    per_gpu_batch_size = int(per_gpu_batch_size)\n",
    "    warmup_steps = int(warmup_steps)\n",
    "    num_epochs = int(num_epochs)\n",
    "\n",
    "    # Load the model\n",
    "    model = SentenceTransformer('multi-qa-distilbert-cos-v1')\n",
    "\n",
    "    # Define the loss function\n",
    "    loss = losses.MultipleNegativesRankingLoss(model=model)\n",
    "\n",
    "    # Setup logging\n",
    "    logger, fh = setup_logging('../models/bayes_opt/training.log')\n",
    "\n",
    "    # Instantiate the custom evaluators\n",
    "    ir_evaluator_val = LoggingInformationRetrievalEvaluator(\n",
    "        queries={q_id: queries[q_id] for q_id in val_queries},\n",
    "        corpus=corpus,\n",
    "        relevant_docs=relevant_docs_val,\n",
    "        name=\"validation-ir-eval\"\n",
    "    )\n",
    "\n",
    "    # Create DataLoader with the suggested batch size\n",
    "    train_dataloader = DataLoader(train_dataset, batch_size=per_gpu_batch_size, shuffle=True)\n",
    "\n",
    "    # Train the model with the evaluator\n",
    "    model.fit(\n",
    "        train_objectives=[(train_dataloader, loss)],\n",
    "        evaluator=ir_evaluator_val,\n",
    "        epochs=num_epochs,\n",
    "        warmup_steps=warmup_steps,\n",
    "        evaluation_steps=500,\n",
    "        save_best_model=False,  # Do not save the model\n",
    "        show_progress_bar=True,\n",
    "        optimizer_params={'lr': learning_rate, 'weight_decay': weight_decay}\n",
    "    )\n",
    "\n",
    "    # Evaluate the model on the validation set\n",
    "    results = ir_evaluator_val(model)\n",
    "    \n",
    "    # Close the logger handlers to release the file\n",
    "    close_logger(logger)\n",
    "    \n",
    "    # Return the metric to be optimized, defaulting to 0 if 'validation-ir-eval_cosine_mrr@10' is not found\n",
    "    return results.get('validation-ir-eval_cosine_mrr@10', 0)\n",
    "\n",
    "# Define the search space\n",
    "pbounds = {\n",
    "    'per_gpu_batch_size': (16, 64),\n",
    "    'weight_decay': (0, 0.3),\n",
    "    'learning_rate': (1e-5, 5e-5),\n",
    "    'warmup_steps': (0, 500),\n",
    "    'num_epochs': (2, 5)\n",
    "}\n",
    "\n",
    "# Create a BayesianOptimization object\n",
    "optimizer = BayesianOptimization(\n",
    "    f=objective,\n",
    "    pbounds=pbounds,\n",
    "    random_state=42,\n",
    "    verbose=2\n",
    ")\n",
    "\n",
    "# Run the optimization with reduced number of trials\n",
    "optimizer.maximize(init_points=5, n_iter=15)\n",
    "\n",
    "# Print the best hyperparameters\n",
    "best_hyperparameters = optimizer.max['params']\n",
    "print(\"Best hyperparameters: \", best_hyperparameters)\n",
    "\n",
    "# Save the best hyperparameters to a file\n",
    "with open('../models/bayes_opt/best_hyperparameters.json', 'w') as f:\n",
    "    json.dump(best_hyperparameters, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load best hyperparameters \n",
    "file_path = '../models/bayes_opt/best_hyperparameters.json'\n",
    "with open(file_path, 'r') as file:\n",
    "    best_hyperparameters = json.load(file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fine-tuning with best hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "75ec9e21d6284d32a3723cd6bde427c3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/52 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 179.9444, 'train_samples_per_second': 14.982, 'train_steps_per_second': 0.289, 'train_loss': 0.24672205631549543, 'epoch': 4.0}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a611a23ca06945a98d7f2d74c77ef387",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Computing widget examples:   0%|          | 0/5 [00:00<?, ?example/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Final model saved to ../models/best-IR-model\n"
     ]
    }
   ],
   "source": [
    "# Use the best hyperparameter for batch size\n",
    "batch_size = int(best_hyperparameters['per_gpu_batch_size'])\n",
    "\n",
    "train_dataloader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
    "val_dataloader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False)\n",
    "test_dataloader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)\n",
    "\n",
    "# Train the final model with the best hyperparameters\n",
    "def train_final_model(best_hyperparameters):\n",
    "    per_gpu_batch_size = int(best_hyperparameters['per_gpu_batch_size'])\n",
    "    weight_decay = best_hyperparameters['weight_decay']\n",
    "    learning_rate = best_hyperparameters['learning_rate']\n",
    "    warmup_steps = int(best_hyperparameters['warmup_steps'])\n",
    "    num_epochs = int(best_hyperparameters['num_epochs'])\n",
    "\n",
    "    # Load the model\n",
    "    model = SentenceTransformer('multi-qa-distilbert-cos-v1')\n",
    "\n",
    "    # Define the loss function\n",
    "    loss = losses.MultipleNegativesRankingLoss(model=model)\n",
    "\n",
    "    # Setup logging\n",
    "    logger, fh = setup_logging('../models/best-IR-model/final_training.log')\n",
    "\n",
    "    # Instantiate the custom evaluators\n",
    "    ir_evaluator_val = LoggingInformationRetrievalEvaluator(\n",
    "        queries={q_id: queries[q_id] for q_id in val_queries},\n",
    "        corpus=corpus,\n",
    "        relevant_docs=relevant_docs_val,\n",
    "        name=\"validation-ir-eval\"\n",
    "    )\n",
    "\n",
    "    # Create DataLoader with the suggested batch size\n",
    "    train_dataloader = DataLoader(train_dataset, batch_size=per_gpu_batch_size, shuffle=True)\n",
    "\n",
    "    # Train the model with the evaluator\n",
    "    model.fit(\n",
    "        train_objectives=[(train_dataloader, loss)],\n",
    "        evaluator=ir_evaluator_val,\n",
    "        epochs=num_epochs,\n",
    "        warmup_steps=warmup_steps,\n",
    "        evaluation_steps=500,\n",
    "        save_best_model=True,  # Save the best model\n",
    "        show_progress_bar=True,\n",
    "        optimizer_params={'lr': learning_rate, 'weight_decay': weight_decay}\n",
    "    )\n",
    "\n",
    "    # Save the trained model\n",
    "    model_save_path = '../models/best-IR-model'\n",
    "    model.save(model_save_path)\n",
    "\n",
    "    # Close the logger handlers to release the file\n",
    "    close_logger(logger)\n",
    "\n",
    "    print(f\"Final model saved to {model_save_path}\")\n",
    "\n",
    "# Train the final model with the best hyperparameters\n",
    "train_final_model(best_hyperparameters)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x250 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x250 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x250 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Read the log file\n",
    "with open('../models/best-IR-model/final_training.log', 'r') as f:\n",
    "    log_contents = f.read()\n",
    "\n",
    "# Extract metrics using regular expressions\n",
    "epoch_pattern = r'Epoch: (\\d+\\.\\d+)'\n",
    "metric_pattern = r'(validation-ir-eval_cosine_(\\w+))@(\\d+): (\\d+\\.\\d+)'\n",
    "\n",
    "epochs = re.findall(epoch_pattern, log_contents)\n",
    "metrics = re.findall(metric_pattern, log_contents)\n",
    "\n",
    "# Create a dictionary to store the metric values for each epoch, grouped by metric type and top_k\n",
    "metric_values = {}\n",
    "for metric_name, metric_type, top_k, metric_value in metrics:\n",
    "    metric_values.setdefault(metric_type, {}).setdefault(top_k, []).append(float(metric_value))\n",
    "\n",
    "# Define colors and line styles for each metric and top_k value\n",
    "colors = {\n",
    "    'accuracy@1': 'cyan', 'accuracy@5': 'cyan',\n",
    "    'recall@1': 'yellow', 'recall@5': 'yellow',\n",
    "    'mrr@10': 'magenta', 'ndcg@10': 'green'\n",
    "}\n",
    "line_styles = {\n",
    "    'accuracy@1': '-', 'accuracy@5': '--',\n",
    "    'recall@1': '-', 'recall@5': '--',\n",
    "    'mrr@10': '-', 'ndcg@10': '--'\n",
    "}\n",
    "\n",
    "def plot_metrics(metric_types, title, filename, y_lim=(0.7, 1.05), y_ticks=np.arange(0.7, 1.05, 0.1), dpi=100):\n",
    "    # Adjust font sizes\n",
    "    plt.rcParams.update({'font.size': 8})  # Adjusting the global font size smaller for smaller plots\n",
    "\n",
    "    # Convert pixel dimensions to inches for figsize\n",
    "    width_in_inches = 400 / dpi\n",
    "    height_in_inches = 250 / dpi\n",
    "\n",
    "    plt.figure(figsize=(width_in_inches, height_in_inches))\n",
    "    for metric_type in metric_types:\n",
    "        for top_k, values in metric_values[metric_type].items():\n",
    "            plt.plot(range(1, len(epochs) + 1), values, color=colors[f'{metric_type}@{top_k}'],\n",
    "                     linestyle=line_styles[f'{metric_type}@{top_k}'], label=f'{metric_type}@{top_k}', alpha=0.7)\n",
    "    plt.title(title, pad=10)\n",
    "    plt.ylabel('Metric Value')\n",
    "    plt.ylim(*y_lim)\n",
    "    plt.xlim(1, len(epochs))\n",
    "    plt.yticks(y_ticks)\n",
    "    plt.grid(True)\n",
    "    mplcyberpunk.add_glow_effects()\n",
    "    handles, labels = plt.gca().get_legend_handles_labels()\n",
    "    unique_labels = list(set(labels))\n",
    "    unique_handles = [handles[labels.index(label)] for label in unique_labels]\n",
    "    plt.legend(unique_handles, unique_labels, bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.xticks(range(1, len(epochs) + 1, 1))\n",
    "    plt.tight_layout(rect=[0, 0.03, 1, 0.95])\n",
    "    plt.savefig(f'../images/{filename}', dpi=dpi)  \n",
    "    plt.show()\n",
    "\n",
    "# Plot accuracy, recall, NDCG and MRR\n",
    "plot_metrics(metric_types=['accuracy'], title='Accuracy', filename='val_accuracy.png')\n",
    "plot_metrics(metric_types=['recall'], title='Recall', filename='val_recall.png')\n",
    "plot_metrics(metric_types=['ndcg', 'mrr'], title='NDCG and MRR', filename='val_ndcg_mrr.png')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "You try to use a model that was created with version 3.0.0.dev0, however, your version is 3.0.0. This might cause unexpected behavior or errors. In that case, try to update to the latest version.\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Load your best model checkpoint\n",
    "best_model_path = '../models/best-IR-model'\n",
    "best_model = SentenceTransformer(best_model_path)\n",
    "\n",
    "# Load the base pre-fine-tuning model and mpnet models\n",
    "base_model = SentenceTransformer('multi-qa-distilbert-cos-v1')\n",
    "mpnet_model = SentenceTransformer('multi-qa-mpnet-base-dot-v1')\n",
    "all_mpnet_model = SentenceTransformer('all-mpnet-base-v2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the evaluator\n",
    "ir_evaluator = InformationRetrievalEvaluator(\n",
    "    queries={q_id: queries[q_id] for q_id in test_queries},\n",
    "    corpus=corpus,\n",
    "    relevant_docs=relevant_docs_test,\n",
    "    name=\"test-ir-eval\"\n",
    ")\n",
    "\n",
    "# Evaluate the models\n",
    "results_best_model = ir_evaluator(best_model)\n",
    "results_base_model = ir_evaluator(base_model)\n",
    "results_mpnet_model = ir_evaluator(mpnet_model)\n",
    "results_all_mpnet_model = ir_evaluator(all_mpnet_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 444.444x155.556 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbUAAACLCAYAAADmrQe0AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAAAxb0lEQVR4nO3ddVxV9x/H8de9l5JWEETsxK7p7O7EmG76szBnByoWdteUWbNwMmNzzq6pqFOxwe7CREJQ+nLh9wfzTgTkYlB+no8Hj8e898TnnnvHh/O953zfioIlGschhBBCZAHK9C5ACCGE+FykqQkhhMgypKkJIYTIMqSpCSGEyDKkqQkhhMgypKkJIYTIMqSpCSGEyDKkqQkhhMgypKllcna5rNO7hCzNwECfHNkt0mRf8l4K8en00ruAd5087E5kZBSxcXEQF0dcHNy558vKNVu5fPXOZ93XqGHdiYiM4ucVmz+4nMeaGaxc+yf/nLz4Wff/Vvbs5rRpUZdaNSpia5MDA319Al+FcPbcVX7fdpCnz14mu+7Afp0wNNBnoZvHZ69r62/z+XnlZo4eP5/qdf/evYLR43/C+9LNz15XWlv+0zg2bN7zUcchNVJ6L3X9vH5OP3zXlOpVyzN45Ow026cQnypDNTWAgcNncfP2QwAUCgXftW3E/JkjaN95JG9Cwz/bfub9tF6n5f7Xa/xn2+f7ypUpxtQJP3Lo6BlmzVvDQ99naDSx5LbLSaP6VVm5ZAJzF7lzPJmGamlhRkRE5BerT4CFhVma7Cel91LXz6sQX7sM19TeFRcXx869Rxk6sDP2uW3Ils2Isc5O3L3nS8UKJZmzYB0nvXzo49SO+nWqYGCgx5lzV1i8dKO2AdaoVp4+Pdthn9sGP79Alv7yO15nLjF+dG8iIiJZ6OZBoYJ5GD28OwXy2/P6dSjHTlxg+arfiY2NS3DGYmFuyqD+31OtSlli4+I4ffYybis28+ZNGM2b1KRpo+o8e+5P3dqVCQuL4PdtB9my9UCSry2XrTXTJw3ExXUJwcFvcB7ajdIlC3Pl2l3OnL9KreoVGDp6LgtnO3Pv/mOePvdPsH7nTs1o3KAqcXGQN08uho+ZT5FCeRnU/3scihfk9etQdu87jsfmPcTGxjF+dG/U6hgKFbSnSKF83L3vy5yF7jx4+DTF96FF01r82Oc7Rros4NadR4me7+DYkC7fN8fIyJA/tx9K8NyHjtnbbf/vhxZY57Dgoe9zFrl5cP3mfdwWuHDqtA+b/tgPQN3a3zCo3/d06OJMhXIOjBzSlSPHztHesQFxcXGsXPMnxtkM6dypOQb6emzYvIeNW/YBkNM6O8MGdqFc2WJERkaza+8xNmzaTWxsHE7dHMmfzw49lYoq35QiMCiE1e5/ccjzDHOmD8PWJgeuLn0pkC83v287yLhRvahYvgTR0WquXb/HvMXrCQ5+k+iYbP1tPmfOXaFOzUpcunKb8VN+jv8s9miHXS5rfJ+8wG35Ji5fvZPovZyz0J2N7rPYs+84jRpUY92GHRQplFf7eVUqFXTu2JzWLepgamrM5Su3WbBkA/4Br3Ad25fQsAgWLtmgrWWT+yxWr9/OYc8zydYAYGuTg7HOTpQqWYQnT/y4dTfxey1ERpehv1MzzmbE/75vQUBgMA8ePQPAPrcNF31u0qrDEE6e9mFA346UKlGIPgOn0qnrGFRKFWNH9QIgX147prkOxN1jJ01a/8iKNVuZMXlQou9InId2xevMZZq2GcDgkbNpWO9bqlYpm6iemVMGY2KSjR96uNCl51jMzUyYNK6f9vlKFUpy8/ZDWrQdxLJVvzOgb0dyWmdP8rUNHdiZVeu24ev7nKWLxnLqzCWatR3E/r9PMaBvR67dvM+9+0/Yvfc4jq3qJ1p/45Z9HDx8mp17jjJ8zHwszE1xWzCGC97Xadl+MM7jFtG8SU06d2yuXadp4xqs/XUHTR0H4HP5NvNmDEdPT/XB96Bpo+r069WBYaPmJdnQvq1chj492zFukhutvxuKmakJxtmMdDpmlSqUYNigLsxZsI7GrQdw4NAp5k4flmJNAAUL2KNQQKsOQ/hl7Z84D+tGvnx2dOjizLQ5q+jf6zssLcxQKhXMnT4M/4BXtPthJINGzKZ+3Sq0d2yo3Va92pXZe/AETdsMZOfeYzgP646Bvj5jJvyE38sgps7+BXePnXTp2BxDAwMcOw7n++5jyJbNkM4dmyVbY4H8uenwP2dmzl+DQ7ECTB7fH7cVm2jedhC/bdnLvBnDsbayTPReAhga6BMZGU2LdoPZtedYgu1+164xzRrXYPiY+Th2HI7vkxfMnjYUgD37/6F+ncqoVPHHsKRDISwtzfnnxMUP1gAwbeJA/F4G0aLtYGbOW0PNquVTfB+EyGgyXFNzWzCG/TuWsX/HMn73mEvpkoUZM+EnoqKitcvs+/skanUM0dFqWjWvzYrVWwl6FUJ4RCRuKzZTp2YlLC3MaFC3Cj6XbnL0+HliY+M4ccqbgcNmEh4RkWCf0dFqvq1chto1KvImNJx2P4zk1OlLCZbJbZeT8mWLs8jNgzeh4bwJDWehmwfVqpTV/lJ49eo123d5oomN5bDnGe1Q4vssLc0oXbIwew+c4PsOTfF9/Jw/tv1NTIyGg4e9CA0N58bN+wBcvX6XggVyp3jcalavQMjrUDZs2kNMjAbfx89Zt2EnLZrV0i7jeewcZ89fJSZGw6p127AwN6Fs6aLJbrN+nSqMHdWLsa5LuHv/cZLLNKr/LX8fOc2NWw9Qq2NYtup31OoYnY5Zk4bVOXjYC5/Lt4iLi+PP7YdxcV2S4msFiI2NxWPzXmJj47hw8Tp6KhUbt+xDrY7B68xlVColNjY5cChWkHx5c+G2YjPR0Wpe+AXw68bdtGlRV7utW3cectLLB01sLAf+PoWZqTHZsycedoyKVlO4UB6aNKqGmakJI8cuZNkvvydb49F/zhMZGU1YWAQtm9XmsOcZLnjfQBMby9Hj57l6/R6NG1RLdv0Dh06h0WgIf29YsnXzOqzfuIunz14SrVazcvVW8uWxxaF4QS543yAqKpqqVcoA8X+UHDl6lmi1+oM12OWyplTJwqxYs5VotZo793zZseeoTu+FEBlJhht+HDxyjvY7taREREQRFhbflCwtzTAyMmTejOG8m58TFRWNXS5rrK0s8XsZlGD9G7ceJNrmxGnL6dOzHYN//AGbnNk5ffYK835aT0BgsHaZ7NnNiYmJwT/glfaxl/5BxMTEYGtjBcCr4NcJtqvRaFAqE//dUKxIfm7f8SUmRsO3lUuz/9Ap7XP6+nqYmmTj2o34pmZlZcnrf4fqPiRHdnNe+AUmeOz5C39tbQBPnvolqC3o1WtyZLdg/qwRlCtTDAA/v0Dt94jlyhTjwYOnNGlUnWs37iW5X6sclnhfuqX9d1RUNEGvQoCUj5lVDksuXbmtfS4uLo6r1++m+Frj96PW/qGjiY1/90P/HXKOi4v/t1KhIJetFfr6+uzZ5qZdVwEJPi/vvm8xMZr4ZRSJ3zePzbvRxGpo27o+o4b14O59XxYs2cC160kfm4CAYO1/57K1omL5EtStXVn7mEqlTPCeJFr/nc/fu3LZWjFqaDdGDO6qfUypVGJna83NWw/Yd/AkjRtU4/TZK9SvW4UxExanWINVDktiYmJ49eq/Y/H8RQBlSxdLtj4hMqIM19RSEvfOr6OQkFCiotX0HzJdOzypUiqxt7fl6VM/Xn4TRJn3zkS6dW7JP6e8EzxWtHBelv/yOwuXbCCPvS0uI3vSr1cHZsxdrV3G72Ugenp62OTMwUv/+EZpa2OFnp4egUEh5M9np/NrMDczITQs/hewoaEBERFR2uca1K1CZGQ0vo+fo1KpaN+mAavX/5XiNv1eBiW6JNw+tw1BQSHaf787FKqvr4dVDkte+gfhPHZhktv8ecVm7t5/zJrlk/E8di7Jqxn9A16Ry/a/xqmnp8Ly34srUjpm/gFB2ORMODw7oG9Hft92kNjYWPT1//t4WpibJlguDt1iAP0DXvH6dSgtOwzRPmZmaoyJibFO67+rYAF7Dnue5bfNe7EwN6VH19ZMGf8jHbo4J7n8uzX6BwTz5/bDLP1li/Yxu1zWvHmT/MVPb5vz+/wDXrF42Sa8zvw3mpA/nx3PnwcAsO/ACdavmk7tGhUJCQnV/kHyoRpMTLKhp6eHtZWltpnmtLb88AERIgPKcMOPqREXF8e+gycY0LcTlhZm6Omp6OPUnp8XuqDSU3Hk2FkqlC1OjWrlUSgU1KxWni6dmmsvUnhr2MAu9OzWBj09FUFBIcTEaAh5HZpgmYCAYM6ev8qwQV0wMzXGzNSY4YO64H3pJi/8AlJVd0BgMHa28Q3o3IVrtHdsQB57W+rW+oahAzprm+Sc6UO59+AxJ718ktxOtFqt/eV86vQljI2N6Na5JXp6KvLlyUX3Lq3Yd/CEdvlG9b+lVMnC6Omp6NerA/4BQR88M1LHxPDg4VM2bdnHWGcnjIwMEi2z7+BJGjWoSrkyxdDTU9HXqT2GhgY6HbMDh7xoVL8qpUoWRqFQ0LZ1fZo0rE5ISCiPn7ygVo2KGBoaYG1tSet3hgtT4/rNB7wKfk3vHm0x0NfHzMyEaa4DGDaws07rq9VqTP89xq2b12WssxNmZia8CQ0jMjIq0eckOXsP/EOLZrUoU6oIAGVKFWH9L9OoWN4BSPhepmTP/n9w6tYGWxsrFAoF7do0YN2KKZibmwDw9Lk/N28/YFD/79l38KRONfi9DOSizw0G9f8eIyMDCuTPTZuW9XSqR4iMJFM3NYAlyzbx+MkL1iyfzJ4/3ShVojAjXRYQHa3m8RM/xk3+mV7dHTmwYxm9e7Rj3GS3RMM6k6Yvp0TxQuz+040/N84nIDCYteu3J9rXlJkrCQ+PZKP7LH73mEd4RCQTpixNdc3Xbtwjb95c5LG3Zc2v2wkIDObXVdNo3rQmcxetxyqHBYvnjebK1TvM++lX7JP4Xg7gyNGzVK1ShtVLXQkNC2fEmAV8U7Eku/90Y8mCMfx95DTuG3Zql/e5fJuBfTuxd9vPFCmUF+exC4mNTfmMx/23najVMfzYu2Oi57wv3WSRmwcTXfqyd9vPGBkZJmjyHzpm3pdustDNg7HOThzYsYzGDarhPHYhanUMGzbtJiwsnB2//8Si2c4cfGeINjU0Gg3O4xZRuFBetm1ewJZf5/Aq+E2Cs/AP2bP/BMMGdmFQ/+9ZuXYrQUEhbF4/mwM7llGqRGGmzFyp03auXLvLnAXrGDm0Gwd3LmeiS19WrtmqvV3j3fcyJRu37OOklw9uC8ZwYMcymjeugfO4RQk+13sPnCCndXb2H/qvqaVUg+v05RgZGrDrj8XMnDw42VtJhMjIFAVLNNZtHEd8Vp07NaN+7coMGTWX8PCk70/S19djyoQfiYqK1vmXZ3LevYVBCCGyqkx/ppZZbdyyD+/Lt/ht7Uy+a9eIvHlsUf17UYmlhRmNG1Zj7fLJxMRomDlvTTpXK4QQmUOmu1AkK1m6cguHj56lTYu6OLash03O7CgUSsLCI7h+8z4r1mxN9vs0IYQQicnwoxBCiCxDhh+FEEJkGWne1IyNjfh11TRy2SaO2ShYwJ5f3Caycd0spk8amOQl5EIIIURy0rSplSpZmBWLx5Mvb9I3Kk8a249lq36nc8+xPHz0jB5dWqdleUIIITK5NG1qbVrUZcGSDUlO/2OTMwcmJtnwuRw/5dKuvceoX7dKWpYnhBAik0vTqx8/dGl6TuvsCeYIDAgIxiZnjrQoSwghRBaRYS7pVygUiR6LTWbuOwCjbNlRJDFZsBBCNxFhgSkvJEQmk2Gamn9AkDbCBeJnp/f3D0p2+ciIV8TPty4yomwmOYgIS/79E0KILyHDNDW/l0FERkZRvmxxfC7fokXTWpw6cymFteQWu4zp3T825D0SQqSddB+/mz9zOA7FCgAwacYKfuzzHR5rZ1DSoRCr3VOOXBFCCCHekhlFxBegeGf4UT5emVV43W+/2LaNj55JcRlDQwN692hLzWrliYqOD4X12LyXf95LD+jr1J4Ojg1x7DQ8UUp4crb+Np9BI2ZTpFAeHIoXTPYPaLtc1nT/X2tmz1+LQ7ECOLaqx+wF63Bb4MLaX7cD4NTNkcEjZ+u036S4LRjD4JFzknyuTcu6vH4dhufxc7g4O1G+TDHcPXbSsN63OI9b9NH7fKtE8YLUrf0Ny1f98cnbSk4uW2t+XuhChy7O/PBdUx4+eobX2ctfbH8ZZvhRCCHeNWvKYJ4996dr7wnExGjIY2/LvJnDMTDQ57BnfFNUKBQ0aViNcxeu0bhBNbbv9kzVPk54+XDiA/Or5rK1xt7OBoCbtx8ye8G6j349yalYvkSSj1tamNGoflUGjYhvmC2a1KR+876o1THs//vjopjeV7CAPTmyW3yWbeli6/ZDLPtpHOe9r6NWx3yRfUhTE0JkOGVLF6VA/tyMnvATMTEaAJ489WPJ0o0MGdhZ29S+rVyawMAQduw5yoA+HZNtamZmJkwa2xdbGysePHyKgYE+AM2b1KRCOQdmzF1Nz66tqVu7MhqNhhu3HjBv0XpGDumKra0Vo4Z155DnmSTPyiwtTJk/awS2OXNw7eZ9Fiz+FbU6hiqVStHHqT0qlRL/gGDmLXInIDCYrb/N5/rN+xQtko9LV24DsGbZJHoNmJJgu+0dG3D0+HkA5kwbilKpZNXPrsxf/CuTx/enQxdnxo/uTWhoOMWK5sfWxooduz3ZsGkPSqWCH3t3pFLFEqhUKjyPncPdY2eC7Zubm9Cre1uMsxnSs2tr/F4GaY8FxMdVeV+6yUWfm8yeNoTbdx5RrGh+oqPUTJqxnOcvAiheND9DB3bGyMiQ0NBw5i/egO/j5xQtko+xzk4A3L33WLtPtTqGy1dv07RRdXbtPZ76D4YOsnxT+5JDKF+aLkM0QmRFJRwKcfvOI21De8v78k3y5ckVnz7+JowWTWtz+NhZLly8jrWVJaVKFuba9XuJttenR1vu3nuM87hFlCtTjAb1Ev5eMDHJRrs2DWjTcRhxcTBySFdscuZgwZINOHVzZN5P66lQziHJWnPb5WTsJDeePPVj6oQfcWxVj4OHvBjQtxODnefw5k0YjRtUY9igLtqA3HMXruI6bRkArZrVTtTQAGpVr8jU2b8AMGbiYk4edqdHP9dEUwzmymXNoBGzyZHdnM2/zuGvnZ7Ur1sZPX0VTv0no1KpmDNtCNWrluPU6f8uvnv9Oow16/+iQjkH1m3YSfMmNZN9PwoXzMPs+Wu5efshQwd2pr1jQ5av+oOxo3rhMnEJL/wCKFOqCJPG9qXXgClMHNOHn1du4ez5q/T4X+sEZ6PePjdp2ay2NDUhhDAyNARApVRibm5CjarlWLzsNzSxsXgeP0fbVvWTbGoVyjkweeYKAC5duc3TZy8TPB8WFsHDR89YvWwSJ718+HP7IV76B2Gf2ybFmi5fvcOTp34AHDzsRfOmtXj67CW2tla4zR8DgEKpIO6dlPkr1+6muN089rb4vUz5XsIzZ68QFxdHYFAIr1+HYmKSjSrflKZYkfxUKBvfiI2MDChcME+CppYar4LfcPP2QwDu3PWlfNni5Mubizy5bZk9dYh2OUtLMywtzchpnZ2z568C8SnsLZvV1i7zwi+QPPa2H1WHLqSpCSEynJu3HtCpfWP09FTExGiwMDfl9ZswSpcqwgu/AIJD3vBdu0agULBi8QQgPinezNSYxcs20ql9Y2pWqwDAavdtxMXFoXjnVpP3zwABBo+cQ5lSRahauSwL5zjrnDaviY397x8KBTExGpRKJZev3GbMxMUA6OmpMDMz0S4WGRmd4nbj4mKTrPN90Wr1O+vEf8+oVCpZunILx05cAOKHGqOi1P9eePPfcUm4vzjenQNDT0/13z6i1QmWVSgUqJRKnj5/SY9+rtrHc1pnJzY2lnc39P5riNFoEh6zz0yaWgZW/01mHTpV4GVyJ72LEJnYpSu3efDwGUMHdGbxso3UqVWJNi3rYW5mwqp18b+MWzSpxdyF6xJcNLFuxRRaNK3Fave/ElzRWLF8CZo1qYHb8s0UL5o/0ZlCLltr5kwfSp+BU7ly7S7W1pYUKZSX23cfoVJ9+M6nMiWLYGtjxUv/IJo3rsHpc1e4fuM+LiN7kjePLY+f+NG5Y3Mcihdg3CS3ROvHaDSolMpEv+gfP/XDLpc1Dx89S/Xxu+h9g1Yt6nDCywc9PRVL5o1h9fq/Eh2Xxg2raV9fcEgohQrmQaFQYGFhSoVyDpy7cC3ZfTx6/BxzMxPKlSnGpSu3adygGj9815Se/Sfx4kUANatX4MQpbxq+N9RrlysnT/89s/0SpKkJITKksZOW0KdnOzasmYEmRkNkZBTPnvtTrkxxHjx8irWVJYePnk2wzuatB+jRtTWb/9if4PHV6/9iwujeeKyZwSPf5zx7nnD48YVfAIc9z7JuxRQio6Lx8wtg74ETqPRUmJpkY/L4/uzYfTTJOh88eorLyJ7ktM7ORZ8b7N53nNjYOGbNX8vUCQNQqZQEvXrNtH+/H3vfsX8usH71NHoPmJLgDO7EKR++qVDyo5raX7s8sbe3wX3lFPT0VBzyPMOJU96Jlrt2/R5OXdswsG8nVq7dSrPGNdi8fjbPX/hz0efGB/ehVscwYepShg7ojKGhAZGRUUyZFX92O3XWSsaN7k2vbo5cvZ5wqLVShRL8k0Qtn0uWv08tM18o0nJXelfwsRQYfneUqMg3ZNb71I4cTfm7FJE+qlQqhc+V24mGxLIaSwszZk4ZzIBhM9O7lM9GX1+PFYvH03/ojKxxSX+92pXp1d0RPT0VBw55sW7DjgTP57K1ZqJLH0yMsxEaFs602at0+qJUCPH1OPuBIbGsJDjkDX8fOU2DulUSnZFmVh3bN2bthh1frKFBGk6TlSO7BYP6d2LQiNl0cRpPuTLFqPJN6QTL9OnZjkOeZ+jRz5Wjx8/Tr1f7tCpPCCEynL92HskyDQ3gt817OfmBm90/hzRrapUrleKC9w2CQ96g0WjY//dJGrwXAqpUKjA2NgLAwFCfqKisPbwghBDi80qz4cdEIaCBiUNAV6/bxgq3CXzXthF6KhX9hkxPYau6RM9IPE36yqzHP7PWnRqZ8/tOIT4kzZqaUplECGhswv+pxo/pw5yF7pw45U3dWt8wa8pguvWZmOT2dA0JjTEy+7iCMwBDo/Su4NMYZuJjn80k66euS0ioyIrSrKm99A+i/DvTzFjlsMQ/4L8QSUsLM/Lns9Nednr0n/M4D+uGpYUZwSFvEm1P15DQ6MjE62YWUbpNOJ4hGcK/Vz9mThJwKkTmlGbfqZ2/eJ1vKpQke3ZzVCoVTRtVTzBlS3DIG6Kj1dr51cqUKkJERFSSDe0/cfKTIX/kPcocP0JkPWl2phYQGMzSlVtYPHcU+gb6nDjpzfGTF3EZ2ZMTp7w54eXDuEluDB/8P4wMDQiPiGTClJ/TqjwhxHu+5Iw2R8wkT+0tXfLUgkPefPJ+0sO7WWrJcermCIC7xw6mTxrE1FkrdZpGLDlpep+a5/FzeB4/l+Cxd/OJbtx6QN9B09KyJCFEBiV5av/lqSWXEJCVxMbGsWf/P/Tq1palv2z56O3INFlCJCE898uUF8qgjJ9l/tlQJE8tYZ7ah/bTvk0DmjaugZGhAQCu05fz4OHTJF+PLjlrb+fB9H38gqKF8/Lo8QtOnfahWeMaWJibMm6SGw8ePaNUicIMHRg/RVZw8BvmLnLn6bOXyWapWVqYMXp4d2z/jc5Z++v2RPesnT5zmaEDOuP+207CwiJS+JQkLc2+UxNCCF3pkqcGJJmnlpS3eWpde0/gzx2HscphkeD5t3lqPfu50uvHKcTFxmnz1G7eesi8n9YnW2tuu5z89PNvdO09AWMjQxxb1cPC3JQBfTsxwmUBTv0nc9jzDMMGddGuc+7CVX7o7sLs+WsBks1Tu3jp5gf3Y2xsRJ1alRg0YjZde0/g6D/ncWxZL9nX07JZbW3OmlP/yZQuWZjqVcsl2neRQnnx2LSbH3qMpUA+O+xsrek/ZAYHD5+mVfM66OmpmDrxRxYv3Uj3PhPZvsuTKeP7AzBxTB9WrN6KU//JPHvur93m0IGdOXDIi14/TmbEmPkM+fEHLC0TXiGtiY3l7j1fKiVz9qoLOVMTQmQaX3OeWlL7+WPb37hOX06DulXIlycX31Yuw517vsm+Hl1z1gICXnHrziMAXvq/4rz3dQD8XgaS2y4nefPk4s2bcK7diD/WnsfPMXpED2xtrJLNUqvyTWkKFrCnZ9c2QHy0TYF8uRO97k/NW5OmJoTIcCRPjUR5akntxyZnDpYuGsvW7YfwOnuZwFchFCuSP9nXk1zO2vvU7x0fjSZhLE5S9x0rFAoiIiOTzVJTKpUMHhmfBA5glcOC4OA3ib5T/NS8NRl+FEJkOO/mqenpqahTqxKrl01i6IDOifLUOnRxpkMXZ9p0HMYj3+faPLUe/Vzp0c+VE14+nL94nWZNagAkm6e2ftU0bt15xCr3bZw9f5UihfKi0Wh0zlNTKBQ0b1yD8xevcf3GfUqVLEzePPH76dyxOaOGdU9y/bd5au97m6f2of2UKF6QJ0/92LL1ANdv3qdq5bIolcpkX8/bnDWVSoWhoQFL5o2hcqVSur8x//J9/AIzMxNKlYgf7q1fpzL+/kG8fh2mzVIDEmSpXfS+QbvW9YH4s1CPNTMwMzdJtO1PzVuTMzUhRIYkeWoJ89SS2o+hgQFtW9fDY+0M1NExXL95n0IF8yT7eqKi1TrlrKVErY7Bdfoyhg7sTDYjQ96EhjNx2jIg+Sy1RT97MGpYd9avmoZSoWDW/LUEBye8D1mpVFC8aH6mz1mV6prekjy1DEzy1NLP7tvpXcHHywpXP36I5KllXbVrVKRUycIsX/XHR29Dhh+FEJnK2QvXsnxDg4R5al8DpVJBi6a1WO/xaX/Np3r4sWABe4oUysuRY+cokN+Oe/ef6LxuSiGhVjkscBnpRE7r7ERERjFl5kpe+AWktkQhhMgS/tp5JL1LSDOxsXHaC2s+RarO1Fo0rcWEMX3o3aMtZqbGzJsxnFbN6+i0ri4hoRNd+nLSy4ce/Vw58PcpBvbrlJryhBBCfOVS1dQ6ODZkwLCZhIdHxs9F1n8ynTo01mndlEJCLcxNKVI4r3ZGgD0H/mHlmq2pKU8IIcRXLlXDj7GxsURF/Xd1TnDIG2I1ut1PkFJIqH1uG/z8Ahn84/dUqlCSl/5BLFi8IYWtSkhoxifHP+3peswz50U8QnxIqpranfuP6diuMfr6ehQtnI+2retz+66vTuumFBKqUilxKF6QdR47cVu+mVbNazNhTJ9kZ6WWkNCMLzOHhBoYpncFH0/XgFMJCRVZUaqa2iI3D7p3aUVkZBRjRzlxwfsGP6/YrNO6KYWEBgWFEB4Rqb1n4uDh0wwb2CXRdt6SkNCMLbOHhEZHpXcFHy8rBZzqEj/zMdEzX1qNauXJY2/Llq0H0ruUr06qmprLiJ5MmbWSX9b+meodnb94nd492pE9uzmvX4fRtFH1BFf2PH3uj9/LQGpUK89JLx+qfVtWO/dY8nQZPpEhlrT37h8bcvzT3uc55vXrfrmkgiNHdbuXLqX4mU+NnvlSihctkN4lfLVS1dTy5LHF2NiI8PDU/zWkU0ioqxujhvegf68OhEdEftJd5UKIzE2X+Bldo2ecujlil8uagvlzkyOHBR6b91KieEFKlSjMvQdPmDh1KRXKOdCnZ1uiotTY2lhx685DZs1bS7Raze6tSzh42ItyZYqhr6/HjLmruXXnEfZ2OXEe1h0LC1PU6hjclm8mPCISx1b1APAPCGLX3uNpdszER9yntm3TQnx9nxMZ9d/4zBDnuTqtm1JIqO+TF5ku2VUI8WXoEj+TVPRMUrP0AxQqmIe+A6dSpnRRlswfQ7c+E3j8+AW/rZtJkUJ5AXAoXoie/Vx55Pucaa4D6dC2IRt/30f27OZcvnqHJcs30al9Y7p1bsX4KT8zfkwflizbyM3bD7G3y8lP80bzfbcxbN8V31yloaW9VDW1pSs/Po1UCCE+h7fxM3p6Kp2jZwDOnr+CJjaWF36BBAYFa+dU9A8I1s6gf+XqHR75PgfgwN8nad2yLht/3wegjWe5c8+XmtUrkM3IkJIOhXAZ6aTdh56eCltbqy/zwoVOUtXUfC7fomqVMlSqUBI9PRU+l25pIwyEEOJzSil+pmG9b3WOngFQq9+JcUnmViSN5r9lFEplgluWotXxU3PFxQEoUKqUREer6dHPVbuMtbUlgYEhn+X1i4+TqqbWuVMz6tb6hoOHvVAoFPzvhxYULGCfKA5cCCE+1bvxM4uXbaROrUq0aVkPczMTVq3bxvcdmjJ34Tr2/31Ku866FVO00TOr3f/SPl5Mxws3ypQqgk3OHPgHvKJZo+qcvXAt2WXDwiJ4/NSPxg2rcfCQF2VLF2X6pEG0+2EEGo0GAwP9j37t4uOlqqk1aVidPgOnaicT3bnnGGuXT5amJoT4IpKLn2naqDoFC+TWOXpGVy/9XzFuVC9sbXJw/uJ1duz68NWUU2auZNSwbnTp1JxYTSwTpvxMTIyGiz43cB3bj9evQ9ny58GPqkV8nFRFz3ismcH/eo3X/lupVOC+cird+kz8IsV9DhI9kx4keiY9ZfXoGfgy8TMVyjng1M1RLlbL5FJ1pnb+4nVmThnM3v3/ANCiWW0u+Nz4IoUJIURyPjQsKL5uqQ4JdWxVj0oVSqJUKDjvfZ2du4+iidVt/sf0IGdq6UHO1NLT13CmJkRyUnWmZmZqjJmpMROnLsXWJgfff9cUo2yGhIVFfKn6hBBCCJ2lKnpm0vj+2v8OeR1KYGAIri59P3tRQgghxMdIVVOztrJkw6Y9AERGRuOxeU+C+JiU1KtdGY81M9i8fjY9u7ZJdrmiRfLhuU+myBJCCJE6qWpqcbFxFC2cT/vv/PnsUMfE6LSuLsnXED8r94jBXeUeDyGEEKmWqu/U3FZsYv6sEQQEBkNcHBYWZkyesUKndd9Nvga0yddnz19NsNzg/t+z5c8DlC1dNDWlCSGEELo3tRrVyvPg4VPa/TCSTh0aU7VyGbwv3+LGrQc6rZ9S8jVAzWrlMTIy4Ojx8zpWJcnXGZ8c/7T3eZKvw3N/ueiZ1Fyhmcfelk3us5g1fy17D5wAYOtv8xk0YjYVyztQoZwDM+au/lKl6sTEJBsTxvRhrOuSRM81b1IzQ9RYuVIp+vf+DpVKSUhIKDPnrcHvZdYLitWpqf3wXVMa1K3C9LmrKZg/N05dHVm87DcK5M/NoP6dWLx0Y4rbSCn5Okd2C7r/rzVDR+k2478kX2d8knydPrJa8nXLprU4+s8FHFvW0za1jMbM1CTBVzMZjZ6eiokufRk4fCaPn/jRqnkdhg/qgksSTTiz06mpNWlUnX6DpxMVFU3/3t9xwstbG6nw29qZOu0opeTrGtXKYWFuyrJF47SPua+cyqARswkNC0+0PUm+ztgk+Tr9ZKXka6VSQdPGNRg+eh6L5o6iaOF83Lnnm+J6uWytmTN9KL6PX1C0cF4ePX7BqdM+NGtcAwtzU8ZNcuPBo2ds/W0+J055U65MMRQKBbMXrOXm7YeMH92b0NBwihXNj62NFTt2e7Jh0x6USgU/9u5IpYolUKlUeB47h7vHTkYO6Yq1lSVzpg1lzMTFieqxz52Tnxe6YGlhxsnTPixf9QcQn9r9TcWSmJka8+ZNOGMnLSHo1WtGDulK6VJFiI2N5cQpH9b+uh0jIwOGD/ofRYvkQ6lU8uf2Q4mibczNTdi4dhZtfxiBWh2DXS5rFs8bTY++rixe+huPn/gBcOfuI9q1rv8Z3qGMR7cLReIgKioagIrlHThz7kqqd3T+4nW+qVCS7NnNUalUNG1UXRvlAPG5Qx27jqZHP1ftrNc9+rkm2dASFCY/GfBH3qPM8ZPxfVu5LCGvQ3nw6BlHjp7FsXU9ndctUigvHpt280OPsRTIZ4edrTX9h8zg4OHTtGpeR7tcaFg4PftPYpX7NiaM6aN9PFcuawaNmE2/wdPo1qUVpibGtGxWGz19FU79J+PUfzKlSxametVyLFiygYDA4CQbGoC9nQ0TpiylRz9XypQqSs3qFbDPbUPBAvb0HzKdH3qMxffJcxo1qEbhgnko4VCIHn1d6T94BnntbTE0NKB7l1bcf/gUp/6T6T9kOm1b16dwoTwJ9vP6dRiXr92hapWyADSqX42/D58mPCJSO0+mUqnAqbsjx09e1PlYZiY6nalpNBpMTYwxMTGiWJH82os7bG1yJIhq+BBdkq+FEOJdLZvW4tCR0wD8feQ0S+aPZukK3XIdAwJecevOIyB+ouLz3tcB8HsZSG67nNrldu45BsBJLx8mjO6NhbkpAGfOXiEuLo7AoBBevw7FxCQbVb4pTbEi+alQNn7UycjIgMIF83D/wdMP1nLy9CXtRXJHjp2lYjkHTpzyZvHSjbRsVpt8ee0oXbIIT5/58+SZHwb6eixdNBavM5dZufZPoqKiqVKpNEZGhjRrVAMAY2MjChfKy737TxLsa//BkzSs9y3/nLxIo/rfMm6ym/Y5AwN9Jo3th1Kh4NeNu3U6jpmNTk1tw+Y9uP8yFZVKya69xwgMCqFe7cr07dUe9w26z9CfUvL1u2o06KHzdoUQWY+FuSnVq5ajZIlCtGkZf4amr69Pk0bVEy3rUKyANqzz5u0HuHvsQv1eYrbOGWr/Tvv3Nj8NIC4OFAoFSqWSpSu3aHMkzc1NiIpSk93SXLusY8t6OLaKr3f7Lk+i1WrtNiF+OzEaDcWL5mfqxAFs+mM/nsfPERsbi0KhIDIymh59XalYoQRVvinNSrcJDBoxC6VSyZRZK7n9b6O2tDQjLCwCl5E9cShWEIDZC9Zy6swlhg7sQqkShYmIjNIOOZqZGjNv5nCePffHdfpynU9IMhudmtrR4+e5eu0uFham2r8KIiOjmLvQHe9LN79ogUKIr1PTRtW5cu0OQ5z/u3isY7vG2obxrpu3HyYI68xla63zfhrW+5Ytfx6kdo2KPHnygjehyX/lcdH7Bq1a1OGElw96eiqWzBvD6vV/cev2Q1Sq+G9ztu/2ZPtuT+06zZvU5NvKZTAzNSYqSk2jelVZ7b6N8uUcuOB9g+27PDExycaood056eVD2dJFcerWhhEuCzh34RrFiuQnX147LvjcoG2r+sxZuA5LCzPWrZjC6PE/JXlicPK0DyOGdGXPv5PPA8ycMpjrN+6zZPkmnY9NZqTzJf0BgcHx96f9y+vs5S9RjxBCANC8aS3Wrt+e4LHd+47j1N2RbEYGn20/JRwKsW7FFKLVaqbO/vBMRn/t8sTe3gb3lVPQ01NxyPMMJ055o1QqeP4igGU/jWXAsFmJ1nvk+4w504dhbm7CoSNnOHvhGvcfPWXm5MGsXzWNmBgNd+/7ktsuJ5ev3uGR7ws2rJ5BVHQ0t+884vSZy3hfusnIId3YsHo6KpWK9R67kr1oZv/Bk7RuXofDnmeA+Mv5K5Yvgbm5Ke4rpwIQ9CqEES4LPvHoZTypnqU/s5FZ+tODzNKfnmSWft29vd/thV9AepciPpNUTZMlhBBCZGSpmiZLCCGykg5dnNO7BPGZyZmaEEKILEOamhBCiCxDmpoQQogsI02/U6tXuzK9ujuip6fiwCEv1m3YkeD5YkXzM2Lw/zAyMiQqKpp5i9Zz9/7jtCxRCCFEJpZmZ2q6hIS6uvRl+ao/6NHXldXufzHBpU8yWxNCCCESS7Om9m5IqEaj0YaEvqVQKNj0x34uXYm/Qej2nUfY2lilVXlCCCGygDQbfkwpJDQuLi7BlC59ndrzT4qzSEtIaMYnxz/tfZ6QUCEyozRraimFhL673NABXXAoVoAhHwgMlZDQjE9CQtNHVgsJFSI10qyppRQSCqCvr8fUiQPIZmTIYOc5hIcnn5IpIaEZm4SEpp+sFBIqRGql2XdqKYWEAowZ0QO1OoaRYxd+sKH9J71DFuUn5fDJ9K5FfnR7n4TIGtLsTC2lkNCHvs9p1rgmj3yfs2bZJO16Tj9OSnKYUgghhHifzNKfgcks/elHZukXInOSGUWEEEJkGdLUhBBCZBnS1IQQQmQZ0tSEEEJkGdLUhBBCZBnS1IQQQmQZ0tSEEEJkGdLUhBBCZBkZKiTU2tqSSWP7kSO7BYFBwbhOX05wcOadP1AIIUTaylAhoSOHdGPvgRN0cRrHgUNeDBvYJa3KE0IIkQVkmJBQlUpFxXIOHDx8GoD9B09S/dtyqFSqtCpRCCFEJpdhQkItzE0Jj4hEo9EAoImNJSw8guyWZgQEBiezVQkJzfjk+Kc9CQkVX68MExKa1PPvL5NYyv9TGh89neIyGdWRzJuxCfuNgMybcmqc3gV8EmlW4uuVZsOPL/2DsLKy1P77/ZDQV8FvMDHOhurfNGuVUolxNiNCXoemVYlCCCEyuQwTEqrRaPC+fIvGDasB0LhhNbwv3dQORwohhBApSdM8tXq1K9Oza2ttSOjSX7ZoQ0JPePlgkzMH40b1wtrKktdvQpkycyV+LyWaXgghhG6yfEioEEKIr4fMKCI+SoVyDrgtcEnVOiqVip/mjqJCOYcvVJUQ4muXpjOKiK9X/nx2jHV2oliR/OldihAiC5OmJj5Zm5Z1adGkFg99nyVqWtt3ebJ9tyctm9XGY/NeOrVvkk5VCiG+BtLUxCdp0rA6zRrXYLjLfMLCIpJdbunKLQDS1IQQX5Q0NfHRChW0Z9woJ6bOXkVYWAQuI3viUKxggmXenqkJIURakKYmPlpERBRTZ/3C6OHd8TpzidkL1qV3SUKIr5xc/Sg+2vMXAZw5d4UTp3wY0KdjepcjhBDS1MSnW7l2KzWqladcmWLpXYoQ4isnN18LIYTIMuRMTQghRJYhTU0IIUSWIU1NCCFEliFNTQghRJYhTU0IIUSWIU1NCCFEliFNTQghRJYhTU0IIUSWIU1NCCFEliFNTQghRJYhTU0IIUSW8X8C0Ex+esIsHAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 444.444x155.556 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 444.444x155.556 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 444.444x155.556 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 444.444x155.556 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Define the metrics and top_k values\n",
    "metric_names = ['accuracy', 'precision', 'recall', 'ndcg', 'mrr']\n",
    "top_k_values = {\n",
    "    'accuracy': [1, 5],\n",
    "    'precision': [1],\n",
    "    'recall': [1, 5],\n",
    "    'ndcg': [10],\n",
    "    'mrr': [10]\n",
    "}\n",
    "\n",
    "# Create a list of model names and their colors\n",
    "model_names = ['QA-distilbert (fine-tuned)', 'QA-distilbert (base model)', 'QA-mpnet', 'All-mpnet-base-v2']\n",
    "model_colors = ['cyan', 'magenta', 'yellow', 'green']\n",
    "\n",
    "# Function to plot metrics\n",
    "def plot_metrics(metric_name, top_k_values, model_names, model_colors, dpi=225):\n",
    "    x = np.arange(len(top_k_values[metric_name]))  # Set the x-axis positions for the bars\n",
    "    width = 0.2  # Width of the bars\n",
    "\n",
    "    # Convert pixel dimensions to inches for figsize\n",
    "    width_in_inches = 1000 / dpi\n",
    "    height_in_inches = 350 / dpi\n",
    "\n",
    "    fig, ax = plt.subplots(figsize=(width_in_inches, height_in_inches))\n",
    "\n",
    "    for j, top_k in enumerate(top_k_values[metric_name]):\n",
    "        metric_key = f'test-ir-eval_cosine_{metric_name}@{top_k}'\n",
    "        \n",
    "        # Extract the metric values for each model\n",
    "        metric_values = [\n",
    "            results_best_model.get(metric_key, 0),\n",
    "            results_base_model.get(metric_key, 0),\n",
    "            results_mpnet_model.get(metric_key, 0),\n",
    "            results_all_mpnet_model.get(metric_key, 0)\n",
    "        ]\n",
    "\n",
    "        # Create the bar chart for the current metric\n",
    "        for i, (model_name, color) in enumerate(zip(model_names, model_colors)):\n",
    "            ax.bar(x[j] + (i - 1.5) * width, metric_values[i], width, color=color, label=model_name, alpha=0.3)\n",
    "\n",
    "    # Set the x-axis labels and tick positions\n",
    "    ax.set_xticks(x)\n",
    "    ax.set_xticklabels([f'k={top_k}' for top_k in top_k_values[metric_name]])\n",
    "\n",
    "    # Set the title and labels for the current subplot\n",
    "    title_map = {'ndcg': 'NDCG', 'mrr': 'MRR'}\n",
    "    metric_title = title_map.get(metric_name, metric_name.capitalize())\n",
    "    ax.set_title(f'{metric_title} @ top-k documents retrieved')\n",
    "    ax.set_ylabel('Score')\n",
    "\n",
    "    # Set y-axis limits and ticks\n",
    "    ax.set_ylim(0, 1)\n",
    "    ax.set_yticks(np.arange(0, 1.1, 0.2))\n",
    "\n",
    "    # Remove duplicate legend entries and show only model names in the correct order\n",
    "    handles, labels = ax.get_legend_handles_labels()\n",
    "    unique_labels, unique_handles = [], []\n",
    "    for model_name in model_names:\n",
    "        if model_name in labels:\n",
    "            unique_labels.append(model_name)\n",
    "            unique_handles.append(handles[labels.index(model_name)])\n",
    "    ax.legend(unique_handles, unique_labels, bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)\n",
    "\n",
    "    # Adjust the layout and save the plot\n",
    "    plt.tight_layout(rect=[0, 0.03, 1, 0.95])\n",
    "    plt.savefig(f'../images/{metric_name}_test_plot.png', dpi=dpi)\n",
    "    plt.show()\n",
    "\n",
    "# Iterate over each metric and plot\n",
    "for metric_name in metric_names:\n",
    "    plot_metrics(metric_name, top_k_values, model_names, model_colors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Accuracy@1</th>\n",
       "      <th>Accuracy@5</th>\n",
       "      <th>Precision@1</th>\n",
       "      <th>Recall@1</th>\n",
       "      <th>Recall@5</th>\n",
       "      <th>NDCG@10</th>\n",
       "      <th>MRR@10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>QA-distilbert (fine-tuned)</th>\n",
       "      <td>0.9103</td>\n",
       "      <td>0.9862</td>\n",
       "      <td>0.9103</td>\n",
       "      <td>0.9103</td>\n",
       "      <td>0.9862</td>\n",
       "      <td>0.9586</td>\n",
       "      <td>0.9449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QA-distilbert (base model)</th>\n",
       "      <td>0.8759</td>\n",
       "      <td>0.9379</td>\n",
       "      <td>0.8759</td>\n",
       "      <td>0.8759</td>\n",
       "      <td>0.9379</td>\n",
       "      <td>0.9142</td>\n",
       "      <td>0.9020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QA-mpnet</th>\n",
       "      <td>0.8621</td>\n",
       "      <td>0.9310</td>\n",
       "      <td>0.8621</td>\n",
       "      <td>0.8621</td>\n",
       "      <td>0.9310</td>\n",
       "      <td>0.9027</td>\n",
       "      <td>0.8892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>All-mpnet-base-v2</th>\n",
       "      <td>0.7379</td>\n",
       "      <td>0.8759</td>\n",
       "      <td>0.7379</td>\n",
       "      <td>0.7379</td>\n",
       "      <td>0.8759</td>\n",
       "      <td>0.8276</td>\n",
       "      <td>0.8028</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Accuracy@1 Accuracy@5 Precision@1 Recall@1  \\\n",
       "QA-distilbert (fine-tuned)     0.9103     0.9862      0.9103   0.9103   \n",
       "QA-distilbert (base model)     0.8759     0.9379      0.8759   0.8759   \n",
       "QA-mpnet                       0.8621     0.9310      0.8621   0.8621   \n",
       "All-mpnet-base-v2              0.7379     0.8759      0.7379   0.7379   \n",
       "\n",
       "                           Recall@5 NDCG@10  MRR@10  \n",
       "QA-distilbert (fine-tuned)   0.9862  0.9586  0.9449  \n",
       "QA-distilbert (base model)   0.9379  0.9142  0.9020  \n",
       "QA-mpnet                     0.9310  0.9027  0.8892  \n",
       "All-mpnet-base-v2            0.8759  0.8276  0.8028  "
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Combine results into a DataFrame\n",
    "data = {\n",
    "    'QA-distilbert (fine-tuned)': [\n",
    "        results_best_model['test-ir-eval_cosine_accuracy@1'],\n",
    "        results_best_model['test-ir-eval_cosine_accuracy@5'],\n",
    "        results_best_model['test-ir-eval_cosine_precision@1'],\n",
    "        results_best_model['test-ir-eval_cosine_recall@1'],\n",
    "        results_best_model['test-ir-eval_cosine_recall@5'],\n",
    "        results_best_model['test-ir-eval_cosine_ndcg@10'],\n",
    "        results_best_model['test-ir-eval_cosine_mrr@10']\n",
    "    ],\n",
    "    'QA-distilbert (base model)': [\n",
    "        results_base_model['test-ir-eval_cosine_accuracy@1'],\n",
    "        results_base_model['test-ir-eval_cosine_accuracy@5'],\n",
    "        results_base_model['test-ir-eval_cosine_precision@1'],\n",
    "        results_base_model['test-ir-eval_cosine_recall@1'],\n",
    "        results_base_model['test-ir-eval_cosine_recall@5'],\n",
    "        results_base_model['test-ir-eval_cosine_ndcg@10'],\n",
    "        results_base_model['test-ir-eval_cosine_mrr@10']\n",
    "    ],\n",
    "    'QA-mpnet': [\n",
    "        results_mpnet_model['test-ir-eval_cosine_accuracy@1'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_accuracy@5'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_precision@1'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_recall@1'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_recall@5'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_ndcg@10'],\n",
    "        results_mpnet_model['test-ir-eval_cosine_mrr@10']\n",
    "    ],\n",
    "    'All-mpnet-base-v2': [\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_accuracy@1'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_accuracy@5'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_precision@1'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_recall@1'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_recall@5'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_ndcg@10'],\n",
    "        results_all_mpnet_model['test-ir-eval_cosine_mrr@10']\n",
    "    ]\n",
    "}\n",
    "\n",
    "# Define the column names (metrics)\n",
    "columns = [\n",
    "    'Accuracy@1', 'Accuracy@5', 'Precision@1', 'Recall@1', 'Recall@5', 'NDCG@10', 'MRR@10'\n",
    "]\n",
    "\n",
    "# Create the DataFrame and transpose it\n",
    "df = pd.DataFrame(data, index=columns).T\n",
    "\n",
    "# Format the DataFrame values to four decimal places\n",
    "df = df.astype(float).map(lambda x: f'{x:.4f}')\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_latex('../images/performance_metrics_table.tex', caption='Performance Metrics for Each Embedding Model', label='tab:performance_metrics', float_format=\"%.4f\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model metrics have been written to ../docs/model_metrics.txt\n"
     ]
    }
   ],
   "source": [
    "num_documents = len(unique_corpus)\n",
    "num_queries_total = len(queries)\n",
    "num_queries_train = len(train_queries)\n",
    "num_queries_val = len(val_queries)\n",
    "num_queries_test = len(test_queries)\n",
    "\n",
    "hyperparameters = {\n",
    "    'per_gpu_batch_size': int(best_hyperparameters['per_gpu_batch_size']),\n",
    "    'weight_decay': best_hyperparameters['weight_decay'],\n",
    "    'learning_rate': best_hyperparameters['learning_rate'],\n",
    "    'warmup_steps': int(best_hyperparameters['warmup_steps']),\n",
    "    'num_epochs': int(best_hyperparameters['num_epochs'])\n",
    "}\n",
    "\n",
    "performance_metrics = {\n",
    "    'Accuracy@1': results_best_model['test-ir-eval_cosine_accuracy@1'],\n",
    "    'Accuracy@5': results_best_model['test-ir-eval_cosine_accuracy@5'],\n",
    "    'Precision@1': results_best_model['test-ir-eval_cosine_precision@1'],\n",
    "    'Recall@1': results_best_model['test-ir-eval_cosine_recall@1'],\n",
    "    'Recall@5': results_best_model['test-ir-eval_cosine_recall@5'],\n",
    "    'NDCG@10': results_best_model['test-ir-eval_cosine_ndcg@10'],\n",
    "    'MRR@10': results_best_model['test-ir-eval_cosine_mrr@10']\n",
    "}\n",
    "\n",
    "# Combine all data into a single string for text file export\n",
    "data = {\n",
    "    'Number of Documents': num_documents,\n",
    "    'Number of Queries (Total)': num_queries_total,\n",
    "    'Number of Queries (Training)': num_queries_train,\n",
    "    'Number of Queries (Validation)': num_queries_val,\n",
    "    'Number of Queries (Test)': num_queries_test,\n",
    "    'Best Hyperparameters': hyperparameters,\n",
    "    'Performance Metrics': performance_metrics\n",
    "}\n",
    "\n",
    "def format_dict(data):\n",
    "    if isinstance(data, dict):\n",
    "        return '\\n'.join([f\"{key}: {value}\" for key, value in data.items()])\n",
    "    return str(data)\n",
    "\n",
    "# Create the content for the text file\n",
    "content = \"\\n\\n\".join([f\"{key}:\\n{format_dict(value)}\" for key, value in data.items()])\n",
    "\n",
    "# Define the text file name\n",
    "text_file_path = '../docs/model_metrics.txt'\n",
    "\n",
    "# Write the content to the text file\n",
    "with open(text_file_path, 'w') as file:\n",
    "    file.write(content)\n",
    "\n",
    "print(f\"Model metrics have been written to {text_file_path}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "chatbot",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}