notebooks/semantic_search.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "aGixSoAgIzLo",
   "metadata": {
    "id": "aGixSoAgIzLo",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pinecone-client in /opt/homebrew/lib/python3.11/site-packages (3.0.1)\n",
      "Collecting pinecone-client\n",
      "  Downloading pinecone_client-3.2.2-py3-none-any.whl.metadata (16 kB)\n",
      "Requirement already satisfied: openai in /opt/homebrew/lib/python3.11/site-packages (1.10.0)\n",
      "Collecting openai\n",
      "  Downloading openai-1.16.1-py3-none-any.whl.metadata (21 kB)\n",
      "Requirement already satisfied: sentence-transformers in /opt/homebrew/lib/python3.11/site-packages (2.2.2)\n",
      "Collecting sentence-transformers\n",
      "  Downloading sentence_transformers-2.6.1-py3-none-any.whl.metadata (11 kB)\n",
      "Requirement already satisfied: tiktoken in /opt/homebrew/lib/python3.11/site-packages (0.5.2)\n",
      "Collecting tiktoken\n",
      "  Downloading tiktoken-0.6.0-cp311-cp311-macosx_11_0_arm64.whl.metadata (6.6 kB)\n",
      "Requirement already satisfied: datasets in /opt/homebrew/lib/python3.11/site-packages (2.18.0)\n",
      "Requirement already satisfied: certifi>=2019.11.17 in /opt/homebrew/lib/python3.11/site-packages (from pinecone-client) (2023.7.22)\n",
      "Requirement already satisfied: tqdm>=4.64.1 in /opt/homebrew/lib/python3.11/site-packages (from pinecone-client) (4.66.1)\n",
      "Requirement already satisfied: typing-extensions>=3.7.4 in /opt/homebrew/lib/python3.11/site-packages (from pinecone-client) (4.8.0)\n",
      "Requirement already satisfied: urllib3>=1.26.0 in /opt/homebrew/lib/python3.11/site-packages (from pinecone-client) (1.26.18)\n",
      "Requirement already satisfied: anyio<5,>=3.5.0 in /opt/homebrew/lib/python3.11/site-packages (from openai) (3.7.1)\n",
      "Requirement already satisfied: distro<2,>=1.7.0 in /opt/homebrew/lib/python3.11/site-packages (from openai) (1.8.0)\n",
      "Requirement already satisfied: httpx<1,>=0.23.0 in /opt/homebrew/lib/python3.11/site-packages (from openai) (0.23.3)\n",
      "Requirement already satisfied: pydantic<3,>=1.9.0 in /opt/homebrew/lib/python3.11/site-packages (from openai) (1.10.12)\n",
      "Requirement already satisfied: sniffio in /opt/homebrew/lib/python3.11/site-packages (from openai) (1.3.0)\n",
      "Requirement already satisfied: transformers<5.0.0,>=4.32.0 in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (4.36.2)\n",
      "Requirement already satisfied: torch>=1.11.0 in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (2.0.1)\n",
      "Requirement already satisfied: numpy in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (1.26.0)\n",
      "Requirement already satisfied: scikit-learn in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (1.3.0)\n",
      "Requirement already satisfied: scipy in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (1.11.4)\n",
      "Requirement already satisfied: huggingface-hub>=0.15.1 in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (0.20.2)\n",
      "Requirement already satisfied: Pillow in /opt/homebrew/lib/python3.11/site-packages (from sentence-transformers) (9.5.0)\n",
      "Requirement already satisfied: regex>=2022.1.18 in /opt/homebrew/lib/python3.11/site-packages (from tiktoken) (2023.8.8)\n",
      "Requirement already satisfied: requests>=2.26.0 in /opt/homebrew/lib/python3.11/site-packages (from tiktoken) (2.31.0)\n",
      "Requirement already satisfied: filelock in /opt/homebrew/lib/python3.11/site-packages (from datasets) (3.12.4)\n",
      "Requirement already satisfied: pyarrow>=12.0.0 in /opt/homebrew/lib/python3.11/site-packages (from datasets) (12.0.1)\n",
      "Requirement already satisfied: pyarrow-hotfix in /opt/homebrew/lib/python3.11/site-packages (from datasets) (0.6)\n",
      "Requirement already satisfied: dill<0.3.9,>=0.3.0 in /opt/homebrew/lib/python3.11/site-packages (from datasets) (0.3.6)\n",
      "Requirement already satisfied: pandas in /opt/homebrew/lib/python3.11/site-packages (from datasets) (2.0.3)\n",
      "Requirement already satisfied: xxhash in /opt/homebrew/lib/python3.11/site-packages (from datasets) (3.2.0)\n",
      "Requirement already satisfied: multiprocess in /opt/homebrew/lib/python3.11/site-packages (from datasets) (0.70.14)\n",
      "Requirement already satisfied: fsspec<=2024.2.0,>=2023.1.0 in /opt/homebrew/lib/python3.11/site-packages (from fsspec[http]<=2024.2.0,>=2023.1.0->datasets) (2023.9.1)\n",
      "Requirement already satisfied: aiohttp in /opt/homebrew/lib/python3.11/site-packages (from datasets) (3.9.1)\n",
      "Requirement already satisfied: packaging in /opt/homebrew/lib/python3.11/site-packages (from datasets) (23.2)\n",
      "Requirement already satisfied: pyyaml>=5.1 in /opt/homebrew/lib/python3.11/site-packages (from datasets) (6.0)\n",
      "Requirement already satisfied: idna>=2.8 in /opt/homebrew/lib/python3.11/site-packages (from anyio<5,>=3.5.0->openai) (3.4)\n",
      "Requirement already satisfied: attrs>=17.3.0 in /opt/homebrew/lib/python3.11/site-packages (from aiohttp->datasets) (23.1.0)\n",
      "Requirement already satisfied: multidict<7.0,>=4.5 in /opt/homebrew/lib/python3.11/site-packages (from aiohttp->datasets) (5.2.0)\n",
      "Requirement already satisfied: yarl<2.0,>=1.0 in /opt/homebrew/lib/python3.11/site-packages (from aiohttp->datasets) (1.9.2)\n",
      "Requirement already satisfied: frozenlist>=1.1.1 in /opt/homebrew/lib/python3.11/site-packages (from aiohttp->datasets) (1.4.0)\n",
      "Requirement already satisfied: aiosignal>=1.1.2 in /opt/homebrew/lib/python3.11/site-packages (from aiohttp->datasets) (1.3.1)\n",
      "Requirement already satisfied: httpcore<0.17.0,>=0.15.0 in /opt/homebrew/lib/python3.11/site-packages (from httpx<1,>=0.23.0->openai) (0.16.3)\n",
      "Requirement already satisfied: rfc3986<2,>=1.3 in /opt/homebrew/lib/python3.11/site-packages (from rfc3986[idna2008]<2,>=1.3->httpx<1,>=0.23.0->openai) (1.5.0)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in /opt/homebrew/lib/python3.11/site-packages (from requests>=2.26.0->tiktoken) (2.1.1)\n",
      "Requirement already satisfied: sympy in /opt/homebrew/lib/python3.11/site-packages (from torch>=1.11.0->sentence-transformers) (1.12)\n",
      "Requirement already satisfied: networkx in /opt/homebrew/lib/python3.11/site-packages (from torch>=1.11.0->sentence-transformers) (3.1)\n",
      "Requirement already satisfied: jinja2 in /opt/homebrew/lib/python3.11/site-packages (from torch>=1.11.0->sentence-transformers) (3.1.2)\n",
      "Requirement already satisfied: tokenizers<0.19,>=0.14 in /opt/homebrew/lib/python3.11/site-packages (from transformers<5.0.0,>=4.32.0->sentence-transformers) (0.15.0)\n",
      "Requirement already satisfied: safetensors>=0.3.1 in /opt/homebrew/lib/python3.11/site-packages (from transformers<5.0.0,>=4.32.0->sentence-transformers) (0.3.3)\n",
      "Requirement already satisfied: python-dateutil>=2.8.2 in /opt/homebrew/lib/python3.11/site-packages (from pandas->datasets) (2.8.2)\n",
      "Requirement already satisfied: pytz>=2020.1 in /opt/homebrew/lib/python3.11/site-packages (from pandas->datasets) (2023.3)\n",
      "Requirement already satisfied: tzdata>=2022.1 in /opt/homebrew/lib/python3.11/site-packages (from pandas->datasets) (2023.3)\n",
      "Requirement already satisfied: joblib>=1.1.1 in /opt/homebrew/lib/python3.11/site-packages (from scikit-learn->sentence-transformers) (1.3.1)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/homebrew/lib/python3.11/site-packages (from scikit-learn->sentence-transformers) (3.2.0)\n",
      "Requirement already satisfied: h11<0.15,>=0.13 in /opt/homebrew/lib/python3.11/site-packages (from httpcore<0.17.0,>=0.15.0->httpx<1,>=0.23.0->openai) (0.14.0)\n",
      "Requirement already satisfied: six>=1.5 in /opt/homebrew/lib/python3.11/site-packages (from python-dateutil>=2.8.2->pandas->datasets) (1.16.0)\n",
      "Requirement already satisfied: MarkupSafe>=2.0 in /opt/homebrew/lib/python3.11/site-packages (from jinja2->torch>=1.11.0->sentence-transformers) (2.1.3)\n",
      "Requirement already satisfied: mpmath>=0.19 in /opt/homebrew/lib/python3.11/site-packages (from sympy->torch>=1.11.0->sentence-transformers) (1.3.0)\n",
      "Downloading pinecone_client-3.2.2-py3-none-any.whl (215 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m215.9/215.9 kB\u001b[0m \u001b[31m5.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading openai-1.16.1-py3-none-any.whl (266 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m266.9/266.9 kB\u001b[0m \u001b[31m14.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading sentence_transformers-2.6.1-py3-none-any.whl (163 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m163.3/163.3 kB\u001b[0m \u001b[31m12.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading tiktoken-0.6.0-cp311-cp311-macosx_11_0_arm64.whl (949 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m949.8/949.8 kB\u001b[0m \u001b[31m13.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n",
      "\u001b[?25hInstalling collected packages: pinecone-client, tiktoken, openai, sentence-transformers\n",
      "  Attempting uninstall: pinecone-client\n",
      "    Found existing installation: pinecone-client 3.0.1\n",
      "    Uninstalling pinecone-client-3.0.1:\n",
      "      Successfully uninstalled pinecone-client-3.0.1\n",
      "  Attempting uninstall: tiktoken\n",
      "    Found existing installation: tiktoken 0.5.2\n",
      "    Uninstalling tiktoken-0.5.2:\n",
      "      Successfully uninstalled tiktoken-0.5.2\n",
      "  Attempting uninstall: openai\n",
      "    Found existing installation: openai 1.10.0\n",
      "    Uninstalling openai-1.10.0:\n",
      "      Successfully uninstalled openai-1.10.0\n",
      "  Attempting uninstall: sentence-transformers\n",
      "    Found existing installation: sentence-transformers 2.2.2\n",
      "    Uninstalling sentence-transformers-2.2.2:\n",
      "      Successfully uninstalled sentence-transformers-2.2.2\n",
      "Successfully installed openai-1.16.1 pinecone-client-3.2.2 sentence-transformers-2.6.1 tiktoken-0.6.0\n",
      "\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m24.0\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3.11 -m pip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install pinecone-client openai sentence-transformers tiktoken datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c12182ab",
   "metadata": {
    "id": "c12182ab"
   },
   "outputs": [],
   "source": [
    "from openai import OpenAI\n",
    "from datetime import datetime\n",
    "import hashlib\n",
    "import re\n",
    "import os\n",
    "from sentence_transformers import CrossEncoder\n",
    "\n",
    "\n",
    "from tqdm import tqdm\n",
    "import numpy as np\n",
    "from torch import nn\n",
    "\n",
    "import logging\n",
    "from pinecone import Pinecone, ServerlessSpec\n",
    "\n",
    "logger = logging.getLogger()\n",
    "logger.setLevel(logging.DEBUG)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "793a562a-af04-46a4-ae9d-7ad1e1b6ea72",
   "metadata": {},
   "source": [
    "# Getting Started with Semantic Search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f0289c54",
   "metadata": {
    "id": "f0289c54"
   },
   "outputs": [],
   "source": [
    "pinecone_key = os.environ.get('PINECONE_API_KEY')\n",
    "client = OpenAI(\n",
    "    api_key=os.environ.get(\"OPENAI_API_KEY\")\n",
    ")\n",
    "\n",
    "INDEX_NAME = 'semantic-search-health-idx'\n",
    "NAMESPACE = 'default'\n",
    "ENGINE = 'text-embedding-3-large'  # has size 3072\n",
    "\n",
    "pc = Pinecone(\n",
    "    api_key=pinecone_key\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3cf993a3-6d51-49f4-8968-60f654d6202d",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "3cf993a3-6d51-49f4-8968-60f654d6202d",
    "outputId": "12a29199-9a84-4855-f07f-7979b15a718b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3072, 2)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# helper functions to get lists of embeddings from the OpenAI API\n",
    "def get_embeddings(texts, engine=ENGINE):\n",
    "    response = client.embeddings.create(\n",
    "        input=texts,\n",
    "        model=engine\n",
    "    )\n",
    "\n",
    "    return [d.embedding for d in list(response.data)]\n",
    "\n",
    "def get_embedding(text, engine=ENGINE):\n",
    "    return get_embeddings([text], engine)[0]\n",
    "\n",
    "len(get_embedding('hi')), len(get_embeddings(['hi', 'hello']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ea70672a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ea70672a",
    "outputId": "7ba237ca-f93f-4a8b-83a1-5133a33b3080"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pinecone.data.index.Index at 0x28d568890>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "if INDEX_NAME not in pc.list_indexes().names():\n",
    "    print(f'Creating index {INDEX_NAME}')\n",
    "    pc.create_index(\n",
    "        name=INDEX_NAME,  # The name of the index\n",
    "        dimension=3072,  # The dimensionality of the vectors for our OpenAI embedder\n",
    "        metric='cosine',  # The similarity metric to use when searching the index\n",
    "        spec=ServerlessSpec(\n",
    "            cloud='aws',\n",
    "            region='us-west-2'\n",
    "        )\n",
    "    )\n",
    "\n",
    "# Store the index as a variable\n",
    "index = pc.Index(name=INDEX_NAME)\n",
    "index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b6103d6c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "b6103d6c",
    "outputId": "549365a0-a4ee-4dff-b20e-8d2e07c97e13"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dimension': 3072,\n",
       " 'index_fullness': 0.0,\n",
       " 'namespaces': {},\n",
       " 'total_vector_count': 0}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.describe_index_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7f2fdfe7",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "7f2fdfe7",
    "outputId": "4c41491f-5ce8-415e-f792-e149e470b267"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'ae76cc4dfd345ecaeea9b8ba0d5c3437'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def my_hash(s):\n",
    "    # Return the MD5 hash of the input string as a hexadecimal string\n",
    "    return hashlib.md5(s.encode()).hexdigest()\n",
    "\n",
    "my_hash('I love to hash it')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ecd86f51",
   "metadata": {
    "id": "ecd86f51"
   },
   "outputs": [],
   "source": [
    "def prepare_for_pinecone(texts, engine=ENGINE):\n",
    "    # Get the current UTC date and time\n",
    "    now = datetime.utcnow()\n",
    "\n",
    "    # Generate vector embeddings for each string in the input list, using the specified engine\n",
    "    embeddings = get_embeddings(texts, engine=engine)\n",
    "\n",
    "    # Create tuples of (hash, embedding, metadata) for each input string and its corresponding vector embedding\n",
    "    # The my_hash() function is used to generate a unique hash for each string, and the datetime.utcnow() function is used to generate the current UTC date and time\n",
    "    return [\n",
    "        (\n",
    "            my_hash(text),  # A unique ID for each string, generated using the my_hash() function\n",
    "            embedding,  # The vector embedding of the string\n",
    "            dict(text=text, date_uploaded=now)  # A dictionary of metadata, including the original text and the current UTC date and time\n",
    "        )\n",
    "        for text, embedding in zip(texts, embeddings)  # Iterate over each input string and its corresponding vector embedding\n",
    "    ]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "4c40d99a",
   "metadata": {
    "id": "4c40d99a"
   },
   "outputs": [],
   "source": [
    "texts = ['hi']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "3e1b73f3",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "3e1b73f3",
    "outputId": "eeba3b54-c2e5-481c-f340-3e3802c79b38"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ID:   49f68a5c8493ec2c0bf489821c21fc3b \n",
      "LEN:  3072 \n",
      "META: {'text': 'hi', 'date_uploaded': datetime.datetime(2024, 4, 3, 14, 55, 29, 838076)}\n"
     ]
    }
   ],
   "source": [
    "_id, embedding, metadata = prepare_for_pinecone(texts)[0]\n",
    "\n",
    "print('ID:  ',_id, '\\nLEN: ', len(embedding), '\\nMETA:', metadata)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "357e1ce3-99d3-4517-abea-045ce72bbb35",
   "metadata": {},
   "source": [
    "## Uploading Text to our DB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "bf47aabd",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "bf47aabd",
    "outputId": "0cdc3643-d204-4d4f-9ed4-3f632a4328fb"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def upload_texts_to_pinecone(texts, namespace=NAMESPACE, batch_size=None, show_progress_bar=False):\n",
    "    # Call the prepare_for_pinecone function to prepare the input texts for indexing\n",
    "    total_upserted = 0\n",
    "    if not batch_size:\n",
    "        batch_size = len(texts)\n",
    "\n",
    "    _range = range(0, len(texts), batch_size)\n",
    "    for i in tqdm(_range) if show_progress_bar else _range:\n",
    "        batch = texts[i: i + batch_size]\n",
    "        prepared_texts = prepare_for_pinecone(batch)\n",
    "\n",
    "        # Use the upsert() method of the index object to upload the prepared texts to Pinecone\n",
    "        total_upserted += index.upsert(\n",
    "            vectors=prepared_texts,\n",
    "            namespace=namespace\n",
    "        )['upserted_count']\n",
    "\n",
    "\n",
    "    return total_upserted\n",
    "\n",
    "# Call the upload_texts_to_pinecone() function with the input texts\n",
    "upload_texts_to_pinecone(texts)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "b65cb2e1-f67d-4606-b0ff-c7b67aab04d8",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "b65cb2e1-f67d-4606-b0ff-c7b67aab04d8",
    "outputId": "721bfc5a-bf16-44f5-fd8e-5b8fddb3ef95"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dimension': 3072,\n",
       " 'index_fullness': 0.0,\n",
       " 'namespaces': {'default': {'vector_count': 978}},\n",
       " 'total_vector_count': 978}"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.describe_index_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "f64fa0cb-d704-42ef-b06f-5ac882a3a55a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f64fa0cb-d704-42ef-b06f-5ac882a3a55a",
    "outputId": "a20a6143-23e8-4f97-dd10-d4314c125bdb"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['hi']"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "texts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "e8ed7e8e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e8ed7e8e",
    "outputId": "35c9667c-897b-450e-a630-022d74ca868e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def query_from_pinecone(query, top_k=3, include_metadata=True):\n",
    "    # get embedding from THE SAME embedder as the documents\n",
    "    query_embedding = get_embedding(query, engine=ENGINE)\n",
    "\n",
    "    return index.query(\n",
    "      vector=query_embedding,\n",
    "      top_k=top_k,\n",
    "      namespace=NAMESPACE,\n",
    "      include_metadata=include_metadata   # gets the metadata (dates, text, etc)\n",
    "    ).get('matches')\n",
    "\n",
    "query_from_pinecone('hello', top_k=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "84a0871f",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "84a0871f",
    "outputId": "cf32ed08-348f-4be0-a647-cf0546ca0c2d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import hashlib\n",
    "\n",
    "def delete_texts_from_pinecone(texts, namespace=NAMESPACE):\n",
    "    # Compute the hash (id) for each text\n",
    "    hashes = [hashlib.md5(text.encode()).hexdigest() for text in texts]\n",
    "\n",
    "    # The ids parameter is used to specify the list of IDs (hashes) to delete\n",
    "    return index.delete(ids=hashes, namespace=namespace)\n",
    "\n",
    "# delete our text\n",
    "delete_texts_from_pinecone(texts)\n",
    "\n",
    "# test that the index is empty\n",
    "query_from_pinecone('hello')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "72729603",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "72729603",
    "outputId": "52312e5a-09d2-4047-c87d-02577e6a8089"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[36661, 1070]"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Importing the tiktoken library\n",
    "import tiktoken\n",
    "\n",
    "# Initializing a tokenizer for the 'cl100k_base' model\n",
    "# This tokenizer is designed to work with the 'ada-002' embedding model\n",
    "tokenizer = tiktoken.get_encoding(\"cl100k_base\")\n",
    "\n",
    "# Using the tokenizer to encode the text 'hey there'\n",
    "# The resulting output is a list of integers representing the encoded text\n",
    "# This is the input format required for embedding using the 'ada-002' model\n",
    "tokenizer.encode('hey there')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bbc147d8",
   "metadata": {
    "id": "bbc147d8"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6d7ed0ec-0db5-4cad-b867-4c0b528b63ed",
   "metadata": {},
   "source": [
    "# Chunking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "84f34d65",
   "metadata": {
    "id": "84f34d65"
   },
   "outputs": [],
   "source": [
    "# Function to split the text into chunks of a maximum number of tokens. Inspired by OpenAI\n",
    "def overlapping_chunks(text, max_tokens = 500, overlapping_factor = 5):\n",
    "    '''\n",
    "    max_tokens: tokens we want per chunk\n",
    "    overlapping_factor: number of sentences to start each chunk with that overlaps with the previous chunk\n",
    "    '''\n",
    "\n",
    "    # Split the text using punctuation\n",
    "    sentences = re.split(r'[.?!]', text)\n",
    "\n",
    "    # Get the number of tokens for each sentence\n",
    "    n_tokens = [len(tokenizer.encode(\" \" + sentence)) for sentence in sentences]\n",
    "\n",
    "    chunks, tokens_so_far, chunk = [], 0, []\n",
    "\n",
    "    # Loop through the sentences and tokens joined together in a tuple\n",
    "    for sentence, token in zip(sentences, n_tokens):\n",
    "\n",
    "        # If the number of tokens so far plus the number of tokens in the current sentence is greater\n",
    "        # than the max number of tokens, then add the chunk to the list of chunks and reset\n",
    "        # the chunk and tokens so far\n",
    "        if tokens_so_far + token > max_tokens:\n",
    "            chunks.append(\". \".join(chunk) + \".\")\n",
    "            if overlapping_factor > 0:\n",
    "                chunk = chunk[-overlapping_factor:]\n",
    "                tokens_so_far = sum([len(tokenizer.encode(c)) for c in chunk])\n",
    "            else:\n",
    "                chunk = []\n",
    "                tokens_so_far = 0\n",
    "\n",
    "        # If the number of tokens in the current sentence is greater than the max number of\n",
    "        # tokens, go to the next sentence\n",
    "        if token > max_tokens:\n",
    "            continue\n",
    "\n",
    "        # Otherwise, add the sentence to the chunk and add the number of tokens to the total\n",
    "        chunk.append(sentence)\n",
    "        tokens_so_far += token + 1\n",
    "    if chunk:\n",
    "        chunks.append(\". \".join(chunk) + \".\")\n",
    "\n",
    "    return chunks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "3164e0ba",
   "metadata": {
    "id": "3164e0ba"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████| 428/428 [06:28<00:00,  1.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "575490\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "import PyPDF2\n",
    "\n",
    "# Open the PDF file in read-binary mode\n",
    "with open('../data/pds2.pdf', 'rb') as file:\n",
    "\n",
    "    # Create a PDF reader object\n",
    "    reader = PyPDF2.PdfReader(file)\n",
    "\n",
    "    # Initialize an empty string to hold the text\n",
    "    principles_of_ds = ''\n",
    "    # Loop through each page in the PDF file\n",
    "    for page in tqdm(reader.pages):\n",
    "        text = page.extract_text()\n",
    "        principles_of_ds += '\\n\\n' + text[text.find(' ]')+2:]\n",
    "\n",
    "# Print the final string containing all the text from the PDF file\n",
    "principles_of_ds = principles_of_ds.strip()\n",
    "\n",
    "print(len(principles_of_ds))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "dfdac6ff",
   "metadata": {
    "id": "dfdac6ff"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "non-overlapping chunking approach has 286 documents with average length 474.1 tokens\n"
     ]
    }
   ],
   "source": [
    "split = overlapping_chunks(principles_of_ds, overlapping_factor=0)\n",
    "avg_length = sum([len(tokenizer.encode(t)) for t in split]) / len(split)\n",
    "print(f'non-overlapping chunking approach has {len(split)} documents with average length {avg_length:.1f} tokens')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d0d7b21-c1f5-4b37-b430-85b966107b6b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "1730b34b",
   "metadata": {
    "id": "1730b34b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "overlapping chunking approach has 392 documents with average length 485.2 tokens\n"
     ]
    }
   ],
   "source": [
    "split = overlapping_chunks(principles_of_ds)\n",
    "avg_length = sum([len(tokenizer.encode(t)) for t in split]) / len(split)\n",
    "print(f'overlapping chunking approach has {len(split)} documents with average length {avg_length:.1f} tokens')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "758cb1ff",
   "metadata": {
    "id": "758cb1ff"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "7fc7dc10",
   "metadata": {
    "id": "7fc7dc10"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(' ', 82259), ('\\n', 9220), ('  ', 1592), ('\\n\\n', 333), ('\\n   ', 250), ('\\n\\n\\n', 82), ('\\n    ', 73), ('\\n ', 46), (' \\n', 39), ('     ', 34)]\n"
     ]
    }
   ],
   "source": [
    "# Importing the Counter and re libraries\n",
    "from collections import Counter\n",
    "import re\n",
    "\n",
    "# Find all occurrences of one or more spaces in 'principles_of_ds'\n",
    "matches = re.findall(r'[\\s]{1,}', principles_of_ds)\n",
    "\n",
    "# The 10 most frequent spaces that occur in the document\n",
    "most_common_spaces = Counter(matches).most_common(10)\n",
    "\n",
    "# Print the most common spaces and their frequencies\n",
    "print(most_common_spaces)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "63ccc253",
   "metadata": {
    "id": "63ccc253"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "custom delimiter approach has 426 documents with average length 316.3 tokens\n"
     ]
    }
   ],
   "source": [
    "# Only keep documents of at least 100 characters split by a custom delimiter\n",
    "split = list(filter(lambda x: len(x) > 50, principles_of_ds.split('\\n\\n')))\n",
    "\n",
    "avg_length = sum([len(tokenizer.encode(t)) for t in split]) / len(split)\n",
    "print(f'custom delimiter approach has {len(split)} documents with average length {avg_length:.1f} tokens')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "2480fe3c",
   "metadata": {
    "id": "2480fe3c"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:04<00:00,  1.00it/s]\n"
     ]
    }
   ],
   "source": [
    "embeddings = None\n",
    "for s in tqdm(range(0, len(split), 100)):\n",
    "    if embeddings is None:\n",
    "        embeddings = np.array(get_embeddings(split[s:s+100], engine=ENGINE))\n",
    "    else:\n",
    "        embeddings = np.vstack([embeddings, np.array(get_embeddings(split[s:s+100], engine=ENGINE))])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "8d6f0d09-5f2c-42e0-bddf-f98c042744f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2974705542613624"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.percentile(cosine_sim_matrix.flatten(), 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "685dc753-ef49-40c9-853f-3eab38f60d09",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "1cfe8fd2",
   "metadata": {
    "id": "1cfe8fd2",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster 0: 3 embeddings\n",
      "Cluster 1: 2 embeddings\n",
      "Cluster 2: 2 embeddings\n",
      "Cluster 3: 2 embeddings\n",
      "Cluster 4: 3 embeddings\n",
      "Cluster 5: 2 embeddings\n",
      "Cluster 6: 3 embeddings\n",
      "Cluster 7: 2 embeddings\n",
      "Cluster 8: 2 embeddings\n",
      "Cluster 9: 3 embeddings\n",
      "Cluster 10: 2 embeddings\n",
      "Cluster 11: 2 embeddings\n",
      "Cluster 12: 2 embeddings\n",
      "Cluster 13: 2 embeddings\n",
      "Cluster 14: 2 embeddings\n",
      "Cluster 15: 2 embeddings\n",
      "Cluster 16: 2 embeddings\n",
      "Cluster 17: 3 embeddings\n",
      "Cluster 18: 2 embeddings\n",
      "Cluster 19: 2 embeddings\n",
      "Cluster 20: 2 embeddings\n",
      "Cluster 21: 2 embeddings\n",
      "Cluster 22: 2 embeddings\n",
      "Cluster 23: 3 embeddings\n",
      "Cluster 24: 3 embeddings\n",
      "Cluster 25: 2 embeddings\n",
      "Cluster 26: 2 embeddings\n",
      "Cluster 27: 2 embeddings\n",
      "Cluster 28: 2 embeddings\n",
      "Cluster 29: 2 embeddings\n",
      "Cluster 30: 2 embeddings\n",
      "Cluster 31: 2 embeddings\n",
      "Cluster 32: 2 embeddings\n",
      "Cluster 33: 3 embeddings\n",
      "Cluster 34: 2 embeddings\n",
      "Cluster 35: 2 embeddings\n",
      "Cluster 36: 2 embeddings\n",
      "Cluster 37: 2 embeddings\n",
      "Cluster 38: 2 embeddings\n",
      "Cluster 39: 2 embeddings\n",
      "Cluster 40: 2 embeddings\n",
      "Cluster 41: 2 embeddings\n",
      "Cluster 42: 2 embeddings\n",
      "Cluster 43: 2 embeddings\n",
      "Cluster 44: 2 embeddings\n",
      "Cluster 45: 2 embeddings\n",
      "Cluster 46: 3 embeddings\n",
      "Cluster 47: 2 embeddings\n",
      "Cluster 48: 3 embeddings\n",
      "Cluster 49: 3 embeddings\n",
      "Cluster 50: 2 embeddings\n",
      "Cluster 51: 2 embeddings\n",
      "Cluster 52: 2 embeddings\n",
      "Cluster 53: 2 embeddings\n",
      "Cluster 54: 2 embeddings\n",
      "Cluster 55: 2 embeddings\n",
      "Cluster 56: 3 embeddings\n",
      "Cluster 57: 2 embeddings\n",
      "Cluster 58: 2 embeddings\n",
      "Cluster 59: 2 embeddings\n",
      "Cluster 60: 2 embeddings\n",
      "Cluster 61: 2 embeddings\n",
      "Cluster 62: 1 embeddings\n",
      "Cluster 63: 2 embeddings\n",
      "Cluster 64: 2 embeddings\n",
      "Cluster 65: 2 embeddings\n",
      "Cluster 66: 1 embeddings\n",
      "Cluster 67: 2 embeddings\n",
      "Cluster 68: 1 embeddings\n",
      "Cluster 69: 2 embeddings\n",
      "Cluster 70: 2 embeddings\n",
      "Cluster 71: 3 embeddings\n",
      "Cluster 72: 3 embeddings\n",
      "Cluster 73: 2 embeddings\n",
      "Cluster 74: 1 embeddings\n",
      "Cluster 75: 2 embeddings\n",
      "Cluster 76: 2 embeddings\n",
      "Cluster 77: 2 embeddings\n",
      "Cluster 78: 2 embeddings\n",
      "Cluster 79: 2 embeddings\n",
      "Cluster 80: 1 embeddings\n",
      "Cluster 81: 1 embeddings\n",
      "Cluster 82: 2 embeddings\n",
      "Cluster 83: 2 embeddings\n",
      "Cluster 84: 3 embeddings\n",
      "Cluster 85: 2 embeddings\n",
      "Cluster 86: 2 embeddings\n",
      "Cluster 87: 2 embeddings\n",
      "Cluster 88: 1 embeddings\n",
      "Cluster 89: 1 embeddings\n",
      "Cluster 90: 1 embeddings\n",
      "Cluster 91: 2 embeddings\n",
      "Cluster 92: 2 embeddings\n",
      "Cluster 93: 2 embeddings\n",
      "Cluster 94: 3 embeddings\n",
      "Cluster 95: 1 embeddings\n",
      "Cluster 96: 2 embeddings\n",
      "Cluster 97: 3 embeddings\n",
      "Cluster 98: 2 embeddings\n",
      "Cluster 99: 2 embeddings\n",
      "Cluster 100: 2 embeddings\n",
      "Cluster 101: 2 embeddings\n",
      "Cluster 102: 3 embeddings\n",
      "Cluster 103: 1 embeddings\n",
      "Cluster 104: 2 embeddings\n",
      "Cluster 105: 2 embeddings\n",
      "Cluster 106: 1 embeddings\n",
      "Cluster 107: 2 embeddings\n",
      "Cluster 108: 2 embeddings\n",
      "Cluster 109: 2 embeddings\n",
      "Cluster 110: 1 embeddings\n",
      "Cluster 111: 1 embeddings\n",
      "Cluster 112: 1 embeddings\n",
      "Cluster 113: 1 embeddings\n",
      "Cluster 114: 2 embeddings\n",
      "Cluster 115: 2 embeddings\n",
      "Cluster 116: 1 embeddings\n",
      "Cluster 117: 1 embeddings\n",
      "Cluster 118: 2 embeddings\n",
      "Cluster 119: 1 embeddings\n",
      "Cluster 120: 1 embeddings\n",
      "Cluster 121: 1 embeddings\n",
      "Cluster 122: 2 embeddings\n",
      "Cluster 123: 1 embeddings\n",
      "Cluster 124: 1 embeddings\n",
      "Cluster 125: 1 embeddings\n",
      "Cluster 126: 1 embeddings\n",
      "Cluster 127: 1 embeddings\n",
      "Cluster 128: 1 embeddings\n",
      "Cluster 129: 1 embeddings\n",
      "Cluster 130: 1 embeddings\n",
      "Cluster 131: 1 embeddings\n",
      "Cluster 132: 1 embeddings\n",
      "Cluster 133: 1 embeddings\n",
      "Cluster 134: 1 embeddings\n",
      "Cluster 135: 2 embeddings\n",
      "Cluster 136: 2 embeddings\n",
      "Cluster 137: 1 embeddings\n",
      "Cluster 138: 1 embeddings\n",
      "Cluster 139: 1 embeddings\n",
      "Cluster 140: 1 embeddings\n",
      "Cluster 141: 1 embeddings\n",
      "Cluster 142: 1 embeddings\n",
      "Cluster 143: 1 embeddings\n",
      "Cluster 144: 1 embeddings\n",
      "Cluster 145: 1 embeddings\n",
      "Cluster 146: 1 embeddings\n",
      "Cluster 147: 1 embeddings\n",
      "Cluster 148: 1 embeddings\n",
      "Cluster 149: 1 embeddings\n",
      "Cluster 150: 1 embeddings\n",
      "Cluster 151: 1 embeddings\n",
      "Cluster 152: 1 embeddings\n",
      "Cluster 153: 1 embeddings\n",
      "Cluster 154: 1 embeddings\n",
      "Cluster 155: 1 embeddings\n",
      "Cluster 156: 3 embeddings\n",
      "Cluster 157: 1 embeddings\n",
      "Cluster 158: 1 embeddings\n",
      "Cluster 159: 1 embeddings\n",
      "Cluster 160: 1 embeddings\n",
      "Cluster 161: 1 embeddings\n",
      "Cluster 162: 1 embeddings\n",
      "Cluster 163: 1 embeddings\n",
      "Cluster 164: 1 embeddings\n",
      "Cluster 165: 1 embeddings\n",
      "Cluster 166: 1 embeddings\n",
      "Cluster 167: 1 embeddings\n",
      "Cluster 168: 1 embeddings\n",
      "Cluster 169: 1 embeddings\n",
      "Cluster 170: 1 embeddings\n",
      "Cluster 171: 2 embeddings\n",
      "Cluster 172: 1 embeddings\n",
      "Cluster 173: 1 embeddings\n",
      "Cluster 174: 1 embeddings\n",
      "Cluster 175: 1 embeddings\n",
      "Cluster 176: 1 embeddings\n",
      "Cluster 177: 1 embeddings\n",
      "Cluster 178: 1 embeddings\n",
      "Cluster 179: 1 embeddings\n",
      "Cluster 180: 1 embeddings\n",
      "Cluster 181: 1 embeddings\n",
      "Cluster 182: 1 embeddings\n",
      "Cluster 183: 1 embeddings\n",
      "Cluster 184: 1 embeddings\n",
      "Cluster 185: 1 embeddings\n",
      "Cluster 186: 2 embeddings\n",
      "Cluster 187: 1 embeddings\n",
      "Cluster 188: 1 embeddings\n",
      "Cluster 189: 1 embeddings\n",
      "Cluster 190: 1 embeddings\n",
      "Cluster 191: 1 embeddings\n",
      "Cluster 192: 1 embeddings\n",
      "Cluster 193: 1 embeddings\n",
      "Cluster 194: 1 embeddings\n",
      "Cluster 195: 3 embeddings\n",
      "Cluster 196: 1 embeddings\n",
      "Cluster 197: 2 embeddings\n",
      "Cluster 198: 1 embeddings\n",
      "Cluster 199: 1 embeddings\n",
      "Cluster 200: 2 embeddings\n",
      "Cluster 201: 1 embeddings\n",
      "Cluster 202: 2 embeddings\n",
      "Cluster 203: 2 embeddings\n",
      "Cluster 204: 1 embeddings\n",
      "Cluster 205: 1 embeddings\n",
      "Cluster 206: 1 embeddings\n",
      "Cluster 207: 1 embeddings\n",
      "Cluster 208: 1 embeddings\n",
      "Cluster 209: 1 embeddings\n",
      "Cluster 210: 1 embeddings\n",
      "Cluster 211: 1 embeddings\n",
      "Cluster 212: 2 embeddings\n",
      "Cluster 213: 1 embeddings\n",
      "Cluster 214: 1 embeddings\n",
      "Cluster 215: 1 embeddings\n",
      "Cluster 216: 1 embeddings\n",
      "Cluster 217: 1 embeddings\n",
      "Cluster 218: 1 embeddings\n",
      "Cluster 219: 1 embeddings\n",
      "Cluster 220: 1 embeddings\n",
      "Cluster 221: 1 embeddings\n",
      "Cluster 222: 1 embeddings\n",
      "Cluster 223: 1 embeddings\n",
      "Cluster 224: 1 embeddings\n",
      "Cluster 225: 1 embeddings\n",
      "Cluster 226: 1 embeddings\n",
      "Cluster 227: 1 embeddings\n",
      "Cluster 228: 1 embeddings\n",
      "Cluster 229: 1 embeddings\n",
      "Cluster 230: 1 embeddings\n",
      "Cluster 231: 1 embeddings\n",
      "Cluster 232: 1 embeddings\n",
      "Cluster 233: 1 embeddings\n",
      "Cluster 234: 1 embeddings\n",
      "Cluster 235: 1 embeddings\n",
      "Cluster 236: 1 embeddings\n",
      "Cluster 237: 1 embeddings\n",
      "Cluster 238: 1 embeddings\n",
      "Cluster 239: 1 embeddings\n",
      "Cluster 240: 1 embeddings\n",
      "Cluster 241: 1 embeddings\n",
      "Cluster 242: 1 embeddings\n",
      "Cluster 243: 1 embeddings\n",
      "Cluster 244: 1 embeddings\n",
      "Cluster 245: 1 embeddings\n",
      "Cluster 246: 1 embeddings\n",
      "Cluster 247: 1 embeddings\n",
      "Cluster 248: 1 embeddings\n",
      "Cluster 249: 1 embeddings\n",
      "Cluster 250: 1 embeddings\n",
      "Cluster 251: 1 embeddings\n",
      "Cluster 252: 1 embeddings\n",
      "Cluster 253: 1 embeddings\n",
      "Cluster 254: 1 embeddings\n",
      "Cluster 255: 1 embeddings\n",
      "Cluster 256: 1 embeddings\n",
      "Cluster 257: 1 embeddings\n",
      "Cluster 258: 1 embeddings\n",
      "Cluster 259: 1 embeddings\n",
      "Cluster 260: 1 embeddings\n",
      "Cluster 261: 1 embeddings\n",
      "Cluster 262: 1 embeddings\n",
      "Cluster 263: 1 embeddings\n",
      "Cluster 264: 1 embeddings\n",
      "Cluster 265: 1 embeddings\n",
      "Cluster 266: 1 embeddings\n",
      "Cluster 267: 1 embeddings\n",
      "Cluster 268: 1 embeddings\n",
      "Cluster 269: 1 embeddings\n",
      "Cluster 270: 1 embeddings\n",
      "Cluster 271: 1 embeddings\n",
      "Cluster 272: 2 embeddings\n",
      "Cluster 273: 1 embeddings\n",
      "Cluster 274: 1 embeddings\n",
      "Cluster 275: 1 embeddings\n",
      "Cluster 276: 1 embeddings\n",
      "Cluster 277: 1 embeddings\n",
      "Cluster 278: 1 embeddings\n",
      "Cluster 279: 1 embeddings\n",
      "Cluster 280: 1 embeddings\n",
      "Cluster 281: 1 embeddings\n",
      "Cluster 282: 1 embeddings\n",
      "Cluster 283: 1 embeddings\n",
      "Cluster 284: 1 embeddings\n",
      "Cluster 285: 1 embeddings\n",
      "Cluster 286: 1 embeddings\n",
      "Cluster 287: 1 embeddings\n",
      "Cluster 288: 1 embeddings\n",
      "Cluster 289: 1 embeddings\n",
      "Cluster 290: 1 embeddings\n",
      "Cluster 291: 1 embeddings\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/cluster/_agglomerative.py:1005: FutureWarning: Attribute `affinity` was deprecated in version 1.2 and will be removed in 1.4. Use `metric` instead\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "import numpy as np\n",
    "\n",
    "# Assume you have a list of text embeddings called `embeddings`\n",
    "# First, compute the cosine similarity matrix between all pairs of embeddings\n",
    "cosine_sim_matrix = cosine_similarity(embeddings)\n",
    "\n",
    "# Instantiate the AgglomerativeClustering model\n",
    "agg_clustering = AgglomerativeClustering(\n",
    "    n_clusters=None,         # the algorithm will determine the optimal number of clusters based on the data\n",
    "    \n",
    "    # clusters will be formed until all pairwise distances between clusters are greater than given value\n",
    "    distance_threshold=np.percentile(cosine_sim_matrix.flatten(), 50),\n",
    "    \n",
    "    affinity='precomputed',  # we are providing a precomputed distance matrix (1 - similarity matrix) as input\n",
    "    linkage='complete'       # form clusters by iteratively merging the smallest clusters based on the maximum distance between their components\n",
    ")\n",
    "\n",
    "# Fit the model to the cosine distance matrix (1 - similarity matrix)\n",
    "agg_clustering.fit(1 - cosine_sim_matrix)\n",
    "\n",
    "# Get the cluster labels for each embedding\n",
    "cluster_labels = agg_clustering.labels_\n",
    "\n",
    "# Print the number of embeddings in each cluster\n",
    "unique_labels, counts = np.unique(cluster_labels, return_counts=True)\n",
    "for label, count in zip(unique_labels, counts):\n",
    "    print(f'Cluster {label}: {count} embeddings')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "0727319b",
   "metadata": {
    "id": "0727319b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Our pruning approach has 292 documents with average length 461.6 tokens\n"
     ]
    }
   ],
   "source": [
    "pruned_documents = []\n",
    "for _label, count in zip(unique_labels, counts):\n",
    "    pruned_documents.append('\\n\\n'.join([text for text, label in zip(split, cluster_labels) if label == _label]))\n",
    "\n",
    "\n",
    "avg_length = sum([len(tokenizer.encode(t)) for t in pruned_documents]) / len(pruned_documents)\n",
    "print(f'Our pruning approach has {len(pruned_documents)} documents with average length {avg_length:.1f} tokens')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "97f80b78",
   "metadata": {
    "id": "97f80b78",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "So, we have way more ham messages than we do spam. Because this is a classification\n",
      "problem, it will be very useful to know our null accuracy rate , which is the percentage\n",
      "chance of predicting a single row correctly if we keep guessing the most common class,\n",
      "ham:\n",
      "df.label.value_counts() / df.shape[0]\n",
      "ham     0.865937\n",
      "spam    0.134063\n",
      "So if we blindly guessed ham, we would be correct about 87% of the time, but we can do\n",
      "better than that. If we have a set of classes, C, and features, xi, then we can use Bayes'\n",
      "theorem to predict the probability that a single row belongs to class C, using the following\n",
      "formula:\n",
      "Let's look at this formula in a little more detail:\n",
      "P(class C | {xi}) : The posterior probability is the probability that the row belongs\n",
      "to class C  given the features {xi}.\n",
      "P({xi} | class C) : This is the likelihood that we would observe these features given\n",
      "that the row was in class C .\n",
      "P(class C) : This is the prior probability. It is the probability that the data point\n",
      "belongs to class C  before we see any data.\n",
      "P({xi}) : This is our normalization constant.\n",
      "For example, imagine we have an email with three words: send cash now.  We'll use\n",
      "Naive Bayes to classify the email as either being spam or ham:\n",
      "We are concerned with the difference of these two numbers. We can use the following\n",
      "criteria to classify any single text sample:\n",
      "If P(spam | send cash now)  is larger than P(ham | send cash now) ,\n",
      "then we will classify the text as spam\n",
      "\n",
      "If P(ham | send cash now)  is larger than P(spam | send cash now) , then\n",
      "we will label the text ham\n",
      "Because both equations have P (send money now)  in the denominator, we can ignore them.\n",
      "So, now we are concerned with the following:\n",
      "Let's figure out the numbers  in this equation:\n",
      "P(spam) = 0.134063\n",
      "P(ham) = 0.865937\n",
      "P(send cash now | spam)\n",
      "P(send cash now | ham)\n",
      "The final two likelihoods might seem like they would not be so difficult to calculate. All we\n",
      "have to do is count the numbers of spam messages that include the send money\n",
      "now phrase and divide that by the total number of spam messages:\n",
      "df.msg = df.msg.apply(lambda x:x.lower())\n",
      "# make all strings lower case so we can search easier\n",
      "df[df.msg.str.contains('send cash now')] .shape\n",
      "(0, 2)\n",
      "Oh no! There are none! There are literally zero texts with the exact phrase send cash now .\n",
      "The hidden problem here is that this phrase is very specific and we can't assume that we\n",
      "will have enough data in the world to have seen this exact phrase many times before.\n",
      "Instead, we can make a naïve assumption  in our Bayes' theorem. If we assume that the\n",
      "features (words) are conditionally independent (meaning that no word affects the existence\n",
      "of another word), then we can rewrite the formula:\n",
      "spams = df[df.label == 'spam']\n",
      "for word in ['send', 'cash', 'now']:\n",
      "    print( word, spams[spams.msg.str.contains(word)].shape[0] /\n",
      "float(spams.shape[0]))\n",
      "\n",
      "P(send|spam) = 0.096\n",
      "P(cash |spam) = 0.091\n",
      "P(now|spam) = 0.280\n",
      "With this, we can calculate the following:\n",
      "P(send cash now| spam) ∗ P(spam)= (.096 ∗.091∗.280) ∗.134 = 0.00032\n",
      "Repeating the same procedure for ham gives us the following:\n",
      "P(send|ham) = 0.03\n",
      "P(cash|ham) = 0.003\n",
      "P(now|ham) = 0.109\n",
      "The fact that these numbers are both very low is not as important as the fact that the spam\n",
      "probability is much larger than the ham calculation. If we calculate 0.00032 / 0.0000084 =\n",
      "38.1 we see that the send cash now  probability for spam is 38 times higher than for spam.\n",
      "Doing this means that we can classify send cash now  as spam! Simple, right?\n",
      "Let's use Python to implement a Naive Bayes classifier without having to do all of these\n",
      "calculations ourselves.\n",
      "First, let's revisit the count vectorizer in scikit-learn, which turns text into numerical data\n",
      "for us. Let's assume that we will train on three documents (sentences):\n",
      "# simple count vectorizer example\n",
      "from sklearn.feature_extraction.text import CountVectorizer\n",
      "# start with a simple example\n",
      "train_simple = ['call you tonight',\n",
      "                'Call me a cab',\n",
      "                'please call me... PLEASE 44!']\n",
      "# learn the 'vocabulary' of the training data\n",
      "vect = CountVectorizer()\n",
      "train_simple_dtm = vect.fit_transform(train_simple)\n",
      "pd.DataFrame(train_simple_dtm.toarray(), columns=vect.get_feature_names())\n"
     ]
    }
   ],
   "source": [
    "print(pruned_documents[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "3a3364dd",
   "metadata": {
    "id": "3a3364dd"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "292"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "upload_texts_to_pinecone(pruned_documents, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8511c945",
   "metadata": {
    "id": "8511c945"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "13cc8bdc",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "13cc8bdc",
    "outputId": "b34ad588-4ae4-4eed-e6bc-e048f780ca12"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30282dd530d9a170ec50c8252d337b33\t0.56\tLet's begin by learning a very  important value  i\n",
      "126c5edbf7271df6e2d9983cd740ba10\t0.50\tWe get this chart:\n",
      "We can see that we have negativ\n",
      "7da124bb8c927e0c392cf75d4bfb2c76\t0.41\t\n",
      "Basic Statistics\n",
      "This chapter will focus on the s\n"
     ]
    }
   ],
   "source": [
    "query = 'How do z scores work?'\n",
    "\n",
    "results_from_pinecone = query_from_pinecone(query, top_k=3)\n",
    "\n",
    "for result_from_pinecone in results_from_pinecone:\n",
    "    print(f\"{result_from_pinecone['id']}\\t{result_from_pinecone['score']:.2f}\\t{result_from_pinecone['metadata']['text'][:50]}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "59780b12-f643-4d78-b298-123f4d7ff384",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'id': '7da124bb8c927e0c392cf75d4bfb2c76',\n",
       " 'metadata': {'date_uploaded': '2024-04-03T15:07:13.510992',\n",
       "              'text': '\\n'\n",
       "                      'Basic Statistics\\n'\n",
       "                      'This chapter will focus on the statistical knowledge '\n",
       "                      'required by any aspiring data scientist.\\n'\n",
       "                      'We will explore ways of sampling and obtaining data '\n",
       "                      'without being affected by bias and\\n'\n",
       "                      'then use measures of statistics to quantify and '\n",
       "                      'visualize our data. Using the z-score and the\\n'\n",
       "                      'empirical rule, we will see how we can standardize data '\n",
       "                      'for the purpose of both graphing\\n'\n",
       "                      'and interpretability.\\n'\n",
       "                      'In this chapter, we will look at the following topics:\\n'\n",
       "                      'How to obtain and sample data\\n'\n",
       "                      'The measures of center, variance, and relative '\n",
       "                      'standing\\n'\n",
       "                      'Normalization of data using the z-score\\n'\n",
       "                      'The empirical rule\\n'\n",
       "                      'What are statistics?\\n'\n",
       "                      'This might seem like an odd question to ask, but I am '\n",
       "                      'frequently surprised by the number\\n'\n",
       "                      'of people who cannot answer this simple and yet '\n",
       "                      'powerful question: what are statistics?\\n'\n",
       "                      'Statistics are the numbers you always see on the news '\n",
       "                      'and in the paper. Statistics are useful\\n'\n",
       "                      'when trying to prove a point or trying to scare '\n",
       "                      'someone, but what are they?\\n'\n",
       "                      'To answer this question, we need to back up for a '\n",
       "                      'minute and talk about why we even\\n'\n",
       "                      'measure them in the first place. The goal of this field '\n",
       "                      'is to try to explain and model the\\n'\n",
       "                      'world around us. To do that, we have to take a look at '\n",
       "                      'the population.\\n'\n",
       "                      'We can define a population  as the entire pool of '\n",
       "                      'subjects of an experiment or a model.\\n'\n",
       "                      '\\n'\n",
       "                      'Summary\\n'\n",
       "                      'In this chapter, we covered much of the basic '\n",
       "                      'statistics required by most data scientists.\\n'\n",
       "                      'Everything from how we obtain/sample data to how to '\n",
       "                      'standardize data according to the z-\\n'\n",
       "                      'score and applications of the empirical rule was '\n",
       "                      'covered. We also reviewed how to take\\n'\n",
       "                      'samples for data analysis. In addition, we reviewed '\n",
       "                      'various statistical measures, such as the\\n'\n",
       "                      'mean and standard deviation, that help describe data.\\n'\n",
       "                      'In the next chapter, we will look at much more advanced '\n",
       "                      'applications of statistics. One\\n'\n",
       "                      'thing that we will consider is how to use hypothesis '\n",
       "                      'tests on data that we can assume to be\\n'\n",
       "                      'normal. As we use these tests, we will also quantify '\n",
       "                      'our errors and identify the best\\n'\n",
       "                      'practices to solve these errors.'},\n",
       " 'score': 0.411503136,\n",
       " 'values': []}"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_from_pinecone"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "c018b64d",
   "metadata": {
    "id": "c018b64d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{}"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "delete_texts_from_pinecone(pruned_documents)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "33c716a0",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "33c716a0",
    "outputId": "6700983d-b77d-4d78-904c-2500ffc8ccd0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dimension': 3072,\n",
       " 'index_fullness': 0.0,\n",
       " 'namespaces': {},\n",
       " 'total_vector_count': 0}"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.describe_index_stats()  # no vectors inside"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "40d020ce-6f8a-4ca5-9aaf-48715df47a33",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ffda611c-743b-4bc2-b7c2-33f2a643af37",
   "metadata": {},
   "source": [
    "# Working with Open-Source Cross-Encoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "074fab6f",
   "metadata": {
    "id": "074fab6f"
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "This example computes the score between a query and all possible\n",
    "sentences in a corpus using a Cross-Encoder for semantic textual similarity (STS).\n",
    "It output then the most similar sentences for the given query.\n",
    "\"\"\"\n",
    "from sentence_transformers.cross_encoder import CrossEncoder\n",
    "import numpy as np\n",
    "from torch import nn\n",
    "\n",
    "# Pre-trained cross encoder\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "25b8e942",
   "metadata": {
    "id": "25b8e942"
   },
   "outputs": [],
   "source": [
    "# remove debug and info logs\n",
    "import logging\n",
    "logging.basicConfig(level=logging.WARNING)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "118a061c-db0a-4903-8260-46c6f15189ae",
   "metadata": {
    "id": "118a061c-db0a-4903-8260-46c6f15189ae"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5865c184-ff37-40c4-b93c-30bb233f1cbd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatasetDict({\n",
       "    test: Dataset({\n",
       "        features: ['id', 'title', 'context', 'question', 'answers'],\n",
       "        num_rows: 1148\n",
       "    })\n",
       "    train: Dataset({\n",
       "        features: ['id', 'title', 'context', 'question', 'answers'],\n",
       "        num_rows: 11590\n",
       "    })\n",
       "})"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "dataset = load_dataset(\"xtreme\", \"MLQA.en.en\")\n",
    "\n",
    "dataset['train'] = dataset['test']\n",
    "\n",
    "# just renaming out of habit, sorry\n",
    "dataset['test'] = dataset['validation']\n",
    "del dataset['validation']\n",
    "\n",
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e622fbfa-2dde-4711-9c46-1390eb3430f9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e622fbfa-2dde-4711-9c46-1390eb3430f9",
    "outputId": "8e7d6133-76ce-493a-963f-f29e05c6e0ba"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'id': 'a4968ca8a18de16aa3859be760e43dbd3af3fce9',\n",
       "  'title': 'Area 51',\n",
       "  'context': 'In 1994, five unnamed civilian contractors and the widows of contractors Walter Kasza and Robert Frost sued the USAF and the United States Environmental Protection Agency. Their suit, in which they were represented by George Washington University law professor Jonathan Turley, alleged they had been present when large quantities of unknown chemicals had been burned in open pits and trenches at Groom. Biopsies taken from the complainants were analyzed by Rutgers University biochemists, who found high levels of dioxin, dibenzofuran, and trichloroethylene in their body fat. The complainants alleged they had sustained skin, liver, and respiratory injuries due to their work at Groom, and that this had contributed to the deaths of Frost and Kasza. The suit sought compensation for the injuries they had sustained, claiming the USAF had illegally handled toxic materials, and that the EPA had failed in its duty to enforce the Resource Conservation and Recovery Act (which governs handling of dangerous materials). They also sought detailed information about the chemicals to which they were allegedly exposed, hoping this would facilitate the medical treatment of survivors. Congressman Lee H. Hamilton, former chairman of the House Intelligence Committee, told 60 Minutes reporter Lesley Stahl, \"The Air Force is classifying all information about Area 51 in order to protect themselves from a lawsuit.\"',\n",
       "  'question': 'Who analyzed the biopsies?',\n",
       "  'answers': {'answer_start': [457],\n",
       "   'text': ['Rutgers University biochemists']}},\n",
       " {'id': 'f251ea56c4f1aa1df270137f7e6d89c0cc1b6ef4',\n",
       "  'title': 'Area 51',\n",
       "  'context': 'In 1994, five unnamed civilian contractors and the widows of contractors Walter Kasza and Robert Frost sued the USAF and the United States Environmental Protection Agency. Their suit, in which they were represented by George Washington University law professor Jonathan Turley, alleged they had been present when large quantities of unknown chemicals had been burned in open pits and trenches at Groom. Biopsies taken from the complainants were analyzed by Rutgers University biochemists, who found high levels of dioxin, dibenzofuran, and trichloroethylene in their body fat. The complainants alleged they had sustained skin, liver, and respiratory injuries due to their work at Groom, and that this had contributed to the deaths of Frost and Kasza. The suit sought compensation for the injuries they had sustained, claiming the USAF had illegally handled toxic materials, and that the EPA had failed in its duty to enforce the Resource Conservation and Recovery Act (which governs handling of dangerous materials). They also sought detailed information about the chemicals to which they were allegedly exposed, hoping this would facilitate the medical treatment of survivors. Congressman Lee H. Hamilton, former chairman of the House Intelligence Committee, told 60 Minutes reporter Lesley Stahl, \"The Air Force is classifying all information about Area 51 in order to protect themselves from a lawsuit.\"',\n",
       "  'question': 'who represented robert frost and walter kasza in their suit?',\n",
       "  'answers': {'answer_start': [218],\n",
       "   'text': ['George Washington University law professor Jonathan Turley']}})"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset['train'][0], dataset['train'][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "0221343a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "0221343a",
    "outputId": "5ab7c522-23ea-467c-be72-fe031432fbdc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "978"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_passages = list(set(dataset['test']['context']))\n",
    "upload_texts_to_pinecone(unique_passages, batch_size=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "796a7c80-7149-430e-b22c-9926c0d1daee",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "796a7c80-7149-430e-b22c-9926c0d1daee",
    "outputId": "5fc883f2-230a-4fbf-d186-a60f5dd3573a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "978"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(unique_passages)  # only 978 unique passages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "33yf7QrWtwt-",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "33yf7QrWtwt-",
    "outputId": "9e8d30a5-b1b3-43f7-ad6d-bec8135fc5ee"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dimension': 3072,\n",
       " 'index_fullness': 0.0,\n",
       " 'namespaces': {'default': {'vector_count': 978}},\n",
       " 'total_vector_count': 978}"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.describe_index_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58231b08-36c3-4e70-8fd4-3b4ee9cd1555",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "b25a8a8f-e873-43e7-9f56-b57bdf328416",
   "metadata": {
    "id": "b25a8a8f-e873-43e7-9f56-b57bdf328416"
   },
   "outputs": [],
   "source": [
    "# Pre-trained cross encoder\n",
    "cross_encoder = CrossEncoder('cross-encoder/ms-marco-MiniLM-L-12-v2', num_labels=1)\n",
    "\n",
    "q_to_hash = {data['question']: my_hash(data['context']) for data in dataset['test']}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "1JmjvJ5YutEx",
   "metadata": {
    "id": "1JmjvJ5YutEx"
   },
   "outputs": [],
   "source": [
    "import logging\n",
    "\n",
    "# Set the logging level to ERROR to suppress DEBUG and INFO logs\n",
    "logging.getLogger().setLevel(logging.ERROR)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "eLYwOGbZvprX",
   "metadata": {
    "id": "eLYwOGbZvprX"
   },
   "outputs": [],
   "source": [
    "from copy import copy\n",
    "\n",
    "def get_results_from_pinecone(query, top_k=3, re_rank_model=None, verbose=True, correct_hash=None):\n",
    "    results_from_pinecone = query_from_pinecone(query, top_k=top_k)\n",
    "\n",
    "    if verbose:\n",
    "        print(\"Query:\", query)\n",
    "\n",
    "    final_results = []\n",
    "\n",
    "    retrieved_correct_position, reranked_correct_position = None, None\n",
    "    for idx, result_from_pinecone in enumerate(results_from_pinecone):\n",
    "        if correct_hash and result_from_pinecone['id'] == correct_hash:\n",
    "            retrieved_correct_position = idx\n",
    "\n",
    "    if re_rank_model is not None:\n",
    "        if verbose:\n",
    "            print('Document ID (Hash)\\t\\tRetrieval Score\\tCE Score\\tText')\n",
    "\n",
    "        sentence_combinations = [[query, result_from_pinecone['metadata']['text']] for result_from_pinecone in results_from_pinecone]\n",
    "\n",
    "        # Compute the similarity scores for these combinations\n",
    "        similarity_scores = re_rank_model.predict(sentence_combinations, activation_fct=nn.Sigmoid())\n",
    "\n",
    "        # Sort the scores in decreasing order\n",
    "        sim_scores_argsort = list(reversed(np.argsort(similarity_scores)))\n",
    "        for rerank_idx, idx in enumerate(sim_scores_argsort):\n",
    "            result_from_pinecone = results_from_pinecone[idx]\n",
    "            if correct_hash and result_from_pinecone['id'] == correct_hash:\n",
    "                reranked_correct_position = rerank_idx\n",
    "            final_results.append({'score': similarity_scores[idx], 'id': result_from_pinecone['id'], 'metadata': result_from_pinecone['metadata']})\n",
    "            if verbose:\n",
    "                print(f\"{result_from_pinecone['id']}\\t{result_from_pinecone['score']:.2f}\\t{similarity_scores[idx]:.6f}\\t{result_from_pinecone['metadata']['text'][:50]}\")\n",
    "        return {'final_results': final_results, 'retrieved_correct_position': retrieved_correct_position, 'reranked_correct_position': reranked_correct_position, 'results_from_pinecone': results_from_pinecone}\n",
    "\n",
    "    if verbose:\n",
    "        print('Document ID (Hash)\\t\\tRetrieval Score\\tText')\n",
    "    for result_from_pinecone in results_from_pinecone:\n",
    "        final_results.append(result_from_pinecone)\n",
    "        if verbose:\n",
    "            print(f\"{result_from_pinecone['id']}\\t{result_from_pinecone['score']:.2f}\\t{result_from_pinecone['metadata']['text'][:50]}\")\n",
    "\n",
    "    return {'final_results': final_results, 'retrieved_correct_position': retrieved_correct_position, 'reranked_correct_position': reranked_correct_position}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "vptFN0wPw0uS",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "vptFN0wPw0uS",
    "outputId": "72a9e946-3b3a-429b-b476-a945e94e26ec"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1148"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_inputs = list(set(dataset['test']['question']))\n",
    "len(unique_inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "caa66ac8-d69c-49a1-9f55-e9ef93379fec",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "caa66ac8-d69c-49a1-9f55-e9ef93379fec",
    "outputId": "575fa7f5-32dc-4a03-8f31-01df1ee11cc4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In what year did Scipione Riva-Rocci invent the sphygmomanometer\n",
      "raw retrieved index of correct result 0\n",
      "reranked index of correct result 0\n",
      "---\n",
      "When did Muntjacs begin appearing?\n",
      "raw retrieved index of correct result 0\n",
      "reranked index of correct result 0\n",
      "---\n",
      "Who aided Germany?\n",
      "raw retrieved index of correct result 2\n",
      "reranked index of correct result 0\n",
      "---\n",
      "What can negatively affect medical devices?\n",
      "raw retrieved index of correct result 0\n",
      "reranked index of correct result 0\n",
      "---\n",
      "what can websites mine?\n",
      "raw retrieved index of correct result 0\n",
      "reranked index of correct result 0\n",
      "---\n"
     ]
    }
   ],
   "source": [
    "for query in unique_inputs[:5]:\n",
    "    \n",
    "    r = get_results_from_pinecone(\n",
    "        query, top_k=10, re_rank_model=cross_encoder, correct_hash=q_to_hash[query],\n",
    "        verbose=False\n",
    "        )\n",
    "    \n",
    "    print(query)\n",
    "    print('raw retrieved index of correct result', r['retrieved_correct_position'])\n",
    "    print('reranked index of correct result',  r['reranked_correct_position'])\n",
    "    print('---')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "a63ec53e",
   "metadata": {
    "id": "a63ec53e"
   },
   "outputs": [],
   "source": [
    "# Let's test the performance re-ranking against validation datapoints\n",
    "# Note we could not use Pinecone here to speed things up\n",
    "#  but it's also a good time to see the latency of the pipeline with Pinecone\n",
    "test_sample = dataset['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "21a3a147",
   "metadata": {
    "id": "21a3a147"
   },
   "outputs": [],
   "source": [
    "TOP_K=50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b98413a3",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "b98413a3",
    "outputId": "c6557744-6998-492f-db27-98b94ed7ec31",
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "logger.setLevel(logging.CRITICAL)\n",
    "\n",
    "predictions = []\n",
    "\n",
    "for question in tqdm(test_sample['question']):\n",
    "    r = get_results_from_pinecone(\n",
    "        question, top_k=TOP_K, re_rank_model=cross_encoder, correct_hash=q_to_hash[question],\n",
    "        verbose=False\n",
    "        )\n",
    "\n",
    "    r['retrieved_correct_position'], r['reranked_correct_position']\n",
    "    predictions.append(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2ff94e7-0b78-40ca-8cf0-42e345439e8b",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "a2ff94e7-0b78-40ca-8cf0-42e345439e8b",
    "outputId": "a2cd7fd3-ae14-43dd-cab8-559f7878fd6f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7543554006968641\n",
      "Accuracy with re-ranking: 0.8327526132404182\n"
     ]
    }
   ],
   "source": [
    "retrieved_accuracy = sum([_['retrieved_correct_position'] == 0 for _ in predictions])/len(predictions)\n",
    "re_ranked_accuracy = sum([_['reranked_correct_position'] == 0 for _ in predictions])/len(predictions)\n",
    "\n",
    "print(f'Accuracy without re-ranking: {retrieved_accuracy}')\n",
    "print(f'Accuracy with re-ranking: {re_ranked_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fT4FuVcf9ONw",
   "metadata": {
    "id": "fT4FuVcf9ONw"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6649afac-9434-440b-b332-de9149a7ebfa",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "6649afac-9434-440b-b332-de9149a7ebfa",
    "outputId": "17b2e526-cd7e-409e-e0f3-2ae4afd29160"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-65b2749d-6c16-4a26-8a9c-7659badc9be6\" class=\"colab-df-container\">\n",
       "    <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>final_results</th>\n",
       "      <th>retrieved_correct_position</th>\n",
       "      <th>reranked_correct_position</th>\n",
       "      <th>results_from_pinecone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '2f90090e21f19450887d5f3ff781e541',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '49201636ad4102735125e146c0dbafa4',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '80d92494d2b06f341842f1855d2938cf',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>[{'id': 'e3fd54f33b021ea3cf88b438fefcada7',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '37ed40c20d4e1b9cc8d6cc27a4d0fff3',\n",
       " '...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "    <div class=\"colab-df-buttons\">\n",
       "\n",
       "  <div class=\"colab-df-container\">\n",
       "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-65b2749d-6c16-4a26-8a9c-7659badc9be6')\"\n",
       "            title=\"Convert this dataframe to an interactive table.\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
       "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
       "  </svg>\n",
       "    </button>\n",
       "\n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    .colab-df-buttons div {\n",
       "      margin-bottom: 4px;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "    <script>\n",
       "      const buttonEl =\n",
       "        document.querySelector('#df-65b2749d-6c16-4a26-8a9c-7659badc9be6 button.colab-df-convert');\n",
       "      buttonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "      async function convertToInteractive(key) {\n",
       "        const element = document.querySelector('#df-65b2749d-6c16-4a26-8a9c-7659badc9be6');\n",
       "        const dataTable =\n",
       "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                    [key], {});\n",
       "        if (!dataTable) return;\n",
       "\n",
       "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "          + ' to learn more about interactive tables.';\n",
       "        element.innerHTML = '';\n",
       "        dataTable['output_type'] = 'display_data';\n",
       "        await google.colab.output.renderOutput(dataTable, element);\n",
       "        const docLink = document.createElement('div');\n",
       "        docLink.innerHTML = docLinkHtml;\n",
       "        element.appendChild(docLink);\n",
       "      }\n",
       "    </script>\n",
       "  </div>\n",
       "\n",
       "\n",
       "<div id=\"df-b850ff8b-8fcc-433e-b65c-488a7c0db270\">\n",
       "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-b850ff8b-8fcc-433e-b65c-488a7c0db270')\"\n",
       "            title=\"Suggest charts\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "     width=\"24px\">\n",
       "    <g>\n",
       "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
       "    </g>\n",
       "</svg>\n",
       "  </button>\n",
       "\n",
       "<style>\n",
       "  .colab-df-quickchart {\n",
       "      --bg-color: #E8F0FE;\n",
       "      --fill-color: #1967D2;\n",
       "      --hover-bg-color: #E2EBFA;\n",
       "      --hover-fill-color: #174EA6;\n",
       "      --disabled-fill-color: #AAA;\n",
       "      --disabled-bg-color: #DDD;\n",
       "  }\n",
       "\n",
       "  [theme=dark] .colab-df-quickchart {\n",
       "      --bg-color: #3B4455;\n",
       "      --fill-color: #D2E3FC;\n",
       "      --hover-bg-color: #434B5C;\n",
       "      --hover-fill-color: #FFFFFF;\n",
       "      --disabled-bg-color: #3B4455;\n",
       "      --disabled-fill-color: #666;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart {\n",
       "    background-color: var(--bg-color);\n",
       "    border: none;\n",
       "    border-radius: 50%;\n",
       "    cursor: pointer;\n",
       "    display: none;\n",
       "    fill: var(--fill-color);\n",
       "    height: 32px;\n",
       "    padding: 0;\n",
       "    width: 32px;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart:hover {\n",
       "    background-color: var(--hover-bg-color);\n",
       "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "    fill: var(--button-hover-fill-color);\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart-complete:disabled,\n",
       "  .colab-df-quickchart-complete:disabled:hover {\n",
       "    background-color: var(--disabled-bg-color);\n",
       "    fill: var(--disabled-fill-color);\n",
       "    box-shadow: none;\n",
       "  }\n",
       "\n",
       "  .colab-df-spinner {\n",
       "    border: 2px solid var(--fill-color);\n",
       "    border-color: transparent;\n",
       "    border-bottom-color: var(--fill-color);\n",
       "    animation:\n",
       "      spin 1s steps(1) infinite;\n",
       "  }\n",
       "\n",
       "  @keyframes spin {\n",
       "    0% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "      border-left-color: var(--fill-color);\n",
       "    }\n",
       "    20% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    30% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    40% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    60% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    80% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "    90% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "  }\n",
       "</style>\n",
       "\n",
       "  <script>\n",
       "    async function quickchart(key) {\n",
       "      const quickchartButtonEl =\n",
       "        document.querySelector('#' + key + ' button');\n",
       "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
       "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
       "      try {\n",
       "        const charts = await google.colab.kernel.invokeFunction(\n",
       "            'suggestCharts', [key], {});\n",
       "      } catch (error) {\n",
       "        console.error('Error during call to suggestCharts:', error);\n",
       "      }\n",
       "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
       "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
       "    }\n",
       "    (() => {\n",
       "      let quickchartButtonEl =\n",
       "        document.querySelector('#df-b850ff8b-8fcc-433e-b65c-488a7c0db270 button');\n",
       "      quickchartButtonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "    })();\n",
       "  </script>\n",
       "</div>\n",
       "\n",
       "    </div>\n",
       "  </div>\n"
      ],
      "text/plain": [
       "                                       final_results  \\\n",
       "0  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "1  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "2  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "3  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "4  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "\n",
       "   retrieved_correct_position  reranked_correct_position  \\\n",
       "0                         0.0                        0.0   \n",
       "1                         0.0                        0.0   \n",
       "2                         0.0                        0.0   \n",
       "3                         0.0                        2.0   \n",
       "4                         0.0                        0.0   \n",
       "\n",
       "                               results_from_pinecone  \n",
       "0  [{'id': '2f90090e21f19450887d5f3ff781e541',\n",
       " '...  \n",
       "1  [{'id': '49201636ad4102735125e146c0dbafa4',\n",
       " '...  \n",
       "2  [{'id': '80d92494d2b06f341842f1855d2938cf',\n",
       " '...  \n",
       "3  [{'id': 'e3fd54f33b021ea3cf88b438fefcada7',\n",
       " '...  \n",
       "4  [{'id': '37ed40c20d4e1b9cc8d6cc27a4d0fff3',\n",
       " '...  "
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "predictions_df = pd.DataFrame(predictions)\n",
    "predictions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "831397d3-606a-4c7f-b77f-6ee85bd24f7e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "831397d3-606a-4c7f-b77f-6ee85bd24f7e",
    "outputId": "041c1a5c-cdef-4668-bb59-2700997c67ab"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "retrieved_correct_position    1.311052\n",
       "reranked_correct_position     0.682709\n",
       "dtype: float64"
      ]
     },
     "execution_count": 288,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions_df[['retrieved_correct_position', 'reranked_correct_position']].mean()  # lower is better"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eyv-6pPIEy0r",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "eyv-6pPIEy0r",
    "outputId": "4d82a213-bd82-44e4-d933-8d39870c2f4d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 866 956\n",
      "3 1009 1057\n",
      "5 1048 1077\n",
      "10 1072 1094\n",
      "25 1108 1120\n",
      "50 1122 1122\n"
     ]
    }
   ],
   "source": [
    "# do recall @ 1, 3, 5, 10\n",
    "for k in (1, 3, 5, 10, 25, 50):\n",
    "    embedding_only_recall = predictions_df[predictions_df['retrieved_correct_position'] < k].shape[0]\n",
    "    reranked_recall = predictions_df[predictions_df['reranked_correct_position'] < k].shape[0]\n",
    "    print(k, embedding_only_recall, reranked_recall)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "vtMoDh0-TDpr",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "vtMoDh0-TDpr",
    "outputId": "8581d271-8699-4925-9b57-341a09a64c1a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1148, 4)"
      ]
     },
     "execution_count": 291,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea30153c",
   "metadata": {
    "id": "ea30153c"
   },
   "source": [
    "# Fine-tuning our cross-encoder re-ranker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "508edaf6",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "508edaf6",
    "outputId": "1938803f-c8bd-478b-a37c-3a3894e1c106"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'id': 'a4968ca8a18de16aa3859be760e43dbd3af3fce9',\n",
       " 'title': 'Area 51',\n",
       " 'context': 'In 1994, five unnamed civilian contractors and the widows of contractors Walter Kasza and Robert Frost sued the USAF and the United States Environmental Protection Agency. Their suit, in which they were represented by George Washington University law professor Jonathan Turley, alleged they had been present when large quantities of unknown chemicals had been burned in open pits and trenches at Groom. Biopsies taken from the complainants were analyzed by Rutgers University biochemists, who found high levels of dioxin, dibenzofuran, and trichloroethylene in their body fat. The complainants alleged they had sustained skin, liver, and respiratory injuries due to their work at Groom, and that this had contributed to the deaths of Frost and Kasza. The suit sought compensation for the injuries they had sustained, claiming the USAF had illegally handled toxic materials, and that the EPA had failed in its duty to enforce the Resource Conservation and Recovery Act (which governs handling of dangerous materials). They also sought detailed information about the chemicals to which they were allegedly exposed, hoping this would facilitate the medical treatment of survivors. Congressman Lee H. Hamilton, former chairman of the House Intelligence Committee, told 60 Minutes reporter Lesley Stahl, \"The Air Force is classifying all information about Area 51 in order to protect themselves from a lawsuit.\"',\n",
       " 'question': 'Who analyzed the biopsies?',\n",
       " 'answers': {'answer_start': [457],\n",
       "  'text': ['Rutgers University biochemists']}}"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset['train'][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "51bf1607",
   "metadata": {
    "id": "51bf1607"
   },
   "outputs": [],
   "source": [
    "from sentence_transformers import InputExample, losses, evaluation\n",
    "from torch.utils.data import DataLoader\n",
    "from random import shuffle\n",
    "\n",
    "shuffled_training_passages = dataset['train']['question'].copy()\n",
    "shuffle(shuffled_training_passages)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6b58d2d0",
   "metadata": {
    "id": "6b58d2d0"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50fd5c19-894a-41e4-bf6f-d9e1197c053f",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "50fd5c19-894a-41e4-bf6f-d9e1197c053f",
    "outputId": "8501a6c6-548c-4928-ecb7-4ed99339063b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9916, 11590)"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_train_passages = list(set(dataset['train']['context']))\n",
    "len(unique_train_passages), len(dataset['train']['context'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce8e076e-4822-4686-b57a-b8c0e09583e5",
   "metadata": {
    "id": "ce8e076e-4822-4686-b57a-b8c0e09583e5"
   },
   "outputs": [],
   "source": [
    "from sentence_transformers import SentenceTransformer\n",
    "\n",
    "bi_encoder = SentenceTransformer(\"sentence-transformers/msmarco-MiniLM-L-6-v3\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c4ae6f1f-0523-40ff-8d5d-f374e4ead044",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 49,
     "referenced_widgets": [
      "dbca2d7da970438892d45cd387434f6f",
      "6f5389fdc3674a05ad79bb9456a5995f",
      "7a5c511f2fe84c66beef7fed87023669",
      "70e45cb4f2f64c499acc785ec64743fe",
      "bb260568c96a4397bba2b1058404eec2",
      "e281ca37b72d422285aa64193265aaa4",
      "687d7bf7373f4541852203637e18d2e3",
      "5a7b319329164b27bf452869e2335aef",
      "e5a88639e7e5430fbe44f73d2e5a4b23",
      "d9f6781147a24bc3aec97b87927bdfbb",
      "d0c85ebbf1a54fdcaa33210aa3e247b8"
     ]
    },
    "id": "c4ae6f1f-0523-40ff-8d5d-f374e4ead044",
    "outputId": "847007cf-bf2e-4d3a-a10f-fd84123c1eb0"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dbca2d7da970438892d45cd387434f6f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Batches:   0%|          | 0/620 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "embeddings = bi_encoder.encode(\n",
    "    unique_train_passages,\n",
    "    show_progress_bar=True,\n",
    "    batch_size=16,\n",
    "    normalize_embeddings=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5eaad504-b944-4e36-987f-34126189c1a9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "5eaad504-b944-4e36-987f-34126189c1a9",
    "outputId": "ac4a626c-862b-4676-f177-4b3c477d93c1"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9916, (9916, 384))"
      ]
     },
     "execution_count": 199,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(unique_train_passages), embeddings.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "361d5ef0-3fe3-4ed8-87c0-1cea3d9abcc2",
   "metadata": {
    "id": "361d5ef0-3fe3-4ed8-87c0-1cea3d9abcc2",
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# use sentence_transformers.util.semantic_search\n",
    "from sentence_transformers.util import semantic_search\n",
    "\n",
    "unique_train_passages = np.array(unique_train_passages)\n",
    "\n",
    "# Function to find most similar document\n",
    "def find_most_similar(text, embeddings, documents, k=3):\n",
    "    query_embedding = bi_encoder.encode([text], show_progress_bar=False)\n",
    "    similarities = semantic_search(query_embedding, embeddings, top_k=k)\n",
    "    return [(documents[sim['corpus_id']], sim['score'], sim['corpus_id']) for sim in similarities[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c7d79fe9-f053-40e7-a2e6-2efd715b43f7",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "c7d79fe9-f053-40e7-a2e6-2efd715b43f7",
    "outputId": "5112eaae-cf89-4779-a87c-dcd03750798c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The ancient Egyptians engaged in trade with their foreign neighbors to obtain rare, exotic goods not found in Egypt. In the Predynastic Period, they established trade with Nubia to obtain gold and incense. They also established trade with Palestine, as evidenced by Palestinian-style oil jugs found in the burials of the First Dynasty pharaohs. An Egyptian colony stationed in southern Canaan dates to slightly before the First Dynasty. Narmer had Egyptian pottery produced in Canaan and exported back to Egypt.By the Second Dynasty at latest, ancient Egyptian trade with Byblos yielded a critical source of quality timber not found in Egypt. By the Fifth Dynasty, trade with Punt provided gold, aromatic resins, ebony, ivory, and wild animals such as monkeys and baboons. Egypt relied on trade with Anatolia for essential quantities of tin as well as supplementary supplies of copper, both metals being necessary for the manufacture of bronze. The ancient Egyptians prized the blue stone lapis lazuli, which had to be imported from far-away Afghanistan. Egypt's Mediterranean trade partners also included Greece and Crete, which provided, among other goods, supplies of olive oil. In exchange for its luxury imports and raw materials, Egypt mainly exported grain, gold, linen, and papyrus, in addition to other finished goods including glass and stone objects.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(\"The ancient Egyptians engaged in trade with their foreign neighbors to obtain rare, exotic goods not found in Egypt. In the Predynastic Period, they established trade with Nubia to obtain gold and incense. They also established trade with Palestine, as evidenced by Palestinian-style oil jugs found in the burials of the First Dynasty pharaohs. An Egyptian colony stationed in southern Canaan dates to slightly before the First Dynasty. Narmer had Egyptian pottery produced in Canaan and exported back to Egypt.By the Second Dynasty at latest, ancient Egyptian trade with Byblos yielded a critical source of quality timber not found in Egypt. By the Fifth Dynasty, trade with Punt provided gold, aromatic resins, ebony, ivory, and wild animals such as monkeys and baboons. Egypt relied on trade with Anatolia for essential quantities of tin as well as supplementary supplies of copper, both metals being necessary for the manufacture of bronze. The ancient Egyptians prized the blue stone lapis lazuli, which had to be imported from far-away Afghanistan. Egypt's Mediterranean trade partners also included Greece and Crete, which provided, among other goods, supplies of olive oil. In exchange for its luxury imports and raw materials, Egypt mainly exported grain, gold, linen, and papyrus, in addition to other finished goods including glass and stone objects.\",\n",
       "  1.0,\n",
       "  0),\n",
       " ('Egypt is rich in building and decorative stone, copper and lead ores, gold, and semiprecious stones. These natural resources allowed the ancient Egyptians to build monuments, sculpt statues, make tools, and fashion jewelry. Embalmers used salts from the Wadi Natrun for mummification, which also provided the gypsum needed to make plaster. Ore-bearing rock formations were found in distant, inhospitable wadis in the eastern desert and the Sinai, requiring large, state-controlled expeditions to obtain natural resources found there. There were extensive gold mines in Nubia, and one of the first maps known is of a gold mine in this region. The Wadi Hammamat was a notable source of granite, greywacke, and gold. Flint was the first mineral collected and used to make tools, and flint handaxes are the earliest pieces of evidence of habitation in the Nile valley. Nodules of the mineral were carefully flaked to make blades and arrowheads of moderate hardness and durability even after copper was adopted for this purpose. Ancient Egyptians were among the first to use minerals such as sulfur as cosmetic substances.The Egyptians worked deposits of the lead ore galena at Gebel Rosas to make net sinkers, plumb bobs, and small figurines. Copper was the most important metal for toolmaking in ancient Egypt and was smelted in furnaces from malachite ore mined in the Sinai. Workers collected gold by washing the nuggets out of sediment in alluvial deposits, or by the more labor-intensive process of grinding and washing gold-bearing quartzite. Iron deposits found in upper Egypt were utilized in the Late Period. High-quality building stones were abundant in Egypt; the ancient Egyptians quarried limestone all along the Nile valley, granite from Aswan, and basalt and sandstone from the wadis of the eastern desert. Deposits of decorative stones such as porphyry, greywacke, alabaster, and carnelian dotted the eastern desert and were collected even before the First Dynasty. In the Ptolemaic and Roman Periods, miners worked deposits of emeralds in Wadi Sikait and amethyst in Wadi el-Hudi.',\n",
       "  0.7423145174980164,\n",
       "  1393),\n",
       " ('The ancient Egyptians viewed men and women, including people from all social classes except slaves, as essentially equal under the law, and even the lowliest peasant was entitled to petition the vizier and his court for redress. Although slaves were mostly used as indentured servants, they were able to buy and sell their servitude, work their way to freedom or nobility, and were usually treated by doctors in the workplace. Both men and women had the right to own and sell property, make contracts, marry and divorce, receive inheritance, and pursue legal disputes in court. Married couples could own property jointly and protect themselves from divorce by agreeing to marriage contracts, which stipulated the financial obligations of the husband to his wife and children should the marriage end. Compared with their counterparts in ancient Greece, Rome, and even more modern places around the world, ancient Egyptian women had a greater range of personal choices and opportunities for achievement. Women such as Hatshepsut and Cleopatra VII even became pharaohs, while others wielded power as Divine Wives of Amun. Despite these freedoms, ancient Egyptian women did not often take part in official roles in the administration, served only secondary roles in the temples, and were not as likely to be as educated as men.',\n",
       "  0.6631918549537659,\n",
       "  1693)]"
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Example usage\n",
    "input_text = unique_train_passages[0]\n",
    "print(input_text)\n",
    "\n",
    "find_most_similar(input_text, embeddings, unique_train_passages)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "262e4ca4-0bfe-4243-9534-49370ed66b40",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "262e4ca4-0bfe-4243-9534-49370ed66b40",
    "outputId": "79506047-cc55-4163-e4c4-85274f0622f0"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11590/11590 [01:54<00:00, 100.86it/s]\n"
     ]
    }
   ],
   "source": [
    "# add some negative examples\n",
    "train_samples = []\n",
    "\n",
    "for train_example in tqdm(dataset['train']):\n",
    "    for i, (passage, score, corpus_idx) in enumerate(find_most_similar(train_example['question'], embeddings, unique_train_passages)):\n",
    "        train_samples.append(\n",
    "            InputExample(\n",
    "                texts=[train_example['question'], passage], label=1 if i == 0 else 0\n",
    "            )\n",
    "        )\n",
    "\n",
    "shuffle(train_samples)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b281241b-54ed-4405-a101-f774efd2317d",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "b281241b-54ed-4405-a101-f774efd2317d",
    "outputId": "649df482-1761-4ba9-c20d-4364e037e2b0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34770"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e7eaa40",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8e7eaa40",
    "outputId": "2d1ab8ba-9f2e-4b24-b104-baf10d58e45f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'guid': '',\n",
       " 'texts': ['How was Mike Leander also known as?',\n",
       "  \"Günter Theodor Netzer (born 14 September 1944 in Mönchengladbach) is a former German football player and team general manager currently working in the media business. He achieved great success in Germany with Borussia Mönchengladbach in the early 1970s, and, after moving to Spain in 1973, with Real Madrid. He played as an attacking midfielder, and is considered one of the greatest passers in the game's history. He was voted Footballer of the Year in Germany twice, in 1972 and 1973. He was the general manager for Hamburger SV during much of the team's golden period from the late 1970s through the early 1980s, when the club won three league titles and the 1983 European Cup.\"],\n",
       " 'label': 0}"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samples[0].__dict__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5fb16949-464e-4885-bf2a-f5e1cc7fe5b1",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "5fb16949-464e-4885-bf2a-f5e1cc7fe5b1",
    "outputId": "a26b3848-0cb8-4bd0-b155-4a8df4b24f19"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'guid': '',\n",
       " 'texts': ['When did US troops finally leave Uzbekistan?',\n",
       "  \"In 1918, Japan continued to extend its influence and privileges in China via the Nishihara Loans. Following the Bolshevik Revolution in Russia, Japan and the United States sent forces to Siberia in 1918 to bolster the armies of the White movement leader Admiral Alexander Kolchak against the Bolshevik Red Army. In this Siberian Intervention, the Imperial Japanese Army initially planned to send more than 70,000 troops to occupy Siberia as far west as Lake Baikal. The plan was scaled back considerably due to opposition from the United States.Toward the end of the war, Japan increasingly filled orders for needed war material for its European allies. The wartime boom helped to diversify the country's industry, increase its exports, and transform Japan from a debtor to a creditor nation for the first time. Exports quadrupled from 1913 to 1918. The massive capital influx into Japan and the subsequent industrial boom led to rapid inflation. In August 1918, rice riots caused by this inflation erupted in towns and cities throughout Japan.\"],\n",
       " 'label': 0}"
      ]
     },
     "execution_count": 205,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samples[1].__dict__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc2bc9d1",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "bc2bc9d1",
    "outputId": "97ead4f4-bce2-4e8b-d379-1baab6ddd641"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warmup-steps: 174\n"
     ]
    }
   ],
   "source": [
    "from sentence_transformers.cross_encoder.evaluation import CECorrelationEvaluator, CEBinaryClassificationEvaluator\n",
    "import math\n",
    "import torch\n",
    "from random import sample\n",
    "\n",
    "logger.setLevel(logging.DEBUG)  # just to get some logs\n",
    "\n",
    "num_epochs = 1\n",
    "\n",
    "model_save_path = './fine_tuned_ir_cross_encoder'\n",
    "\n",
    "train_dataloader = DataLoader(train_samples[:int(len(train_samples)*.8)], shuffle=True, batch_size=16)\n",
    "\n",
    "# An evaluator for training performance\n",
    "evaluator = CECorrelationEvaluator.from_input_examples(train_samples[int(len(train_samples)*.8):], name='test')\n",
    "\n",
    "# Rule of thumb for warmup steps\n",
    "warmup_steps = math.ceil(len(train_dataloader) * num_epochs * 0.1)  # 10% of train data for warm-up\n",
    "print(f\"Warmup-steps: {warmup_steps}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce4e8079-1335-4d53-944d-3ed0f6537781",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 137,
     "referenced_widgets": [
      "f36d8f08f6284b848622fd560196cc78",
      "608d1345487046ceb64b0b31a4287ae0",
      "43c8db4be4984f6da1540de6bc89b616",
      "6cb81ad9ee8141aa9d0fbefc276f244d",
      "8e989ddf9d274876a3ba06056e7147a1",
      "fce9aa1662c34941afadd6141aba9180",
      "098016d32abc4f45be0a4f747cbc3f11",
      "dc26c84713fb45f5a6c5b870c96d9d9e",
      "a6ac9d7aed4c493db567b015ab8502a5",
      "80e18075a957476591fa1ba2585b7a73",
      "e1d8ec56e5614d77a6999a42283be94b",
      "21316ef8a22d40ab9109ccf0fa7fa645",
      "458bcab796d642c9afd2e534085b991f",
      "266b42f10bd74b9f907588c1251c8d6c",
      "ba968ae891a2400fa0db9f331b85b66c",
      "5d09ef7e47f446a3a3a51458195a0fe3",
      "352af51e6e7943a7a41d3bbe41885d8d",
      "2cb4c1bff01b4f669ecd8583187bab9f",
      "b070922627a04195a41899a169984b08",
      "85a67e7f5d2845329d63f0922bb5d9a6",
      "5b3c84192618410ea1793239c4514c6e",
      "7bd0527b8e174aec9911c5b32dc048fd"
     ]
    },
    "id": "ce4e8079-1335-4d53-944d-3ed0f6537781",
    "outputId": "446b8d35-7bc9-48d5-c9ab-60dc6b3762fc"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f36d8f08f6284b848622fd560196cc78",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Batches:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'guid': '', 'texts': ['What can be inferred about Zionism?', 'Determines that Zionism is a form of racism and racial discrimination.'], 'label': 1} 0.9391304\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "21316ef8a22d40ab9109ccf0fa7fa645",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Batches:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'guid': '', 'texts': ['How was Mike Leander also known as?', \"Günter Theodor Netzer (born 14 September 1944 in Mönchengladbach) is a former German football player and team general manager currently working in the media business. He achieved great success in Germany with Borussia Mönchengladbach in the early 1970s, and, after moving to Spain in 1973, with Real Madrid. He played as an attacking midfielder, and is considered one of the greatest passers in the game's history. He was voted Footballer of the Year in Germany twice, in 1972 and 1973. He was the general manager for Hamburger SV during much of the team's golden period from the late 1970s through the early 1980s, when the club won three league titles and the 1983 European Cup.\"], 'label': 0} 0.00016999716\n"
     ]
    }
   ],
   "source": [
    "for t in train_samples:\n",
    "    if t.label == 1:\n",
    "        print(t.__dict__, cross_encoder.predict(t.texts, activation_fct=nn.Sigmoid()))\n",
    "        break\n",
    "for t in train_samples:\n",
    "    if t.label == 0:\n",
    "        print(t.__dict__, cross_encoder.predict(t.texts, activation_fct=nn.Sigmoid()))\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f981bd03-bb31-492d-aabf-d5e6c7647704",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f981bd03-bb31-492d-aabf-d5e6c7647704",
    "outputId": "f918fbbb-80b1-44b5-eee7-219d4e632dc7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3844874561927453"
      ]
     },
     "execution_count": 251,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluator(cross_encoder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0de4dbc8",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 81,
     "referenced_widgets": [
      "e31ae00bfabe4ed5b058d7dbe55a80f6",
      "b9fbabd33a2a4d0a94055cd086d83228",
      "b281e2a5a8204b52827f01316d637952",
      "cdd38d9a948a4a4c8408dcb06d7c584a",
      "b86aa89087cb425198afa2db30d86600",
      "168948926c2340abb8b52f8dd67a5e73",
      "8396e1788d5540a79a9d195e2753351a",
      "084ebc36eb9d4616acc1a0843673ac4f",
      "252dfe4d46604a089c263bb0eae17a3a",
      "2aa32422748943759daf5ef05077bda0",
      "e7c672eb19574b32811bdcc3f6ac0b34",
      "37395cf0cac548c783ccbdac230c4a3c",
      "8085fe243d814d5da2b4d6b665a6aa51",
      "0295736b3f254a8590d911792f2c19db",
      "ad079ced956841be88a3ab6f2fdcd88d",
      "953b16307b514bc995efc40ff8f4698b",
      "4d26773cf0ec45fc9ac3185f8141cfe6",
      "d7959d1933e5466095bed23bb268afd5",
      "3f1ba16f17c94fb8a409d152b9a6045a",
      "c485139a148e4207957e96b4275eeed2",
      "6f96c8a415744597bcaa83141b3a457c",
      "8e41a0a7932b4f0b8a448847ff3f14ff"
     ]
    },
    "id": "0de4dbc8",
    "outputId": "1190bca8-f5ae-4907-cd50-1d6771ef1e22"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e31ae00bfabe4ed5b058d7dbe55a80f6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Epoch:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "37395cf0cac548c783ccbdac230c4a3c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Iteration:   0%|          | 0/1739 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train the model\n",
    "from sentence_transformers import InputExample, losses, evaluation\n",
    "\n",
    "# you may turn on debug for more logs here e.g. logger.setLevel(logging.DEBUG)\n",
    "cross_encoder.fit(\n",
    "    train_dataloader=train_dataloader,\n",
    "    loss_fct=nn.BCEWithLogitsLoss(),  # this is the default loss if num_labels is 1 otherwise CrossEntropyLoss\n",
    "    evaluator=evaluator,\n",
    "    epochs=num_epochs,\n",
    "    warmup_steps=warmup_steps,\n",
    "    output_path=model_save_path,\n",
    "    use_amp=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9_a5R9sVBvHN",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "9_a5R9sVBvHN",
    "outputId": "fdba13f3-8eaa-4bf8-8056-3168674bd0a8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.42047041059290036"
      ]
     },
     "execution_count": 260,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluator(cross_encoder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f747537",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2f747537",
    "outputId": "bb2caf80-5b00-4f57-aaf0-245fc586ef5f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9973792\n",
      "5.941634\n"
     ]
    }
   ],
   "source": [
    "finetuned = CrossEncoder(model_save_path)\n",
    "\n",
    "print(finetuned.predict(['hello', 'hi'], activation_fct=nn.Sigmoid()))\n",
    "print(finetuned.predict(['hello', 'hi'], activation_fct=nn.Identity()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "KGSvctdu4pRR",
   "metadata": {
    "id": "KGSvctdu4pRR"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "94cf736e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "94cf736e",
    "outputId": "28ee86d1-ada9-44d5-9fde-eaede674f0ac",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  9%|▊         | 100/1148 [02:38<27:50,  1.59s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.78\n",
      "Accuracy with re-ranking: 0.84\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 17%|█▋        | 200/1148 [05:18<26:10,  1.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.765\n",
      "Accuracy with re-ranking: 0.84\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 26%|██▌       | 300/1148 [07:57<22:00,  1.56s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7666666666666667\n",
      "Accuracy with re-ranking: 0.8266666666666667\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 35%|███▍      | 400/1148 [10:37<19:44,  1.58s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7625\n",
      "Accuracy with re-ranking: 0.84\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 44%|████▎     | 500/1148 [13:17<17:16,  1.60s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.764\n",
      "Accuracy with re-ranking: 0.85\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 52%|█████▏    | 600/1148 [16:00<14:35,  1.60s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7683333333333333\n",
      "Accuracy with re-ranking: 0.8566666666666667\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 61%|██████    | 700/1148 [18:40<12:28,  1.67s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7471428571428571\n",
      "Accuracy with re-ranking: 0.8428571428571429\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 70%|██████▉   | 800/1148 [21:21<09:13,  1.59s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.74875\n",
      "Accuracy with re-ranking: 0.835\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 78%|███████▊  | 900/1148 [25:36<06:43,  1.63s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7388888888888889\n",
      "Accuracy with re-ranking: 0.8333333333333334\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 87%|████████▋ | 1000/1148 [28:16<04:00,  1.62s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.741\n",
      "Accuracy with re-ranking: 0.835\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 96%|█████████▌| 1100/1148 [30:57<01:16,  1.60s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7509090909090909\n",
      "Accuracy with re-ranking: 0.8436363636363636\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1148/1148 [32:14<00:00,  1.68s/it]\n"
     ]
    }
   ],
   "source": [
    "logger.setLevel(logging.CRITICAL)\n",
    "\n",
    "ft_predictions = []\n",
    "\n",
    "for question in tqdm(test_sample['question']):\n",
    "    r = get_results_from_pinecone(\n",
    "        question, top_k=TOP_K, re_rank_model=finetuned, correct_hash=q_to_hash[question],\n",
    "        verbose=False\n",
    "        )\n",
    "\n",
    "    r['retrieved_correct_position'], r['reranked_correct_position']\n",
    "    ft_predictions.append(r)\n",
    "    if len(ft_predictions) % 100 == 0:\n",
    "        retrieved_accuracy = sum([_['retrieved_correct_position'] == 0 for _ in ft_predictions])/len(ft_predictions)\n",
    "        re_ranked_accuracy = sum([_['reranked_correct_position'] == 0 for _ in ft_predictions])/len(ft_predictions)\n",
    "\n",
    "        print(f'Accuracy without re-ranking: {retrieved_accuracy}')\n",
    "        print(f'Accuracy with re-ranking: {re_ranked_accuracy}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a07b782d-ef97-4c0a-be53-aefea5936676",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "a07b782d-ef97-4c0a-be53-aefea5936676",
    "outputId": "b4daa149-efbf-4383-c233-ccd7b2bcfc17"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.7543554006968641\n",
      "Accuracy with re-ranking: 0.8493031358885017\n"
     ]
    }
   ],
   "source": [
    "retrieved_accuracy = sum([_['retrieved_correct_position'] == 0 for _ in ft_predictions])/len(ft_predictions)\n",
    "re_ranked_accuracy = sum([_['reranked_correct_position'] == 0 for _ in ft_predictions])/len(ft_predictions)\n",
    "\n",
    "print(f'Accuracy without re-ranking: {retrieved_accuracy}')\n",
    "print(f'Accuracy with re-ranking: {re_ranked_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb46e242",
   "metadata": {
    "id": "bb46e242"
   },
   "outputs": [],
   "source": [
    "# Re-ranking got slightly better after 1 epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0cfcd92e-31ad-4bf7-9e30-b6d260a06871",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "0cfcd92e-31ad-4bf7-9e30-b6d260a06871",
    "outputId": "ea88ceee-5307-491e-afe8-a5a8a43dba97"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-04533850-9071-4c72-8201-4d2b69823f96\" class=\"colab-df-container\">\n",
       "    <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>final_results</th>\n",
       "      <th>retrieved_correct_position</th>\n",
       "      <th>reranked_correct_position</th>\n",
       "      <th>results_from_pinecone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '2f90090e21f19450887d5f3ff781e541',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '49201636ad4102735125e146c0dbafa4',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '80d92494d2b06f341842f1855d2938cf',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': 'e3fd54f33b021ea3cf88b438fefcada7',\n",
       " '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>{'final_results': [{'score': 0.8746502, 'id': ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[{'id': '37ed40c20d4e1b9cc8d6cc27a4d0fff3',\n",
       " '...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "    <div class=\"colab-df-buttons\">\n",
       "\n",
       "  <div class=\"colab-df-container\">\n",
       "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-04533850-9071-4c72-8201-4d2b69823f96')\"\n",
       "            title=\"Convert this dataframe to an interactive table.\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
       "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
       "  </svg>\n",
       "    </button>\n",
       "\n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    .colab-df-buttons div {\n",
       "      margin-bottom: 4px;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "    <script>\n",
       "      const buttonEl =\n",
       "        document.querySelector('#df-04533850-9071-4c72-8201-4d2b69823f96 button.colab-df-convert');\n",
       "      buttonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "      async function convertToInteractive(key) {\n",
       "        const element = document.querySelector('#df-04533850-9071-4c72-8201-4d2b69823f96');\n",
       "        const dataTable =\n",
       "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                    [key], {});\n",
       "        if (!dataTable) return;\n",
       "\n",
       "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "          + ' to learn more about interactive tables.';\n",
       "        element.innerHTML = '';\n",
       "        dataTable['output_type'] = 'display_data';\n",
       "        await google.colab.output.renderOutput(dataTable, element);\n",
       "        const docLink = document.createElement('div');\n",
       "        docLink.innerHTML = docLinkHtml;\n",
       "        element.appendChild(docLink);\n",
       "      }\n",
       "    </script>\n",
       "  </div>\n",
       "\n",
       "\n",
       "<div id=\"df-93c5c343-5687-4e9a-a551-b87fee369a61\">\n",
       "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-93c5c343-5687-4e9a-a551-b87fee369a61')\"\n",
       "            title=\"Suggest charts\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "     width=\"24px\">\n",
       "    <g>\n",
       "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
       "    </g>\n",
       "</svg>\n",
       "  </button>\n",
       "\n",
       "<style>\n",
       "  .colab-df-quickchart {\n",
       "      --bg-color: #E8F0FE;\n",
       "      --fill-color: #1967D2;\n",
       "      --hover-bg-color: #E2EBFA;\n",
       "      --hover-fill-color: #174EA6;\n",
       "      --disabled-fill-color: #AAA;\n",
       "      --disabled-bg-color: #DDD;\n",
       "  }\n",
       "\n",
       "  [theme=dark] .colab-df-quickchart {\n",
       "      --bg-color: #3B4455;\n",
       "      --fill-color: #D2E3FC;\n",
       "      --hover-bg-color: #434B5C;\n",
       "      --hover-fill-color: #FFFFFF;\n",
       "      --disabled-bg-color: #3B4455;\n",
       "      --disabled-fill-color: #666;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart {\n",
       "    background-color: var(--bg-color);\n",
       "    border: none;\n",
       "    border-radius: 50%;\n",
       "    cursor: pointer;\n",
       "    display: none;\n",
       "    fill: var(--fill-color);\n",
       "    height: 32px;\n",
       "    padding: 0;\n",
       "    width: 32px;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart:hover {\n",
       "    background-color: var(--hover-bg-color);\n",
       "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "    fill: var(--button-hover-fill-color);\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart-complete:disabled,\n",
       "  .colab-df-quickchart-complete:disabled:hover {\n",
       "    background-color: var(--disabled-bg-color);\n",
       "    fill: var(--disabled-fill-color);\n",
       "    box-shadow: none;\n",
       "  }\n",
       "\n",
       "  .colab-df-spinner {\n",
       "    border: 2px solid var(--fill-color);\n",
       "    border-color: transparent;\n",
       "    border-bottom-color: var(--fill-color);\n",
       "    animation:\n",
       "      spin 1s steps(1) infinite;\n",
       "  }\n",
       "\n",
       "  @keyframes spin {\n",
       "    0% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "      border-left-color: var(--fill-color);\n",
       "    }\n",
       "    20% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    30% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    40% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    60% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    80% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "    90% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "  }\n",
       "</style>\n",
       "\n",
       "  <script>\n",
       "    async function quickchart(key) {\n",
       "      const quickchartButtonEl =\n",
       "        document.querySelector('#' + key + ' button');\n",
       "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
       "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
       "      try {\n",
       "        const charts = await google.colab.kernel.invokeFunction(\n",
       "            'suggestCharts', [key], {});\n",
       "      } catch (error) {\n",
       "        console.error('Error during call to suggestCharts:', error);\n",
       "      }\n",
       "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
       "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
       "    }\n",
       "    (() => {\n",
       "      let quickchartButtonEl =\n",
       "        document.querySelector('#df-93c5c343-5687-4e9a-a551-b87fee369a61 button');\n",
       "      quickchartButtonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "    })();\n",
       "  </script>\n",
       "</div>\n",
       "\n",
       "    </div>\n",
       "  </div>\n"
      ],
      "text/plain": [
       "                                       final_results  \\\n",
       "0  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "1  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "2  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "3  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "4  {'final_results': [{'score': 0.8746502, 'id': ...   \n",
       "\n",
       "   retrieved_correct_position  reranked_correct_position  \\\n",
       "0                         0.0                        0.0   \n",
       "1                         0.0                        0.0   \n",
       "2                         0.0                        0.0   \n",
       "3                         0.0                        0.0   \n",
       "4                         0.0                        0.0   \n",
       "\n",
       "                               results_from_pinecone  \n",
       "0  [{'id': '2f90090e21f19450887d5f3ff781e541',\n",
       " '...  \n",
       "1  [{'id': '49201636ad4102735125e146c0dbafa4',\n",
       " '...  \n",
       "2  [{'id': '80d92494d2b06f341842f1855d2938cf',\n",
       " '...  \n",
       "3  [{'id': 'e3fd54f33b021ea3cf88b438fefcada7',\n",
       " '...  \n",
       "4  [{'id': '37ed40c20d4e1b9cc8d6cc27a4d0fff3',\n",
       " '...  "
      ]
     },
     "execution_count": 276,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ft_predictions_df = pd.DataFrame(ft_predictions)\n",
    "ft_predictions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9Sr2q5IN9LMz",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "9Sr2q5IN9LMz",
    "outputId": "4a670895-a2df-4300-9b8c-5d549c2e8517"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "retrieved_correct_position    26\n",
       "reranked_correct_position     26\n",
       "dtype: int64"
      ]
     },
     "execution_count": 277,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ft_predictions_df[['retrieved_correct_position', 'reranked_correct_position']].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "75b414c6",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "75b414c6",
    "outputId": "30220b68-5830-407f-e3ed-678ee4378a58"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "retrieved_correct_position    1.310160\n",
       "reranked_correct_position     0.644385\n",
       "dtype: float64"
      ]
     },
     "execution_count": 278,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ft_predictions_df[['retrieved_correct_position', 'reranked_correct_position']].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "BFuWRhWzUoe8",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 564
    },
    "id": "BFuWRhWzUoe8",
    "outputId": "8e6cbcbc-8f01-4f98-dd5a-969f32552e5a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2535cbae",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2535cbae",
    "outputId": "97f405f3-c446-47a2-e645-9440945df769"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 866 975\n",
      "3 1009 1056\n",
      "5 1048 1077\n",
      "10 1072 1101\n",
      "25 1108 1117\n",
      "50 1122 1122\n"
     ]
    }
   ],
   "source": [
    "# do recall @ 1, 3, 5, 10\n",
    "for k in (1, 3, 5, 10, 25, 50):\n",
    "    embedding_only_recall = ft_predictions_df[ft_predictions_df['retrieved_correct_position'] < k].shape[0]\n",
    "    reranked_recall = ft_predictions_df[ft_predictions_df['reranked_correct_position'] < k].shape[0]\n",
    "    print(k, embedding_only_recall, reranked_recall)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "o--xzdNXYxDG",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "o--xzdNXYxDG",
    "outputId": "4adfe46d-0bb3-4bff-f4a9-0536a467a233"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1148, 4)"
      ]
     },
     "execution_count": 290,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ft_predictions_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f14809e",
   "metadata": {
    "id": "3f14809e"
   },
   "source": [
    "# OPEN SOURCE ALTERNATIVE TO EMBEDDING"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "99138055",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 67,
     "referenced_widgets": [
      "c7800a5a8f1b4c3fa541df6b5ba6281f",
      "9d726c5753564258bbe1c281fa8bfefa",
      "ea3f511248734bbab197277b68346826",
      "3838d1bfccb5402d91ecf96f51f9c2d7",
      "ecc8ed6e869d4b01b7905ca951d0068c",
      "66f8508878e94e4296e2ffb009afdb75",
      "999f219c7c9449f0a48feb59eddd23eb",
      "7ab68b0af00248f5bbc15a39bbf38166",
      "b6ecc1cc5b2e4332b56fdc1614823d3e",
      "e732511315484861b39d0efb86b9cfb5",
      "3df968e9ff654de7bfc95f08c88a99c6"
     ]
    },
    "id": "99138055",
    "outputId": "2414eff0-be1c-44fb-be4c-a77d11ae4c57"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c7800a5a8f1b4c3fa541df6b5ba6281f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Batches:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(2, 768)"
      ]
     },
     "execution_count": 390,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "docs = [\"Around 9 Million people live in London\", \"London is known for its financial district\"]\n",
    "bi_encoder = SentenceTransformer(\"sentence-transformers/all-mpnet-base-v2\")\n",
    "doc_emb = bi_encoder.encode(docs, batch_size=32, show_progress_bar=True)\n",
    "\n",
    "doc_emb.shape#  == ('2, 768')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a1140c18",
   "metadata": {
    "id": "a1140c18"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be37c48b",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 49,
     "referenced_widgets": [
      "0caf250cdae643479160eedc708ffeea",
      "fa6089718c574ee19bd3c378e356ea6b",
      "f78b642ffb784f9097e0cbf97c44b9bd",
      "7c322a232739486a937883ac787f8750",
      "1bc5237b0db94f4aa124c8c79a891955",
      "2c83393c2b7745889765d8efb4890763",
      "6854fe2978954c7b9e30202a6ea34b57",
      "4f4986d008434fc9912c496636b40b38",
      "348701c9256e4eb48b8ffb0d9c0d6c83",
      "11e5c44fa73c4d5d88561036f2a0a070",
      "9781773e5d4d4e0ab8e3ff3140bd0983"
     ]
    },
    "id": "be37c48b",
    "outputId": "aeec917e-853f-4b24-facc-c8c42468f3fd"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0caf250cdae643479160eedc708ffeea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Batches:   0%|          | 0/36 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Encode query and documents\n",
    "docs = dataset['test']['context']\n",
    "doc_emb = bi_encoder.encode(docs, batch_size=32, show_progress_bar=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae0afee7",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ae0afee7",
    "outputId": "2ea967ab-1dba-473e-ca08-9927fbe24da5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Where are some destinations flights from Paris Orly operate to?\n",
      "Query: Where are some destinations flights from Paris Orly operate to?\n",
      "Document ID (Hash)\t\tRetrieval Score\tCE Score\tText\n",
      "2cefacc9828ba29e819c9a69e0b7b392\t0.69\t0.991276\tParis Orly Airport (French: Aéroport de Paris-Orly\n",
      "bcea700839a1cefc7fbe752fcfbee880\t0.24\t0.111188\tIran Air Flight 277 was a passenger flight which c\n",
      "20d2ff7276da85340534dabf110da251\t0.24\t0.096094\tAfter the Treaty of Paris in 1763, which deprived \n"
     ]
    }
   ],
   "source": [
    "from random import sample\n",
    "\n",
    "query = sample(dataset['test']['question'], 1)[0]\n",
    "print(query)\n",
    "final_results = get_results_from_pinecone(query, top_k=3, re_rank_model=finetuned)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45c71fb0",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "45c71fb0",
    "outputId": "98df154f-93f7-4d5a-b849-e2f8350e4965"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Where are some destinations flights from Paris Orly operate to?\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('Paris Orly Airport (French: Aéroport de Paris-Orly), commonly referred to as Orly (IATA: ORY, ICAO: LFPO), is an international airport located partially in Orly and partially in Villeneuve-le-Roi, 7 NM (13 km; 8.1 mi) south of Paris, France. It serves as a secondary hub for domestic and overseas territories flights of Air France and as the homebase for Transavia France.  Flights operate to destinations in Europe, the Middle East, Africa, the Caribbean, North America, East Asia and Southeast Asia. Prior to the opening of Charles de Gaulle Airport in March 1974, Orly was the main airport of Paris. Even with the shift of most international traffic to Charles de Gaulle Airport, Orly remains the busiest French airport for domestic traffic and the second busiest French airport overall in passenger traffic, with 33,120,685 passengers in 2018. The airport is operated by Groupe ADP under the brand Paris Aéroport. Since February 2018, the CEO of the airport has been Régis Lacote.',\n",
       "  0.6465857028961182,\n",
       "  933),\n",
       " ('Sevastopol International Airport (Ukrainian: Міжнародний аеропорт Севастополь \"Бельбек\", (IATA: UKS, ICAO: UKFB) is a military airfield in Belbek, a village near Sevastopol in Crimea.',\n",
       "  0.3177761733531952,\n",
       "  1020),\n",
       " ('After the Treaty of Paris in 1763, which deprived France of almost all her possessions in the Americas other than Guiana and a few islands, Louis XV sent thousands of settlers to Guiana who were lured there with stories of plentiful gold and easy fortunes to be made. Instead they found a land filled with hostile natives and tropical diseases. One and a half years later only a few hundred survived. These fled to three small islands which could be seen off shore and named them the Iles de Salut (or \"Islands of Salvation\"). The largest was called Royal Island, another St. Joseph (after the patron saint of the expedition), and the smallest of the islands, surrounded by strong currents, Île du Diable (the infamous \"Devil\\'s Island\"). When the survivors of this ill-fated expedition returned home, the terrible stories they told of the colony left a lasting impression in France.',\n",
       "  0.3029923439025879,\n",
       "  349)]"
      ]
     },
     "execution_count": 393,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(query)\n",
    "find_most_similar(query, doc_emb, docs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93280be8",
   "metadata": {
    "id": "93280be8"
   },
   "outputs": [],
   "source": [
    "def eval_ranking_open_source(query, top_k=3, re_rank_model=None):\n",
    "    ans = {'retrieved_correct_position': None}\n",
    "    correct_hash = q_to_hash[query]\n",
    "    results = find_most_similar(query, doc_emb, docs, k=top_k)\n",
    "    for idx, (passage, score, doc_idx) in enumerate(results):\n",
    "        if correct_hash == my_hash(passage):\n",
    "            ans['retrieved_correct_position'] =  idx\n",
    "    if re_rank_model is not None:\n",
    "        ans['reranked_correct_position'] = None\n",
    "        sentence_combinations = [(query, r[0]) for r in results]\n",
    "\n",
    "        # Compute the similarity scores for these combinations\n",
    "        similarity_scores = re_rank_model.predict(sentence_combinations, activation_fct=nn.Sigmoid())\n",
    "\n",
    "        # Sort the scores in decreasing order\n",
    "        sim_scores_argsort = list(reversed(np.argsort(similarity_scores)))\n",
    "        for i, idx in enumerate(sim_scores_argsort):\n",
    "            r = results[idx]\n",
    "            if correct_hash and my_hash(r[0]) == correct_hash:\n",
    "                ans['reranked_correct_position'] = i\n",
    "\n",
    "    return ans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06f198cd",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "06f198cd",
    "outputId": "ef371189-4a07-43b7-8761-da5ad013fdeb",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'retrieved_correct_position': 0, 'reranked_correct_position': 0}"
      ]
     },
     "execution_count": 420,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eval_ranking_open_source(query, top_k=TOP_K, re_rank_model=finetuned)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8faf0c1a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 224
    },
    "id": "8faf0c1a",
    "outputId": "33c6657e-f529-4cc7-9ee5-1dff61e4172f",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1148/1148 [13:31<00:00,  1.42it/s]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-da5f3659-2adf-46d9-b477-ddc7ef9a51f3\" class=\"colab-df-container\">\n",
       "    <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>retrieved_correct_position</th>\n",
       "      <th>reranked_correct_position</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "    <div class=\"colab-df-buttons\">\n",
       "\n",
       "  <div class=\"colab-df-container\">\n",
       "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-da5f3659-2adf-46d9-b477-ddc7ef9a51f3')\"\n",
       "            title=\"Convert this dataframe to an interactive table.\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
       "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
       "  </svg>\n",
       "    </button>\n",
       "\n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    .colab-df-buttons div {\n",
       "      margin-bottom: 4px;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "    <script>\n",
       "      const buttonEl =\n",
       "        document.querySelector('#df-da5f3659-2adf-46d9-b477-ddc7ef9a51f3 button.colab-df-convert');\n",
       "      buttonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "      async function convertToInteractive(key) {\n",
       "        const element = document.querySelector('#df-da5f3659-2adf-46d9-b477-ddc7ef9a51f3');\n",
       "        const dataTable =\n",
       "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                    [key], {});\n",
       "        if (!dataTable) return;\n",
       "\n",
       "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "          + ' to learn more about interactive tables.';\n",
       "        element.innerHTML = '';\n",
       "        dataTable['output_type'] = 'display_data';\n",
       "        await google.colab.output.renderOutput(dataTable, element);\n",
       "        const docLink = document.createElement('div');\n",
       "        docLink.innerHTML = docLinkHtml;\n",
       "        element.appendChild(docLink);\n",
       "      }\n",
       "    </script>\n",
       "  </div>\n",
       "\n",
       "\n",
       "<div id=\"df-e7255059-6dad-48f1-9b6b-21545152f562\">\n",
       "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-e7255059-6dad-48f1-9b6b-21545152f562')\"\n",
       "            title=\"Suggest charts\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "     width=\"24px\">\n",
       "    <g>\n",
       "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
       "    </g>\n",
       "</svg>\n",
       "  </button>\n",
       "\n",
       "<style>\n",
       "  .colab-df-quickchart {\n",
       "      --bg-color: #E8F0FE;\n",
       "      --fill-color: #1967D2;\n",
       "      --hover-bg-color: #E2EBFA;\n",
       "      --hover-fill-color: #174EA6;\n",
       "      --disabled-fill-color: #AAA;\n",
       "      --disabled-bg-color: #DDD;\n",
       "  }\n",
       "\n",
       "  [theme=dark] .colab-df-quickchart {\n",
       "      --bg-color: #3B4455;\n",
       "      --fill-color: #D2E3FC;\n",
       "      --hover-bg-color: #434B5C;\n",
       "      --hover-fill-color: #FFFFFF;\n",
       "      --disabled-bg-color: #3B4455;\n",
       "      --disabled-fill-color: #666;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart {\n",
       "    background-color: var(--bg-color);\n",
       "    border: none;\n",
       "    border-radius: 50%;\n",
       "    cursor: pointer;\n",
       "    display: none;\n",
       "    fill: var(--fill-color);\n",
       "    height: 32px;\n",
       "    padding: 0;\n",
       "    width: 32px;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart:hover {\n",
       "    background-color: var(--hover-bg-color);\n",
       "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "    fill: var(--button-hover-fill-color);\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart-complete:disabled,\n",
       "  .colab-df-quickchart-complete:disabled:hover {\n",
       "    background-color: var(--disabled-bg-color);\n",
       "    fill: var(--disabled-fill-color);\n",
       "    box-shadow: none;\n",
       "  }\n",
       "\n",
       "  .colab-df-spinner {\n",
       "    border: 2px solid var(--fill-color);\n",
       "    border-color: transparent;\n",
       "    border-bottom-color: var(--fill-color);\n",
       "    animation:\n",
       "      spin 1s steps(1) infinite;\n",
       "  }\n",
       "\n",
       "  @keyframes spin {\n",
       "    0% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "      border-left-color: var(--fill-color);\n",
       "    }\n",
       "    20% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    30% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    40% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    60% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    80% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "    90% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "  }\n",
       "</style>\n",
       "\n",
       "  <script>\n",
       "    async function quickchart(key) {\n",
       "      const quickchartButtonEl =\n",
       "        document.querySelector('#' + key + ' button');\n",
       "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
       "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
       "      try {\n",
       "        const charts = await google.colab.kernel.invokeFunction(\n",
       "            'suggestCharts', [key], {});\n",
       "      } catch (error) {\n",
       "        console.error('Error during call to suggestCharts:', error);\n",
       "      }\n",
       "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
       "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
       "    }\n",
       "    (() => {\n",
       "      let quickchartButtonEl =\n",
       "        document.querySelector('#df-e7255059-6dad-48f1-9b6b-21545152f562 button');\n",
       "      quickchartButtonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "    })();\n",
       "  </script>\n",
       "</div>\n",
       "\n",
       "    </div>\n",
       "  </div>\n"
      ],
      "text/plain": [
       "   retrieved_correct_position  reranked_correct_position\n",
       "0                         0.0                        0.0\n",
       "1                         0.0                        0.0\n",
       "2                         0.0                        0.0\n",
       "3                         1.0                        0.0\n",
       "4                         0.0                        0.0"
      ]
     },
     "execution_count": 423,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logger.setLevel(logging.CRITICAL)\n",
    "os_predictions = []\n",
    "\n",
    "for i, question in tqdm(enumerate(test_sample), total=len(test_sample)):\n",
    "    os_predictions.append(eval_ranking_open_source(question['question'], top_k=TOP_K, re_rank_model=finetuned))\n",
    "\n",
    "os_predictions_df = pd.DataFrame(os_predictions)\n",
    "os_predictions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6eaeb9b4",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "6eaeb9b4",
    "outputId": "01787cbe-fd05-4cf6-83bc-66c8573d0dd6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy without re-ranking: 0.5017421602787456\n",
      "Accuracy with re-ranking: 0.6193379790940766\n"
     ]
    }
   ],
   "source": [
    "raw_accuracy = sum([p['retrieved_correct_position'] == 0 for p in os_predictions])/len(os_predictions)\n",
    "reranked_accuracy = sum([p['reranked_correct_position'] == 0 for p in os_predictions])/len(os_predictions)\n",
    "\n",
    "print(f'Accuracy without re-ranking: {raw_accuracy}')\n",
    "print(f'Accuracy with re-ranking: {reranked_accuracy}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "T00iM9SyaUFI",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "T00iM9SyaUFI",
    "outputId": "c9dda451-d289-4734-c94c-86dcd0fb155c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 576 711\n",
      "3 887 1006\n",
      "5 948 1055\n",
      "10 1008 1087\n",
      "25 1073 1103\n",
      "50 1107 1107\n"
     ]
    }
   ],
   "source": [
    "# do recall @ 1, 3, 5, 10\n",
    "OPEN_SOURCE_RETRIEVAL = []\n",
    "OPEN_SOURCE_RETRIEVAL_PLUS_FT_CE = []\n",
    "for k in (1, 3, 5, 10, 25, 50):\n",
    "    embedding_only_recall = os_predictions_df[os_predictions_df['retrieved_correct_position'] < k].shape[0]\n",
    "    reranked_recall = os_predictions_df[os_predictions_df['reranked_correct_position'] < k].shape[0]\n",
    "    print(k, embedding_only_recall, reranked_recall)\n",
    "    OPEN_SOURCE_RETRIEVAL.append(embedding_only_recall / os_predictions_df.shape[0])\n",
    "    OPEN_SOURCE_RETRIEVAL_PLUS_FT_CE.append(reranked_recall / os_predictions_df.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "sZD_4weEahOm",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 564
    },
    "id": "sZD_4weEahOm",
    "outputId": "df7327c7-50a5-454f-9daa-69cb7e9ca1b2"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "X = [1, 3, 5, 10, 25, 50]\n",
    "\n",
    "OPENAI_RETRIEVAL = [866 / predictions_df.shape[0], 1009/ predictions_df.shape[0], 1048/ predictions_df.shape[0], 1072/ predictions_df.shape[0], 1108/ predictions_df.shape[0], 1122/ predictions_df.shape[0]]\n",
    "# OPEN_SOURCE_RETRIEVAL = [663 / predictions_df.shape[0], 795/ predictions_df.shape[0], 838/ predictions_df.shape[0], 891/ predictions_df.shape[0], 966/ predictions_df.shape[0], 1004/ predictions_df.shape[0]]\n",
    "OLD_CROSS_ENCODER = [956 / predictions_df.shape[0], 1057/ predictions_df.shape[0], 1077/ predictions_df.shape[0], 1094/ predictions_df.shape[0], 1120/ predictions_df.shape[0], 1122/ predictions_df.shape[0]]\n",
    "NEW_CROSS_ENCODER = [975 / predictions_df.shape[0], 1056/ predictions_df.shape[0], 1077/ predictions_df.shape[0], 1101/ predictions_df.shape[0], 1117/ predictions_df.shape[0], 1122/ predictions_df.shape[0]]\n",
    "\n",
    "# Creating the plot\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.plot(X, OPENAI_RETRIEVAL, label='OAI Retrieval Only', marker='o')\n",
    "plt.plot(X, OPEN_SOURCE_RETRIEVAL, label='OS Retrieval Only', marker='*')\n",
    "plt.plot(X, OPEN_SOURCE_RETRIEVAL_PLUS_FT_CE, label='OS + Finetuned CE', marker='v')\n",
    "plt.plot(X, OLD_CROSS_ENCODER, label='OAI + Pretrained CE', marker='s')\n",
    "plt.plot(X, NEW_CROSS_ENCODER, label='OAI + Finetuned CE', marker='^')\n",
    "\n",
    "# Adding titles and labels\n",
    "plt.title('Comparing embedding models + pre-trained vs fine-tuned CE (all retrieved 50 results then re-ranked)')\n",
    "plt.xlabel('Recall @')\n",
    "plt.ylabel('Performance')\n",
    "plt.xticks(X)\n",
    "plt.yticks([i/100 for i in range(70, 101, 5)])  # Adjusting y-ticks to start from 0.75\n",
    "\n",
    "# Adding legend\n",
    "plt.legend()\n",
    "\n",
    "# Show the plot\n",
    "plt.grid(True)\n",
    "# plt.show()\n",
    "\n",
    "plt.savefig('recall_at_k.png', dpi=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "LFaZAKPhaymM",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 275
    },
    "id": "LFaZAKPhaymM",
    "outputId": "a58e68c7-f61e-420a-aef9-c8336bb48702"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-4363c439-77de-4ee1-b439-f4d0dbd94b7a\" class=\"colab-df-container\">\n",
       "    <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>RECALL @</th>\n",
       "      <th>OS_Retrieval_Only</th>\n",
       "      <th>OS_Retrieval_Plus_Finetuned_CE</th>\n",
       "      <th>OAI_Retrieval_Only</th>\n",
       "      <th>OAI_Retrieval_Plus_Pretrained_CE</th>\n",
       "      <th>OAI_Retrieval_Plus_Finetuned_CE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.501742</td>\n",
       "      <td>0.619338</td>\n",
       "      <td>0.754355</td>\n",
       "      <td>0.832753</td>\n",
       "      <td>0.849303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>0.772648</td>\n",
       "      <td>0.876307</td>\n",
       "      <td>0.878920</td>\n",
       "      <td>0.920732</td>\n",
       "      <td>0.919861</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>0.825784</td>\n",
       "      <td>0.918990</td>\n",
       "      <td>0.912892</td>\n",
       "      <td>0.938153</td>\n",
       "      <td>0.938153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>0.878049</td>\n",
       "      <td>0.946864</td>\n",
       "      <td>0.933798</td>\n",
       "      <td>0.952962</td>\n",
       "      <td>0.959059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>25</td>\n",
       "      <td>0.934669</td>\n",
       "      <td>0.960801</td>\n",
       "      <td>0.965157</td>\n",
       "      <td>0.975610</td>\n",
       "      <td>0.972997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>50</td>\n",
       "      <td>0.964286</td>\n",
       "      <td>0.964286</td>\n",
       "      <td>0.977352</td>\n",
       "      <td>0.977352</td>\n",
       "      <td>0.977352</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "    <div class=\"colab-df-buttons\">\n",
       "\n",
       "  <div class=\"colab-df-container\">\n",
       "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4363c439-77de-4ee1-b439-f4d0dbd94b7a')\"\n",
       "            title=\"Convert this dataframe to an interactive table.\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
       "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
       "  </svg>\n",
       "    </button>\n",
       "\n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    .colab-df-buttons div {\n",
       "      margin-bottom: 4px;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "    <script>\n",
       "      const buttonEl =\n",
       "        document.querySelector('#df-4363c439-77de-4ee1-b439-f4d0dbd94b7a button.colab-df-convert');\n",
       "      buttonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "      async function convertToInteractive(key) {\n",
       "        const element = document.querySelector('#df-4363c439-77de-4ee1-b439-f4d0dbd94b7a');\n",
       "        const dataTable =\n",
       "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                    [key], {});\n",
       "        if (!dataTable) return;\n",
       "\n",
       "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "          + ' to learn more about interactive tables.';\n",
       "        element.innerHTML = '';\n",
       "        dataTable['output_type'] = 'display_data';\n",
       "        await google.colab.output.renderOutput(dataTable, element);\n",
       "        const docLink = document.createElement('div');\n",
       "        docLink.innerHTML = docLinkHtml;\n",
       "        element.appendChild(docLink);\n",
       "      }\n",
       "    </script>\n",
       "  </div>\n",
       "\n",
       "\n",
       "<div id=\"df-37d7a99c-666e-4136-93f6-2ddeb94c2b86\">\n",
       "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-37d7a99c-666e-4136-93f6-2ddeb94c2b86')\"\n",
       "            title=\"Suggest charts\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "     width=\"24px\">\n",
       "    <g>\n",
       "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
       "    </g>\n",
       "</svg>\n",
       "  </button>\n",
       "\n",
       "<style>\n",
       "  .colab-df-quickchart {\n",
       "      --bg-color: #E8F0FE;\n",
       "      --fill-color: #1967D2;\n",
       "      --hover-bg-color: #E2EBFA;\n",
       "      --hover-fill-color: #174EA6;\n",
       "      --disabled-fill-color: #AAA;\n",
       "      --disabled-bg-color: #DDD;\n",
       "  }\n",
       "\n",
       "  [theme=dark] .colab-df-quickchart {\n",
       "      --bg-color: #3B4455;\n",
       "      --fill-color: #D2E3FC;\n",
       "      --hover-bg-color: #434B5C;\n",
       "      --hover-fill-color: #FFFFFF;\n",
       "      --disabled-bg-color: #3B4455;\n",
       "      --disabled-fill-color: #666;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart {\n",
       "    background-color: var(--bg-color);\n",
       "    border: none;\n",
       "    border-radius: 50%;\n",
       "    cursor: pointer;\n",
       "    display: none;\n",
       "    fill: var(--fill-color);\n",
       "    height: 32px;\n",
       "    padding: 0;\n",
       "    width: 32px;\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart:hover {\n",
       "    background-color: var(--hover-bg-color);\n",
       "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "    fill: var(--button-hover-fill-color);\n",
       "  }\n",
       "\n",
       "  .colab-df-quickchart-complete:disabled,\n",
       "  .colab-df-quickchart-complete:disabled:hover {\n",
       "    background-color: var(--disabled-bg-color);\n",
       "    fill: var(--disabled-fill-color);\n",
       "    box-shadow: none;\n",
       "  }\n",
       "\n",
       "  .colab-df-spinner {\n",
       "    border: 2px solid var(--fill-color);\n",
       "    border-color: transparent;\n",
       "    border-bottom-color: var(--fill-color);\n",
       "    animation:\n",
       "      spin 1s steps(1) infinite;\n",
       "  }\n",
       "\n",
       "  @keyframes spin {\n",
       "    0% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "      border-left-color: var(--fill-color);\n",
       "    }\n",
       "    20% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    30% {\n",
       "      border-color: transparent;\n",
       "      border-left-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    40% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-top-color: var(--fill-color);\n",
       "    }\n",
       "    60% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "    }\n",
       "    80% {\n",
       "      border-color: transparent;\n",
       "      border-right-color: var(--fill-color);\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "    90% {\n",
       "      border-color: transparent;\n",
       "      border-bottom-color: var(--fill-color);\n",
       "    }\n",
       "  }\n",
       "</style>\n",
       "\n",
       "  <script>\n",
       "    async function quickchart(key) {\n",
       "      const quickchartButtonEl =\n",
       "        document.querySelector('#' + key + ' button');\n",
       "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
       "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
       "      try {\n",
       "        const charts = await google.colab.kernel.invokeFunction(\n",
       "            'suggestCharts', [key], {});\n",
       "      } catch (error) {\n",
       "        console.error('Error during call to suggestCharts:', error);\n",
       "      }\n",
       "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
       "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
       "    }\n",
       "    (() => {\n",
       "      let quickchartButtonEl =\n",
       "        document.querySelector('#df-37d7a99c-666e-4136-93f6-2ddeb94c2b86 button');\n",
       "      quickchartButtonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "    })();\n",
       "  </script>\n",
       "</div>\n",
       "\n",
       "  <div id=\"id_20b5e92e-5da3-4334-a0e4-03dad314ab26\">\n",
       "    <style>\n",
       "      .colab-df-generate {\n",
       "        background-color: #E8F0FE;\n",
       "        border: none;\n",
       "        border-radius: 50%;\n",
       "        cursor: pointer;\n",
       "        display: none;\n",
       "        fill: #1967D2;\n",
       "        height: 32px;\n",
       "        padding: 0 0 0 0;\n",
       "        width: 32px;\n",
       "      }\n",
       "\n",
       "      .colab-df-generate:hover {\n",
       "        background-color: #E2EBFA;\n",
       "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "        fill: #174EA6;\n",
       "      }\n",
       "\n",
       "      [theme=dark] .colab-df-generate {\n",
       "        background-color: #3B4455;\n",
       "        fill: #D2E3FC;\n",
       "      }\n",
       "\n",
       "      [theme=dark] .colab-df-generate:hover {\n",
       "        background-color: #434B5C;\n",
       "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "        fill: #FFFFFF;\n",
       "      }\n",
       "    </style>\n",
       "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('results_df')\"\n",
       "            title=\"Generate code using this dataframe.\"\n",
       "            style=\"display:none;\">\n",
       "\n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
       "  </svg>\n",
       "    </button>\n",
       "    <script>\n",
       "      (() => {\n",
       "      const buttonEl =\n",
       "        document.querySelector('#id_20b5e92e-5da3-4334-a0e4-03dad314ab26 button.colab-df-generate');\n",
       "      buttonEl.style.display =\n",
       "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "      buttonEl.onclick = () => {\n",
       "        google.colab.notebook.generateWithVariable('results_df');\n",
       "      }\n",
       "      })();\n",
       "    </script>\n",
       "  </div>\n",
       "\n",
       "    </div>\n",
       "  </div>\n"
      ],
      "text/plain": [
       "   RECALL @  OS_Retrieval_Only  OS_Retrieval_Plus_Finetuned_CE  \\\n",
       "0         1           0.501742                        0.619338   \n",
       "1         3           0.772648                        0.876307   \n",
       "2         5           0.825784                        0.918990   \n",
       "3        10           0.878049                        0.946864   \n",
       "4        25           0.934669                        0.960801   \n",
       "5        50           0.964286                        0.964286   \n",
       "\n",
       "   OAI_Retrieval_Only  OAI_Retrieval_Plus_Pretrained_CE  \\\n",
       "0            0.754355                          0.832753   \n",
       "1            0.878920                          0.920732   \n",
       "2            0.912892                          0.938153   \n",
       "3            0.933798                          0.952962   \n",
       "4            0.965157                          0.975610   \n",
       "5            0.977352                          0.977352   \n",
       "\n",
       "   OAI_Retrieval_Plus_Finetuned_CE  \n",
       "0                         0.849303  \n",
       "1                         0.919861  \n",
       "2                         0.938153  \n",
       "3                         0.959059  \n",
       "4                         0.972997  \n",
       "5                         0.977352  "
      ]
     },
     "execution_count": 448,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show results as a table\n",
    "\n",
    "results_df = pd.DataFrame({'RECALL @': [1, 3, 5, 10, 25, 50], 'OS_Retrieval_Only': OPEN_SOURCE_RETRIEVAL, 'OS_Retrieval_Plus_Finetuned_CE': OPEN_SOURCE_RETRIEVAL_PLUS_FT_CE   , 'OAI_Retrieval_Only': OPENAI_RETRIEVAL    , 'OAI_Retrieval_Plus_Pretrained_CE': OLD_CROSS_ENCODER, 'OAI_Retrieval_Plus_Finetuned_CE': NEW_CROSS_ENCODER})\n",
    "results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "WpumO13hi_Vr",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 568
    },
    "id": "WpumO13hi_Vr",
    "outputId": "d00fbc68-d2dd-462e-c6a9-2744d178c1c9"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Assuming 'results_df' is your DataFrame\n",
    "# Set 'RECALL @' as the index if it's intended to be used for x-axis labels\n",
    "results_df.set_index('RECALL @', inplace=True)\n",
    "\n",
    "# Plot the bar chart\n",
    "ax = results_df.plot.bar(\n",
    "    xlabel='RECALL @',\n",
    "    ylabel='Performance',\n",
    "    title='Comparing embedding models + pre-trained vs fine-tuned CE (all retrieved 50 results then re-ranked)',\n",
    "    figsize=(10, 6)\n",
    ")\n",
    "\n",
    "# Move the legend outside of the plot\n",
    "ax.legend(loc='upper left', bbox_to_anchor=(1, 1))\n",
    "\n",
    "# Adjust subplot parameters to fit the legend outside of the plot\n",
    "plt.subplots_adjust(right=0.75)\n",
    "\n",
    "plt.savefig('recall_at_k_bar.png', dpi=1000)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a6df164-60c7-4a36-b246-c0c92297a496",
   "metadata": {},
   "source": [
    "# Using a deployed open source embedder (~$45 a month)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "INH8fgQUBdka",
   "metadata": {
    "id": "INH8fgQUBdka"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "HF_TOKEN = userdata.get('HF_TOKEN')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "1721eda4-eee5-4a62-90e4-24577e905a93",
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "\n",
    "API_URL = \"https://rnlg7ggtdkep4h0i.us-east-1.aws.endpoints.huggingface.cloud\"\n",
    "headers = {\n",
    "\t\"Accept\" : \"application/json\",\n",
    "\t\"Authorization\": f\"Bearer {HF_TOKEN}\",\n",
    "\t\"Content-Type\": \"application/json\" \n",
    "}\n",
    "\n",
    "def query(payload):\n",
    "\tresponse = requests.post(API_URL, headers=headers, json=payload)\n",
    "\treturn response.json()\n",
    "\n",
    "output = query({\n",
    "\t\"inputs\": [\"1\", '2'],\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "81c513bc-882c-474e-9efb-2696e4731f35",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(output['embeddings'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "1ad100e6-0d6c-4f77-bd2d-a063976d8b69",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "768"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(output['embeddings'][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "ee28d40b-0973-40c3-8230-abf007471a72",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.012244473211467266,\n",
       " -0.009157375432550907,\n",
       " 0.004602040629833937,\n",
       " -0.024496469646692276,\n",
       " -0.004610566888004541]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output['embeddings'][0][:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eeb65a98-6559-4895-ac65-682d7979bd66",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "gpuType": "T4",
   "machine_shape": "hm",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "0295736b3f254a8590d911792f2c19db": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_3f1ba16f17c94fb8a409d152b9a6045a",
      "max": 1739,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_c485139a148e4207957e96b4275eeed2",
      "value": 1739
     }
    },
    "084ebc36eb9d4616acc1a0843673ac4f": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "098016d32abc4f45be0a4f747cbc3f11": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "0caf250cdae643479160eedc708ffeea": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_fa6089718c574ee19bd3c378e356ea6b",
       "IPY_MODEL_f78b642ffb784f9097e0cbf97c44b9bd",
       "IPY_MODEL_7c322a232739486a937883ac787f8750"
      ],
      "layout": "IPY_MODEL_1bc5237b0db94f4aa124c8c79a891955"
     }
    },
    "11e5c44fa73c4d5d88561036f2a0a070": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "168948926c2340abb8b52f8dd67a5e73": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1bc5237b0db94f4aa124c8c79a891955": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "21316ef8a22d40ab9109ccf0fa7fa645": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_458bcab796d642c9afd2e534085b991f",
       "IPY_MODEL_266b42f10bd74b9f907588c1251c8d6c",
       "IPY_MODEL_ba968ae891a2400fa0db9f331b85b66c"
      ],
      "layout": "IPY_MODEL_5d09ef7e47f446a3a3a51458195a0fe3"
     }
    },
    "252dfe4d46604a089c263bb0eae17a3a": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "266b42f10bd74b9f907588c1251c8d6c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_b070922627a04195a41899a169984b08",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_85a67e7f5d2845329d63f0922bb5d9a6",
      "value": 1
     }
    },
    "2aa32422748943759daf5ef05077bda0": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2c83393c2b7745889765d8efb4890763": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2cb4c1bff01b4f669ecd8583187bab9f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "348701c9256e4eb48b8ffb0d9c0d6c83": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "352af51e6e7943a7a41d3bbe41885d8d": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "37395cf0cac548c783ccbdac230c4a3c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_8085fe243d814d5da2b4d6b665a6aa51",
       "IPY_MODEL_0295736b3f254a8590d911792f2c19db",
       "IPY_MODEL_ad079ced956841be88a3ab6f2fdcd88d"
      ],
      "layout": "IPY_MODEL_953b16307b514bc995efc40ff8f4698b"
     }
    },
    "3838d1bfccb5402d91ecf96f51f9c2d7": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e732511315484861b39d0efb86b9cfb5",
      "placeholder": "​",
      "style": "IPY_MODEL_3df968e9ff654de7bfc95f08c88a99c6",
      "value": " 1/1 [00:00&lt;00:00, 33.52it/s]"
     }
    },
    "3df968e9ff654de7bfc95f08c88a99c6": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "3f1ba16f17c94fb8a409d152b9a6045a": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "43c8db4be4984f6da1540de6bc89b616": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_dc26c84713fb45f5a6c5b870c96d9d9e",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_a6ac9d7aed4c493db567b015ab8502a5",
      "value": 1
     }
    },
    "458bcab796d642c9afd2e534085b991f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_352af51e6e7943a7a41d3bbe41885d8d",
      "placeholder": "​",
      "style": "IPY_MODEL_2cb4c1bff01b4f669ecd8583187bab9f",
      "value": "Batches: 100%"
     }
    },
    "4d26773cf0ec45fc9ac3185f8141cfe6": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "4f4986d008434fc9912c496636b40b38": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5a7b319329164b27bf452869e2335aef": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5b3c84192618410ea1793239c4514c6e": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5d09ef7e47f446a3a3a51458195a0fe3": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "608d1345487046ceb64b0b31a4287ae0": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_fce9aa1662c34941afadd6141aba9180",
      "placeholder": "​",
      "style": "IPY_MODEL_098016d32abc4f45be0a4f747cbc3f11",
      "value": "Batches: 100%"
     }
    },
    "66f8508878e94e4296e2ffb009afdb75": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "6854fe2978954c7b9e30202a6ea34b57": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "687d7bf7373f4541852203637e18d2e3": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "6cb81ad9ee8141aa9d0fbefc276f244d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_80e18075a957476591fa1ba2585b7a73",
      "placeholder": "​",
      "style": "IPY_MODEL_e1d8ec56e5614d77a6999a42283be94b",
      "value": " 1/1 [00:00&lt;00:00, 27.76it/s]"
     }
    },
    "6f5389fdc3674a05ad79bb9456a5995f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e281ca37b72d422285aa64193265aaa4",
      "placeholder": "​",
      "style": "IPY_MODEL_687d7bf7373f4541852203637e18d2e3",
      "value": "Batches: 100%"
     }
    },
    "6f96c8a415744597bcaa83141b3a457c": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "70e45cb4f2f64c499acc785ec64743fe": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_d9f6781147a24bc3aec97b87927bdfbb",
      "placeholder": "​",
      "style": "IPY_MODEL_d0c85ebbf1a54fdcaa33210aa3e247b8",
      "value": " 620/620 [00:30&lt;00:00, 100.95it/s]"
     }
    },
    "7a5c511f2fe84c66beef7fed87023669": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5a7b319329164b27bf452869e2335aef",
      "max": 620,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_e5a88639e7e5430fbe44f73d2e5a4b23",
      "value": 620
     }
    },
    "7ab68b0af00248f5bbc15a39bbf38166": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "7bd0527b8e174aec9911c5b32dc048fd": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "7c322a232739486a937883ac787f8750": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_11e5c44fa73c4d5d88561036f2a0a070",
      "placeholder": "​",
      "style": "IPY_MODEL_9781773e5d4d4e0ab8e3ff3140bd0983",
      "value": " 36/36 [00:16&lt;00:00,  5.64it/s]"
     }
    },
    "8085fe243d814d5da2b4d6b665a6aa51": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_4d26773cf0ec45fc9ac3185f8141cfe6",
      "placeholder": "​",
      "style": "IPY_MODEL_d7959d1933e5466095bed23bb268afd5",
      "value": "Iteration: 100%"
     }
    },
    "80e18075a957476591fa1ba2585b7a73": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "8396e1788d5540a79a9d195e2753351a": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "85a67e7f5d2845329d63f0922bb5d9a6": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "8e41a0a7932b4f0b8a448847ff3f14ff": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "8e989ddf9d274876a3ba06056e7147a1": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "953b16307b514bc995efc40ff8f4698b": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "9781773e5d4d4e0ab8e3ff3140bd0983": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "999f219c7c9449f0a48feb59eddd23eb": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "9d726c5753564258bbe1c281fa8bfefa": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_66f8508878e94e4296e2ffb009afdb75",
      "placeholder": "​",
      "style": "IPY_MODEL_999f219c7c9449f0a48feb59eddd23eb",
      "value": "Batches: 100%"
     }
    },
    "a6ac9d7aed4c493db567b015ab8502a5": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "ad079ced956841be88a3ab6f2fdcd88d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_6f96c8a415744597bcaa83141b3a457c",
      "placeholder": "​",
      "style": "IPY_MODEL_8e41a0a7932b4f0b8a448847ff3f14ff",
      "value": " 1739/1739 [10:05&lt;00:00,  2.86it/s]"
     }
    },
    "b070922627a04195a41899a169984b08": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "b281e2a5a8204b52827f01316d637952": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_084ebc36eb9d4616acc1a0843673ac4f",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_252dfe4d46604a089c263bb0eae17a3a",
      "value": 1
     }
    },
    "b6ecc1cc5b2e4332b56fdc1614823d3e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "b86aa89087cb425198afa2db30d86600": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "b9fbabd33a2a4d0a94055cd086d83228": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_168948926c2340abb8b52f8dd67a5e73",
      "placeholder": "​",
      "style": "IPY_MODEL_8396e1788d5540a79a9d195e2753351a",
      "value": "Epoch: 100%"
     }
    },
    "ba968ae891a2400fa0db9f331b85b66c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5b3c84192618410ea1793239c4514c6e",
      "placeholder": "​",
      "style": "IPY_MODEL_7bd0527b8e174aec9911c5b32dc048fd",
      "value": " 1/1 [00:00&lt;00:00, 28.53it/s]"
     }
    },
    "bb260568c96a4397bba2b1058404eec2": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "c485139a148e4207957e96b4275eeed2": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "c7800a5a8f1b4c3fa541df6b5ba6281f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_9d726c5753564258bbe1c281fa8bfefa",
       "IPY_MODEL_ea3f511248734bbab197277b68346826",
       "IPY_MODEL_3838d1bfccb5402d91ecf96f51f9c2d7"
      ],
      "layout": "IPY_MODEL_ecc8ed6e869d4b01b7905ca951d0068c"
     }
    },
    "cdd38d9a948a4a4c8408dcb06d7c584a": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2aa32422748943759daf5ef05077bda0",
      "placeholder": "​",
      "style": "IPY_MODEL_e7c672eb19574b32811bdcc3f6ac0b34",
      "value": " 1/1 [11:44&lt;00:00, 704.92s/it]"
     }
    },
    "d0c85ebbf1a54fdcaa33210aa3e247b8": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "d7959d1933e5466095bed23bb268afd5": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "d9f6781147a24bc3aec97b87927bdfbb": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "dbca2d7da970438892d45cd387434f6f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_6f5389fdc3674a05ad79bb9456a5995f",
       "IPY_MODEL_7a5c511f2fe84c66beef7fed87023669",
       "IPY_MODEL_70e45cb4f2f64c499acc785ec64743fe"
      ],
      "layout": "IPY_MODEL_bb260568c96a4397bba2b1058404eec2"
     }
    },
    "dc26c84713fb45f5a6c5b870c96d9d9e": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e1d8ec56e5614d77a6999a42283be94b": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "e281ca37b72d422285aa64193265aaa4": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e31ae00bfabe4ed5b058d7dbe55a80f6": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_b9fbabd33a2a4d0a94055cd086d83228",
       "IPY_MODEL_b281e2a5a8204b52827f01316d637952",
       "IPY_MODEL_cdd38d9a948a4a4c8408dcb06d7c584a"
      ],
      "layout": "IPY_MODEL_b86aa89087cb425198afa2db30d86600"
     }
    },
    "e5a88639e7e5430fbe44f73d2e5a4b23": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "e732511315484861b39d0efb86b9cfb5": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e7c672eb19574b32811bdcc3f6ac0b34": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "ea3f511248734bbab197277b68346826": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_7ab68b0af00248f5bbc15a39bbf38166",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_b6ecc1cc5b2e4332b56fdc1614823d3e",
      "value": 1
     }
    },
    "ecc8ed6e869d4b01b7905ca951d0068c": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f36d8f08f6284b848622fd560196cc78": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_608d1345487046ceb64b0b31a4287ae0",
       "IPY_MODEL_43c8db4be4984f6da1540de6bc89b616",
       "IPY_MODEL_6cb81ad9ee8141aa9d0fbefc276f244d"
      ],
      "layout": "IPY_MODEL_8e989ddf9d274876a3ba06056e7147a1"
     }
    },
    "f78b642ffb784f9097e0cbf97c44b9bd": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_4f4986d008434fc9912c496636b40b38",
      "max": 36,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_348701c9256e4eb48b8ffb0d9c0d6c83",
      "value": 36
     }
    },
    "fa6089718c574ee19bd3c378e356ea6b": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2c83393c2b7745889765d8efb4890763",
      "placeholder": "​",
      "style": "IPY_MODEL_6854fe2978954c7b9e30202a6ea34b57",
      "value": "Batches: 100%"
     }
    },
    "fce9aa1662c34941afadd6141aba9180": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}