Demo_1_Solutions_for_Missing_Content_in_RAG.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "authorship_tag": "ABX9TyPsJLb1ErDGe1LB56inJND4",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/dipanjanS/improving-RAG-systems-dhs2024/blob/main/Demo_1_Solutions_for_Missing_Content_in_RAG.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Solutions for Missing Content in RAG\n",
        "\n",
        "Here we will explore the following strategies\n",
        "\n",
        "- Better Data Cleaning\n",
        "- Better Prompting\n",
        "- Agentic RAG with Tools\n"
      ],
      "metadata": {
        "id": "b8kkhS-UgJAR"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Exploring Useful Data Loaders"
      ],
      "metadata": {
        "id": "mjyiD8I5r7jB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Install OpenAI, HuggingFace and LangChain dependencies"
      ],
      "metadata": {
        "id": "L1KvMtf54l0d"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install langchain\n",
        "!pip install langchain-openai\n",
        "!pip install langchain-community"
      ],
      "metadata": {
        "id": "2evPp14fy258",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "794909d6-9d36-4e1a-ecc3-e23c5ee08257"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting langchain\n",
            "  Downloading langchain-0.2.11-py3-none-any.whl.metadata (7.1 kB)\n",
            "Requirement already satisfied: PyYAML>=5.3 in /usr/local/lib/python3.10/dist-packages (from langchain) (6.0.1)\n",
            "Requirement already satisfied: SQLAlchemy<3,>=1.4 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.0.31)\n",
            "Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /usr/local/lib/python3.10/dist-packages (from langchain) (3.9.5)\n",
            "Requirement already satisfied: async-timeout<5.0.0,>=4.0.0 in /usr/local/lib/python3.10/dist-packages (from langchain) (4.0.3)\n",
            "Collecting langchain-core<0.3.0,>=0.2.23 (from langchain)\n",
            "  Downloading langchain_core-0.2.24-py3-none-any.whl.metadata (6.2 kB)\n",
            "Collecting langchain-text-splitters<0.3.0,>=0.2.0 (from langchain)\n",
            "  Downloading langchain_text_splitters-0.2.2-py3-none-any.whl.metadata (2.1 kB)\n",
            "Collecting langsmith<0.2.0,>=0.1.17 (from langchain)\n",
            "  Downloading langsmith-0.1.93-py3-none-any.whl.metadata (13 kB)\n",
            "Requirement already satisfied: numpy<2,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain) (1.25.2)\n",
            "Requirement already satisfied: pydantic<3,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.8.2)\n",
            "Requirement already satisfied: requests<3,>=2 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.31.0)\n",
            "Requirement already satisfied: tenacity!=8.4.0,<9.0.0,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from langchain) (8.5.0)\n",
            "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.3.1)\n",
            "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (23.2.0)\n",
            "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.4.1)\n",
            "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (6.0.5)\n",
            "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.9.4)\n",
            "Collecting jsonpatch<2.0,>=1.33 (from langchain-core<0.3.0,>=0.2.23->langchain)\n",
            "  Downloading jsonpatch-1.33-py2.py3-none-any.whl.metadata (3.0 kB)\n",
            "Requirement already satisfied: packaging<25,>=23.2 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.23->langchain) (24.1)\n",
            "Collecting orjson<4.0.0,>=3.9.14 (from langsmith<0.2.0,>=0.1.17->langchain)\n",
            "  Downloading orjson-3.10.6-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (50 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m50.4/50.4 kB\u001b[0m \u001b[31m986.5 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: annotated-types>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain) (0.7.0)\n",
            "Requirement already satisfied: pydantic-core==2.20.1 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain) (2.20.1)\n",
            "Requirement already satisfied: typing-extensions>=4.6.1 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain) (4.12.2)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (3.7)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (2024.7.4)\n",
            "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.10/dist-packages (from SQLAlchemy<3,>=1.4->langchain) (3.0.3)\n",
            "Collecting jsonpointer>=1.9 (from jsonpatch<2.0,>=1.33->langchain-core<0.3.0,>=0.2.23->langchain)\n",
            "  Downloading jsonpointer-3.0.0-py2.py3-none-any.whl.metadata (2.3 kB)\n",
            "Downloading langchain-0.2.11-py3-none-any.whl (990 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m990.3/990.3 kB\u001b[0m \u001b[31m20.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading langchain_core-0.2.24-py3-none-any.whl (377 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m377.3/377.3 kB\u001b[0m \u001b[31m24.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading langchain_text_splitters-0.2.2-py3-none-any.whl (25 kB)\n",
            "Downloading langsmith-0.1.93-py3-none-any.whl (139 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m139.8/139.8 kB\u001b[0m \u001b[31m10.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading jsonpatch-1.33-py2.py3-none-any.whl (12 kB)\n",
            "Downloading orjson-3.10.6-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (141 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m141.1/141.1 kB\u001b[0m \u001b[31m9.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading jsonpointer-3.0.0-py2.py3-none-any.whl (7.6 kB)\n",
            "Installing collected packages: orjson, jsonpointer, jsonpatch, langsmith, langchain-core, langchain-text-splitters, langchain\n",
            "Successfully installed jsonpatch-1.33 jsonpointer-3.0.0 langchain-0.2.11 langchain-core-0.2.24 langchain-text-splitters-0.2.2 langsmith-0.1.93 orjson-3.10.6\n",
            "Collecting langchain-openai\n",
            "  Downloading langchain_openai-0.1.19-py3-none-any.whl.metadata (2.6 kB)\n",
            "Requirement already satisfied: langchain-core<0.3.0,>=0.2.24 in /usr/local/lib/python3.10/dist-packages (from langchain-openai) (0.2.24)\n",
            "Collecting openai<2.0.0,>=1.32.0 (from langchain-openai)\n",
            "  Downloading openai-1.37.1-py3-none-any.whl.metadata (22 kB)\n",
            "Collecting tiktoken<1,>=0.7 (from langchain-openai)\n",
            "  Downloading tiktoken-0.7.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.6 kB)\n",
            "Requirement already satisfied: PyYAML>=5.3 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (6.0.1)\n",
            "Requirement already satisfied: jsonpatch<2.0,>=1.33 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (1.33)\n",
            "Requirement already satisfied: langsmith<0.2.0,>=0.1.75 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (0.1.93)\n",
            "Requirement already satisfied: packaging<25,>=23.2 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (24.1)\n",
            "Requirement already satisfied: pydantic<3,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (2.8.2)\n",
            "Requirement already satisfied: tenacity!=8.4.0,<9.0.0,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.24->langchain-openai) (8.5.0)\n",
            "Requirement already satisfied: anyio<5,>=3.5.0 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.32.0->langchain-openai) (3.7.1)\n",
            "Requirement already satisfied: distro<2,>=1.7.0 in /usr/lib/python3/dist-packages (from openai<2.0.0,>=1.32.0->langchain-openai) (1.7.0)\n",
            "Collecting httpx<1,>=0.23.0 (from openai<2.0.0,>=1.32.0->langchain-openai)\n",
            "  Downloading httpx-0.27.0-py3-none-any.whl.metadata (7.2 kB)\n",
            "Requirement already satisfied: sniffio in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.32.0->langchain-openai) (1.3.1)\n",
            "Requirement already satisfied: tqdm>4 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.32.0->langchain-openai) (4.66.4)\n",
            "Requirement already satisfied: typing-extensions<5,>=4.7 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.32.0->langchain-openai) (4.12.2)\n",
            "Requirement already satisfied: regex>=2022.1.18 in /usr/local/lib/python3.10/dist-packages (from tiktoken<1,>=0.7->langchain-openai) (2024.5.15)\n",
            "Requirement already satisfied: requests>=2.26.0 in /usr/local/lib/python3.10/dist-packages (from tiktoken<1,>=0.7->langchain-openai) (2.31.0)\n",
            "Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.10/dist-packages (from anyio<5,>=3.5.0->openai<2.0.0,>=1.32.0->langchain-openai) (3.7)\n",
            "Requirement already satisfied: exceptiongroup in /usr/local/lib/python3.10/dist-packages (from anyio<5,>=3.5.0->openai<2.0.0,>=1.32.0->langchain-openai) (1.2.2)\n",
            "Requirement already satisfied: certifi in /usr/local/lib/python3.10/dist-packages (from httpx<1,>=0.23.0->openai<2.0.0,>=1.32.0->langchain-openai) (2024.7.4)\n",
            "Collecting httpcore==1.* (from httpx<1,>=0.23.0->openai<2.0.0,>=1.32.0->langchain-openai)\n",
            "  Downloading httpcore-1.0.5-py3-none-any.whl.metadata (20 kB)\n",
            "Collecting h11<0.15,>=0.13 (from httpcore==1.*->httpx<1,>=0.23.0->openai<2.0.0,>=1.32.0->langchain-openai)\n",
            "  Downloading h11-0.14.0-py3-none-any.whl.metadata (8.2 kB)\n",
            "Requirement already satisfied: jsonpointer>=1.9 in /usr/local/lib/python3.10/dist-packages (from jsonpatch<2.0,>=1.33->langchain-core<0.3.0,>=0.2.24->langchain-openai) (3.0.0)\n",
            "Requirement already satisfied: orjson<4.0.0,>=3.9.14 in /usr/local/lib/python3.10/dist-packages (from langsmith<0.2.0,>=0.1.75->langchain-core<0.3.0,>=0.2.24->langchain-openai) (3.10.6)\n",
            "Requirement already satisfied: annotated-types>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain-core<0.3.0,>=0.2.24->langchain-openai) (0.7.0)\n",
            "Requirement already satisfied: pydantic-core==2.20.1 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain-core<0.3.0,>=0.2.24->langchain-openai) (2.20.1)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests>=2.26.0->tiktoken<1,>=0.7->langchain-openai) (3.3.2)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests>=2.26.0->tiktoken<1,>=0.7->langchain-openai) (2.0.7)\n",
            "Downloading langchain_openai-0.1.19-py3-none-any.whl (47 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m47.1/47.1 kB\u001b[0m \u001b[31m3.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading openai-1.37.1-py3-none-any.whl (337 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m337.0/337.0 kB\u001b[0m \u001b[31m17.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading tiktoken-0.7.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.1/1.1 MB\u001b[0m \u001b[31m44.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading httpx-0.27.0-py3-none-any.whl (75 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m75.6/75.6 kB\u001b[0m \u001b[31m5.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading httpcore-1.0.5-py3-none-any.whl (77 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m77.9/77.9 kB\u001b[0m \u001b[31m7.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading h11-0.14.0-py3-none-any.whl (58 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.3/58.3 kB\u001b[0m \u001b[31m5.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: h11, tiktoken, httpcore, httpx, openai, langchain-openai\n",
            "Successfully installed h11-0.14.0 httpcore-1.0.5 httpx-0.27.0 langchain-openai-0.1.19 openai-1.37.1 tiktoken-0.7.0\n",
            "Collecting langchain-community\n",
            "  Downloading langchain_community-0.2.10-py3-none-any.whl.metadata (2.7 kB)\n",
            "Requirement already satisfied: PyYAML>=5.3 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (6.0.1)\n",
            "Requirement already satisfied: SQLAlchemy<3,>=1.4 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (2.0.31)\n",
            "Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (3.9.5)\n",
            "Collecting dataclasses-json<0.7,>=0.5.7 (from langchain-community)\n",
            "  Downloading dataclasses_json-0.6.7-py3-none-any.whl.metadata (25 kB)\n",
            "Requirement already satisfied: langchain<0.3.0,>=0.2.9 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (0.2.11)\n",
            "Requirement already satisfied: langchain-core<0.3.0,>=0.2.23 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (0.2.24)\n",
            "Requirement already satisfied: langsmith<0.2.0,>=0.1.0 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (0.1.93)\n",
            "Requirement already satisfied: numpy<2,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (1.25.2)\n",
            "Requirement already satisfied: requests<3,>=2 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (2.31.0)\n",
            "Requirement already satisfied: tenacity!=8.4.0,<9.0.0,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from langchain-community) (8.5.0)\n",
            "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (1.3.1)\n",
            "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (23.2.0)\n",
            "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (1.4.1)\n",
            "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (6.0.5)\n",
            "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (1.9.4)\n",
            "Requirement already satisfied: async-timeout<5.0,>=4.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community) (4.0.3)\n",
            "Collecting marshmallow<4.0.0,>=3.18.0 (from dataclasses-json<0.7,>=0.5.7->langchain-community)\n",
            "  Downloading marshmallow-3.21.3-py3-none-any.whl.metadata (7.1 kB)\n",
            "Collecting typing-inspect<1,>=0.4.0 (from dataclasses-json<0.7,>=0.5.7->langchain-community)\n",
            "  Downloading typing_inspect-0.9.0-py3-none-any.whl.metadata (1.5 kB)\n",
            "Requirement already satisfied: langchain-text-splitters<0.3.0,>=0.2.0 in /usr/local/lib/python3.10/dist-packages (from langchain<0.3.0,>=0.2.9->langchain-community) (0.2.2)\n",
            "Requirement already satisfied: pydantic<3,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain<0.3.0,>=0.2.9->langchain-community) (2.8.2)\n",
            "Requirement already satisfied: jsonpatch<2.0,>=1.33 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.23->langchain-community) (1.33)\n",
            "Requirement already satisfied: packaging<25,>=23.2 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.3.0,>=0.2.23->langchain-community) (24.1)\n",
            "Requirement already satisfied: orjson<4.0.0,>=3.9.14 in /usr/local/lib/python3.10/dist-packages (from langsmith<0.2.0,>=0.1.0->langchain-community) (3.10.6)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community) (3.7)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community) (2024.7.4)\n",
            "Requirement already satisfied: typing-extensions>=4.6.0 in /usr/local/lib/python3.10/dist-packages (from SQLAlchemy<3,>=1.4->langchain-community) (4.12.2)\n",
            "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.10/dist-packages (from SQLAlchemy<3,>=1.4->langchain-community) (3.0.3)\n",
            "Requirement already satisfied: jsonpointer>=1.9 in /usr/local/lib/python3.10/dist-packages (from jsonpatch<2.0,>=1.33->langchain-core<0.3.0,>=0.2.23->langchain-community) (3.0.0)\n",
            "Requirement already satisfied: annotated-types>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain<0.3.0,>=0.2.9->langchain-community) (0.7.0)\n",
            "Requirement already satisfied: pydantic-core==2.20.1 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1->langchain<0.3.0,>=0.2.9->langchain-community) (2.20.1)\n",
            "Collecting mypy-extensions>=0.3.0 (from typing-inspect<1,>=0.4.0->dataclasses-json<0.7,>=0.5.7->langchain-community)\n",
            "  Downloading mypy_extensions-1.0.0-py3-none-any.whl.metadata (1.1 kB)\n",
            "Downloading langchain_community-0.2.10-py3-none-any.whl (2.3 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.3/2.3 MB\u001b[0m \u001b[31m35.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading dataclasses_json-0.6.7-py3-none-any.whl (28 kB)\n",
            "Downloading marshmallow-3.21.3-py3-none-any.whl (49 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m49.2/49.2 kB\u001b[0m \u001b[31m3.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading typing_inspect-0.9.0-py3-none-any.whl (8.8 kB)\n",
            "Downloading mypy_extensions-1.0.0-py3-none-any.whl (4.7 kB)\n",
            "Installing collected packages: mypy-extensions, marshmallow, typing-inspect, dataclasses-json, langchain-community\n",
            "Successfully installed dataclasses-json-0.6.7 langchain-community-0.2.10 marshmallow-3.21.3 mypy-extensions-1.0.0 typing-inspect-0.9.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# takes 2 - 5 mins to install on Colab\n",
        "!pip install \"unstructured[all-docs]\""
      ],
      "metadata": {
        "id": "CB6lHzbz5a10",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "6bdb528c-ae6f-4177-c893-77b923a69041"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting unstructured[all-docs]\n",
            "  Downloading unstructured-0.15.0-py3-none-any.whl.metadata (29 kB)\n",
            "Requirement already satisfied: chardet in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (5.2.0)\n",
            "Collecting filetype (from unstructured[all-docs])\n",
            "  Downloading filetype-1.2.0-py2.py3-none-any.whl.metadata (6.5 kB)\n",
            "Collecting python-magic (from unstructured[all-docs])\n",
            "  Downloading python_magic-0.4.27-py2.py3-none-any.whl.metadata (5.8 kB)\n",
            "Requirement already satisfied: lxml in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (4.9.4)\n",
            "Requirement already satisfied: nltk in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (3.8.1)\n",
            "Requirement already satisfied: tabulate in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (0.9.0)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (2.31.0)\n",
            "Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (4.12.3)\n",
            "Collecting emoji (from unstructured[all-docs])\n",
            "  Downloading emoji-2.12.1-py3-none-any.whl.metadata (5.4 kB)\n",
            "Requirement already satisfied: dataclasses-json in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (0.6.7)\n",
            "Collecting python-iso639 (from unstructured[all-docs])\n",
            "  Downloading python_iso639-2024.4.27-py3-none-any.whl.metadata (13 kB)\n",
            "Collecting langdetect (from unstructured[all-docs])\n",
            "  Downloading langdetect-1.0.9.tar.gz (981 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m981.5/981.5 kB\u001b[0m \u001b[31m29.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: numpy<2 in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (1.25.2)\n",
            "Collecting rapidfuzz (from unstructured[all-docs])\n",
            "  Downloading rapidfuzz-3.9.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (12 kB)\n",
            "Collecting backoff (from unstructured[all-docs])\n",
            "  Downloading backoff-2.2.1-py3-none-any.whl.metadata (14 kB)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (4.12.2)\n",
            "Collecting unstructured-client (from unstructured[all-docs])\n",
            "  Downloading unstructured_client-0.24.1-py3-none-any.whl.metadata (12 kB)\n",
            "Requirement already satisfied: wrapt in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (1.14.1)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (4.66.4)\n",
            "Requirement already satisfied: psutil in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (5.9.5)\n",
            "Collecting pillow-heif (from unstructured[all-docs])\n",
            "  Downloading pillow_heif-0.17.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (9.9 kB)\n",
            "Requirement already satisfied: openpyxl in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (3.1.5)\n",
            "Collecting google-cloud-vision (from unstructured[all-docs])\n",
            "  Downloading google_cloud_vision-3.7.3-py2.py3-none-any.whl.metadata (5.2 kB)\n",
            "Collecting unstructured-inference==0.7.36 (from unstructured[all-docs])\n",
            "  Downloading unstructured_inference-0.7.36-py3-none-any.whl.metadata (5.9 kB)\n",
            "Collecting python-docx>=1.1.2 (from unstructured[all-docs])\n",
            "  Downloading python_docx-1.1.2-py3-none-any.whl.metadata (2.0 kB)\n",
            "Collecting pikepdf (from unstructured[all-docs])\n",
            "  Downloading pikepdf-9.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (8.2 kB)\n",
            "Requirement already satisfied: xlrd in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (2.0.1)\n",
            "Collecting python-pptx<=0.6.23 (from unstructured[all-docs])\n",
            "  Downloading python_pptx-0.6.23-py3-none-any.whl.metadata (18 kB)\n",
            "Collecting effdet (from unstructured[all-docs])\n",
            "  Downloading effdet-0.4.1-py3-none-any.whl.metadata (33 kB)\n",
            "Collecting pypandoc (from unstructured[all-docs])\n",
            "  Downloading pypandoc-1.13-py3-none-any.whl.metadata (16 kB)\n",
            "Collecting pypdf (from unstructured[all-docs])\n",
            "  Downloading pypdf-4.3.1-py3-none-any.whl.metadata (7.4 kB)\n",
            "Collecting pytesseract (from unstructured[all-docs])\n",
            "  Downloading pytesseract-0.3.10-py3-none-any.whl.metadata (11 kB)\n",
            "Collecting onnx (from unstructured[all-docs])\n",
            "  Downloading onnx-1.16.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (16 kB)\n",
            "Collecting python-oxmsg (from unstructured[all-docs])\n",
            "  Downloading python_oxmsg-0.0.1-py3-none-any.whl.metadata (5.0 kB)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (3.3)\n",
            "Requirement already satisfied: markdown in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (3.6)\n",
            "Collecting pdf2image (from unstructured[all-docs])\n",
            "  Downloading pdf2image-1.17.0-py3-none-any.whl.metadata (6.2 kB)\n",
            "Collecting unstructured.pytesseract>=0.3.12 (from unstructured[all-docs])\n",
            "  Downloading unstructured.pytesseract-0.3.12-py3-none-any.whl.metadata (11 kB)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.10/dist-packages (from unstructured[all-docs]) (2.0.3)\n",
            "Collecting pdfminer.six (from unstructured[all-docs])\n",
            "  Downloading pdfminer.six-20240706-py3-none-any.whl.metadata (4.1 kB)\n",
            "Collecting layoutparser (from unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading layoutparser-0.3.4-py3-none-any.whl.metadata (7.7 kB)\n",
            "Collecting python-multipart (from unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading python_multipart-0.0.9-py3-none-any.whl.metadata (2.5 kB)\n",
            "Requirement already satisfied: huggingface-hub in /usr/local/lib/python3.10/dist-packages (from unstructured-inference==0.7.36->unstructured[all-docs]) (0.23.5)\n",
            "Requirement already satisfied: opencv-python!=4.7.0.68 in /usr/local/lib/python3.10/dist-packages (from unstructured-inference==0.7.36->unstructured[all-docs]) (4.10.0.84)\n",
            "Collecting onnxruntime>=1.17.0 (from unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading onnxruntime-1.18.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (4.3 kB)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (from unstructured-inference==0.7.36->unstructured[all-docs]) (3.7.1)\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (from unstructured-inference==0.7.36->unstructured[all-docs]) (2.3.1+cu121)\n",
            "Collecting timm (from unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading timm-1.0.7-py3-none-any.whl.metadata (47 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m47.5/47.5 kB\u001b[0m \u001b[31m3.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: transformers>=4.25.1 in /usr/local/lib/python3.10/dist-packages (from unstructured-inference==0.7.36->unstructured[all-docs]) (4.42.4)\n",
            "Requirement already satisfied: Pillow>=3.3.2 in /usr/local/lib/python3.10/dist-packages (from python-pptx<=0.6.23->unstructured[all-docs]) (9.4.0)\n",
            "Collecting XlsxWriter>=0.5.7 (from python-pptx<=0.6.23->unstructured[all-docs])\n",
            "  Downloading XlsxWriter-3.2.0-py3-none-any.whl.metadata (2.6 kB)\n",
            "Requirement already satisfied: packaging>=21.3 in /usr/local/lib/python3.10/dist-packages (from unstructured.pytesseract>=0.3.12->unstructured[all-docs]) (24.1)\n",
            "Requirement already satisfied: soupsieve>1.2 in /usr/local/lib/python3.10/dist-packages (from beautifulsoup4->unstructured[all-docs]) (2.5)\n",
            "Requirement already satisfied: marshmallow<4.0.0,>=3.18.0 in /usr/local/lib/python3.10/dist-packages (from dataclasses-json->unstructured[all-docs]) (3.21.3)\n",
            "Requirement already satisfied: typing-inspect<1,>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from dataclasses-json->unstructured[all-docs]) (0.9.0)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.10/dist-packages (from effdet->unstructured[all-docs]) (0.18.1+cu121)\n",
            "Requirement already satisfied: pycocotools>=2.0.2 in /usr/local/lib/python3.10/dist-packages (from effdet->unstructured[all-docs]) (2.0.8)\n",
            "Collecting omegaconf>=2.0 (from effdet->unstructured[all-docs])\n",
            "  Downloading omegaconf-2.3.0-py3-none-any.whl.metadata (3.9 kB)\n",
            "Requirement already satisfied: google-api-core!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1 in /usr/local/lib/python3.10/dist-packages (from google-api-core[grpc]!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1->google-cloud-vision->unstructured[all-docs]) (2.19.1)\n",
            "Requirement already satisfied: google-auth!=2.24.0,!=2.25.0,<3.0.0dev,>=2.14.1 in /usr/local/lib/python3.10/dist-packages (from google-cloud-vision->unstructured[all-docs]) (2.27.0)\n",
            "Requirement already satisfied: proto-plus<2.0.0dev,>=1.22.3 in /usr/local/lib/python3.10/dist-packages (from google-cloud-vision->unstructured[all-docs]) (1.24.0)\n",
            "Requirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<6.0.0dev,>=3.20.2 in /usr/local/lib/python3.10/dist-packages (from google-cloud-vision->unstructured[all-docs]) (3.20.3)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.10/dist-packages (from langdetect->unstructured[all-docs]) (1.16.0)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.10/dist-packages (from nltk->unstructured[all-docs]) (8.1.7)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.10/dist-packages (from nltk->unstructured[all-docs]) (1.4.2)\n",
            "Requirement already satisfied: regex>=2021.8.3 in /usr/local/lib/python3.10/dist-packages (from nltk->unstructured[all-docs]) (2024.5.15)\n",
            "Requirement already satisfied: et-xmlfile in /usr/local/lib/python3.10/dist-packages (from openpyxl->unstructured[all-docs]) (1.1.0)\n",
            "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/dist-packages (from pandas->unstructured[all-docs]) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas->unstructured[all-docs]) (2024.1)\n",
            "Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas->unstructured[all-docs]) (2024.1)\n",
            "Requirement already satisfied: charset-normalizer>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from pdfminer.six->unstructured[all-docs]) (3.3.2)\n",
            "Requirement already satisfied: cryptography>=36.0.0 in /usr/local/lib/python3.10/dist-packages (from pdfminer.six->unstructured[all-docs]) (42.0.8)\n",
            "Collecting Pillow>=3.3.2 (from python-pptx<=0.6.23->unstructured[all-docs])\n",
            "  Downloading pillow-10.4.0-cp310-cp310-manylinux_2_28_x86_64.whl.metadata (9.2 kB)\n",
            "Collecting Deprecated (from pikepdf->unstructured[all-docs])\n",
            "  Downloading Deprecated-1.2.14-py2.py3-none-any.whl.metadata (5.4 kB)\n",
            "Collecting olefile (from python-oxmsg->unstructured[all-docs])\n",
            "  Downloading olefile-0.47-py2.py3-none-any.whl.metadata (9.7 kB)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->unstructured[all-docs]) (3.7)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->unstructured[all-docs]) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->unstructured[all-docs]) (2024.7.4)\n",
            "Collecting deepdiff>=6.0 (from unstructured-client->unstructured[all-docs])\n",
            "  Downloading deepdiff-7.0.1-py3-none-any.whl.metadata (6.8 kB)\n",
            "Requirement already satisfied: httpx>=0.27.0 in /usr/local/lib/python3.10/dist-packages (from unstructured-client->unstructured[all-docs]) (0.27.0)\n",
            "Collecting jsonpath-python>=1.0.6 (from unstructured-client->unstructured[all-docs])\n",
            "  Downloading jsonpath_python-1.0.6-py3-none-any.whl.metadata (12 kB)\n",
            "Requirement already satisfied: mypy-extensions>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from unstructured-client->unstructured[all-docs]) (1.0.0)\n",
            "Requirement already satisfied: nest-asyncio>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from unstructured-client->unstructured[all-docs]) (1.6.0)\n",
            "Collecting requests-toolbelt>=1.0.0 (from unstructured-client->unstructured[all-docs])\n",
            "  Downloading requests_toolbelt-1.0.0-py2.py3-none-any.whl.metadata (14 kB)\n",
            "Requirement already satisfied: cffi>=1.12 in /usr/local/lib/python3.10/dist-packages (from cryptography>=36.0.0->pdfminer.six->unstructured[all-docs]) (1.16.0)\n",
            "Collecting ordered-set<4.2.0,>=4.1.0 (from deepdiff>=6.0->unstructured-client->unstructured[all-docs])\n",
            "  Downloading ordered_set-4.1.0-py3-none-any.whl.metadata (5.3 kB)\n",
            "Requirement already satisfied: googleapis-common-protos<2.0.dev0,>=1.56.2 in /usr/local/lib/python3.10/dist-packages (from google-api-core!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1->google-api-core[grpc]!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1->google-cloud-vision->unstructured[all-docs]) (1.63.2)\n",
            "Requirement already satisfied: grpcio<2.0dev,>=1.33.2 in /usr/local/lib/python3.10/dist-packages (from google-api-core[grpc]!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1->google-cloud-vision->unstructured[all-docs]) (1.64.1)\n",
            "Requirement already satisfied: grpcio-status<2.0.dev0,>=1.33.2 in /usr/local/lib/python3.10/dist-packages (from google-api-core[grpc]!=2.0.*,!=2.1.*,!=2.10.*,!=2.2.*,!=2.3.*,!=2.4.*,!=2.5.*,!=2.6.*,!=2.7.*,!=2.8.*,!=2.9.*,<3.0.0dev,>=1.34.1->google-cloud-vision->unstructured[all-docs]) (1.48.2)\n",
            "Requirement already satisfied: cachetools<6.0,>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from google-auth!=2.24.0,!=2.25.0,<3.0.0dev,>=2.14.1->google-cloud-vision->unstructured[all-docs]) (5.4.0)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from google-auth!=2.24.0,!=2.25.0,<3.0.0dev,>=2.14.1->google-cloud-vision->unstructured[all-docs]) (0.4.0)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.10/dist-packages (from google-auth!=2.24.0,!=2.25.0,<3.0.0dev,>=2.14.1->google-cloud-vision->unstructured[all-docs]) (4.9)\n",
            "Requirement already satisfied: anyio in /usr/local/lib/python3.10/dist-packages (from httpx>=0.27.0->unstructured-client->unstructured[all-docs]) (3.7.1)\n",
            "Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.10/dist-packages (from httpx>=0.27.0->unstructured-client->unstructured[all-docs]) (1.0.5)\n",
            "Requirement already satisfied: sniffio in /usr/local/lib/python3.10/dist-packages (from httpx>=0.27.0->unstructured-client->unstructured[all-docs]) (1.3.1)\n",
            "Requirement already satisfied: h11<0.15,>=0.13 in /usr/local/lib/python3.10/dist-packages (from httpcore==1.*->httpx>=0.27.0->unstructured-client->unstructured[all-docs]) (0.14.0)\n",
            "Collecting antlr4-python3-runtime==4.9.* (from omegaconf>=2.0->effdet->unstructured[all-docs])\n",
            "  Downloading antlr4-python3-runtime-4.9.3.tar.gz (117 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m117.0/117.0 kB\u001b[0m \u001b[31m9.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: PyYAML>=5.1.0 in /usr/local/lib/python3.10/dist-packages (from omegaconf>=2.0->effdet->unstructured[all-docs]) (6.0.1)\n",
            "Collecting coloredlogs (from onnxruntime>=1.17.0->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading coloredlogs-15.0.1-py2.py3-none-any.whl.metadata (12 kB)\n",
            "Requirement already satisfied: flatbuffers in /usr/local/lib/python3.10/dist-packages (from onnxruntime>=1.17.0->unstructured-inference==0.7.36->unstructured[all-docs]) (24.3.25)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from onnxruntime>=1.17.0->unstructured-inference==0.7.36->unstructured[all-docs]) (1.13.1)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->unstructured-inference==0.7.36->unstructured[all-docs]) (1.2.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib->unstructured-inference==0.7.36->unstructured[all-docs]) (0.12.1)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->unstructured-inference==0.7.36->unstructured[all-docs]) (4.53.1)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->unstructured-inference==0.7.36->unstructured[all-docs]) (1.4.5)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->unstructured-inference==0.7.36->unstructured[all-docs]) (3.1.2)\n",
            "Requirement already satisfied: safetensors in /usr/local/lib/python3.10/dist-packages (from timm->unstructured-inference==0.7.36->unstructured[all-docs]) (0.4.3)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch->unstructured-inference==0.7.36->unstructured[all-docs]) (3.15.4)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch->unstructured-inference==0.7.36->unstructured[all-docs]) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch->unstructured-inference==0.7.36->unstructured[all-docs]) (2024.6.1)\n",
            "Collecting nvidia-cuda-nvrtc-cu12==12.1.105 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cuda_nvrtc_cu12-12.1.105-py3-none-manylinux1_x86_64.whl.metadata (1.5 kB)\n",
            "Collecting nvidia-cuda-runtime-cu12==12.1.105 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cuda_runtime_cu12-12.1.105-py3-none-manylinux1_x86_64.whl.metadata (1.5 kB)\n",
            "Collecting nvidia-cuda-cupti-cu12==12.1.105 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cuda_cupti_cu12-12.1.105-py3-none-manylinux1_x86_64.whl.metadata (1.6 kB)\n",
            "Collecting nvidia-cudnn-cu12==8.9.2.26 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cudnn_cu12-8.9.2.26-py3-none-manylinux1_x86_64.whl.metadata (1.6 kB)\n",
            "Collecting nvidia-cublas-cu12==12.1.3.1 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cublas_cu12-12.1.3.1-py3-none-manylinux1_x86_64.whl.metadata (1.5 kB)\n",
            "Collecting nvidia-cufft-cu12==11.0.2.54 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cufft_cu12-11.0.2.54-py3-none-manylinux1_x86_64.whl.metadata (1.5 kB)\n",
            "Collecting nvidia-curand-cu12==10.3.2.106 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_curand_cu12-10.3.2.106-py3-none-manylinux1_x86_64.whl.metadata (1.5 kB)\n",
            "Collecting nvidia-cusolver-cu12==11.4.5.107 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cusolver_cu12-11.4.5.107-py3-none-manylinux1_x86_64.whl.metadata (1.6 kB)\n",
            "Collecting nvidia-cusparse-cu12==12.1.0.106 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_cusparse_cu12-12.1.0.106-py3-none-manylinux1_x86_64.whl.metadata (1.6 kB)\n",
            "Collecting nvidia-nccl-cu12==2.20.5 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_nccl_cu12-2.20.5-py3-none-manylinux2014_x86_64.whl.metadata (1.8 kB)\n",
            "Collecting nvidia-nvtx-cu12==12.1.105 (from torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Using cached nvidia_nvtx_cu12-12.1.105-py3-none-manylinux1_x86_64.whl.metadata (1.7 kB)\n",
            "Requirement already satisfied: triton==2.3.1 in /usr/local/lib/python3.10/dist-packages (from torch->unstructured-inference==0.7.36->unstructured[all-docs]) (2.3.1)\n",
            "Collecting nvidia-nvjitlink-cu12 (from nvidia-cusolver-cu12==11.4.5.107->torch->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading nvidia_nvjitlink_cu12-12.5.82-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB)\n",
            "Requirement already satisfied: tokenizers<0.20,>=0.19 in /usr/local/lib/python3.10/dist-packages (from transformers>=4.25.1->unstructured-inference==0.7.36->unstructured[all-docs]) (0.19.1)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from layoutparser->unstructured-inference==0.7.36->unstructured[all-docs]) (1.13.1)\n",
            "Collecting iopath (from layoutparser->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading iopath-0.1.10.tar.gz (42 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m42.2/42.2 kB\u001b[0m \u001b[31m3.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting pdfplumber (from layoutparser->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading pdfplumber-0.11.2-py3-none-any.whl.metadata (40 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m40.1/40.1 kB\u001b[0m \u001b[31m3.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: pycparser in /usr/local/lib/python3.10/dist-packages (from cffi>=1.12->cryptography>=36.0.0->pdfminer.six->unstructured[all-docs]) (2.22)\n",
            "Requirement already satisfied: pyasn1<0.7.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth!=2.24.0,!=2.25.0,<3.0.0dev,>=2.14.1->google-cloud-vision->unstructured[all-docs]) (0.6.0)\n",
            "Requirement already satisfied: exceptiongroup in /usr/local/lib/python3.10/dist-packages (from anyio->httpx>=0.27.0->unstructured-client->unstructured[all-docs]) (1.2.2)\n",
            "Collecting humanfriendly>=9.1 (from coloredlogs->onnxruntime>=1.17.0->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading humanfriendly-10.0-py2.py3-none-any.whl.metadata (9.2 kB)\n",
            "Collecting portalocker (from iopath->layoutparser->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading portalocker-2.10.1-py3-none-any.whl.metadata (8.5 kB)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch->unstructured-inference==0.7.36->unstructured[all-docs]) (2.1.5)\n",
            "Collecting pdfminer.six (from unstructured[all-docs])\n",
            "  Downloading pdfminer.six-20231228-py3-none-any.whl.metadata (4.2 kB)\n",
            "Collecting pypdfium2>=4.18.0 (from pdfplumber->layoutparser->unstructured-inference==0.7.36->unstructured[all-docs])\n",
            "  Downloading pypdfium2-4.30.0-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (48 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m48.5/48.5 kB\u001b[0m \u001b[31m3.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy->onnxruntime>=1.17.0->unstructured-inference==0.7.36->unstructured[all-docs]) (1.3.0)\n",
            "Downloading unstructured_inference-0.7.36-py3-none-any.whl (56 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m56.4/56.4 kB\u001b[0m \u001b[31m3.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading python_docx-1.1.2-py3-none-any.whl (244 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m244.3/244.3 kB\u001b[0m \u001b[31m17.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading python_pptx-0.6.23-py3-none-any.whl (471 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m471.6/471.6 kB\u001b[0m \u001b[31m29.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading unstructured.pytesseract-0.3.12-py3-none-any.whl (14 kB)\n",
            "Downloading backoff-2.2.1-py3-none-any.whl (15 kB)\n",
            "Downloading effdet-0.4.1-py3-none-any.whl (112 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m112.5/112.5 kB\u001b[0m \u001b[31m9.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading emoji-2.12.1-py3-none-any.whl (431 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m431.4/431.4 kB\u001b[0m \u001b[31m22.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading filetype-1.2.0-py2.py3-none-any.whl (19 kB)\n",
            "Downloading google_cloud_vision-3.7.3-py2.py3-none-any.whl (466 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m466.4/466.4 kB\u001b[0m \u001b[31m31.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading onnx-1.16.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (15.9 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m15.9/15.9 MB\u001b[0m \u001b[31m42.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pdf2image-1.17.0-py3-none-any.whl (11 kB)\n",
            "Downloading pikepdf-9.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.4/2.4 MB\u001b[0m \u001b[31m40.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pillow_heif-0.17.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (7.5 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m7.5/7.5 MB\u001b[0m \u001b[31m75.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pypandoc-1.13-py3-none-any.whl (21 kB)\n",
            "Downloading pypdf-4.3.1-py3-none-any.whl (295 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m295.8/295.8 kB\u001b[0m \u001b[31m21.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pytesseract-0.3.10-py3-none-any.whl (14 kB)\n",
            "Downloading python_iso639-2024.4.27-py3-none-any.whl (274 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m274.7/274.7 kB\u001b[0m \u001b[31m21.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading python_magic-0.4.27-py2.py3-none-any.whl (13 kB)\n",
            "Downloading python_oxmsg-0.0.1-py3-none-any.whl (31 kB)\n",
            "Downloading rapidfuzz-3.9.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.4 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.4/3.4 MB\u001b[0m \u001b[31m41.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading unstructured-0.15.0-py3-none-any.whl (2.1 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.1/2.1 MB\u001b[0m \u001b[31m57.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading unstructured_client-0.24.1-py3-none-any.whl (41 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m41.9/41.9 kB\u001b[0m \u001b[31m3.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading deepdiff-7.0.1-py3-none-any.whl (80 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m80.8/80.8 kB\u001b[0m \u001b[31m7.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading jsonpath_python-1.0.6-py3-none-any.whl (7.6 kB)\n",
            "Downloading omegaconf-2.3.0-py3-none-any.whl (79 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m79.5/79.5 kB\u001b[0m \u001b[31m7.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading onnxruntime-1.18.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (6.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.8/6.8 MB\u001b[0m \u001b[31m52.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pillow-10.4.0-cp310-cp310-manylinux_2_28_x86_64.whl (4.5 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.5/4.5 MB\u001b[0m \u001b[31m61.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading requests_toolbelt-1.0.0-py2.py3-none-any.whl (54 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m54.5/54.5 kB\u001b[0m \u001b[31m4.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading timm-1.0.7-py3-none-any.whl (2.3 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.3/2.3 MB\u001b[0m \u001b[31m38.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hUsing cached nvidia_cublas_cu12-12.1.3.1-py3-none-manylinux1_x86_64.whl (410.6 MB)\n",
            "Using cached nvidia_cuda_cupti_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (14.1 MB)\n",
            "Using cached nvidia_cuda_nvrtc_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (23.7 MB)\n",
            "Using cached nvidia_cuda_runtime_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (823 kB)\n",
            "Using cached nvidia_cudnn_cu12-8.9.2.26-py3-none-manylinux1_x86_64.whl (731.7 MB)\n",
            "Using cached nvidia_cufft_cu12-11.0.2.54-py3-none-manylinux1_x86_64.whl (121.6 MB)\n",
            "Using cached nvidia_curand_cu12-10.3.2.106-py3-none-manylinux1_x86_64.whl (56.5 MB)\n",
            "Using cached nvidia_cusolver_cu12-11.4.5.107-py3-none-manylinux1_x86_64.whl (124.2 MB)\n",
            "Using cached nvidia_cusparse_cu12-12.1.0.106-py3-none-manylinux1_x86_64.whl (196.0 MB)\n",
            "Using cached nvidia_nccl_cu12-2.20.5-py3-none-manylinux2014_x86_64.whl (176.2 MB)\n",
            "Using cached nvidia_nvtx_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (99 kB)\n",
            "Downloading XlsxWriter-3.2.0-py3-none-any.whl (159 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m159.9/159.9 kB\u001b[0m \u001b[31m343.3 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading Deprecated-1.2.14-py2.py3-none-any.whl (9.6 kB)\n",
            "Downloading layoutparser-0.3.4-py3-none-any.whl (19.2 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m19.2/19.2 MB\u001b[0m \u001b[31m68.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading olefile-0.47-py2.py3-none-any.whl (114 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m114.6/114.6 kB\u001b[0m \u001b[31m7.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading python_multipart-0.0.9-py3-none-any.whl (22 kB)\n",
            "Downloading ordered_set-4.1.0-py3-none-any.whl (7.6 kB)\n",
            "Downloading coloredlogs-15.0.1-py2.py3-none-any.whl (46 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m46.0/46.0 kB\u001b[0m \u001b[31m3.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pdfplumber-0.11.2-py3-none-any.whl (58 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.0/58.0 kB\u001b[0m \u001b[31m5.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pdfminer.six-20231228-py3-none-any.whl (5.6 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.6/5.6 MB\u001b[0m \u001b[31m71.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading humanfriendly-10.0-py2.py3-none-any.whl (86 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m86.8/86.8 kB\u001b[0m \u001b[31m6.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading pypdfium2-4.30.0-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.8/2.8 MB\u001b[0m \u001b[31m52.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading nvidia_nvjitlink_cu12-12.5.82-py3-none-manylinux2014_x86_64.whl (21.3 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m21.3/21.3 MB\u001b[0m \u001b[31m54.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading portalocker-2.10.1-py3-none-any.whl (18 kB)\n",
            "Building wheels for collected packages: langdetect, antlr4-python3-runtime, iopath\n",
            "  Building wheel for langdetect (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for langdetect: filename=langdetect-1.0.9-py3-none-any.whl size=993221 sha256=44f5c128e8e41877e90ddef48fa056fc55044b4f3ee4a9f890d5a31b9caa5789\n",
            "  Stored in directory: /root/.cache/pip/wheels/95/03/7d/59ea870c70ce4e5a370638b5462a7711ab78fba2f655d05106\n",
            "  Building wheel for antlr4-python3-runtime (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for antlr4-python3-runtime: filename=antlr4_python3_runtime-4.9.3-py3-none-any.whl size=144552 sha256=b83bf3353aba6731ebbafaea4fe0ab54ddc9c3b8d7fe2e0ae78940e438951055\n",
            "  Stored in directory: /root/.cache/pip/wheels/12/93/dd/1f6a127edc45659556564c5730f6d4e300888f4bca2d4c5a88\n",
            "  Building wheel for iopath (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for iopath: filename=iopath-0.1.10-py3-none-any.whl size=31531 sha256=befb8f4ae33e80f5526d786982a0885a9eb6672b92c270d101c5895aaa734b97\n",
            "  Stored in directory: /root/.cache/pip/wheels/9a/a3/b6/ac0fcd1b4ed5cfeb3db92e6a0e476cfd48ed0df92b91080c1d\n",
            "Successfully built langdetect antlr4-python3-runtime iopath\n",
            "Installing collected packages: filetype, antlr4-python3-runtime, XlsxWriter, rapidfuzz, python-multipart, python-magic, python-iso639, python-docx, pypdfium2, pypdf, pypandoc, portalocker, Pillow, ordered-set, onnx, omegaconf, olefile, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufft-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, langdetect, jsonpath-python, humanfriendly, emoji, Deprecated, backoff, unstructured.pytesseract, requests-toolbelt, python-pptx, python-oxmsg, pytesseract, pillow-heif, pikepdf, pdf2image, nvidia-cusparse-cu12, nvidia-cudnn-cu12, iopath, deepdiff, coloredlogs, unstructured-client, pdfminer.six, onnxruntime, nvidia-cusolver-cu12, unstructured, pdfplumber, layoutparser, google-cloud-vision, timm, unstructured-inference, effdet\n",
            "  Attempting uninstall: Pillow\n",
            "    Found existing installation: Pillow 9.4.0\n",
            "    Uninstalling Pillow-9.4.0:\n",
            "      Successfully uninstalled Pillow-9.4.0\n",
            "Successfully installed Deprecated-1.2.14 Pillow-10.4.0 XlsxWriter-3.2.0 antlr4-python3-runtime-4.9.3 backoff-2.2.1 coloredlogs-15.0.1 deepdiff-7.0.1 effdet-0.4.1 emoji-2.12.1 filetype-1.2.0 google-cloud-vision-3.7.3 humanfriendly-10.0 iopath-0.1.10 jsonpath-python-1.0.6 langdetect-1.0.9 layoutparser-0.3.4 nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-nccl-cu12-2.20.5 nvidia-nvjitlink-cu12-12.5.82 nvidia-nvtx-cu12-12.1.105 olefile-0.47 omegaconf-2.3.0 onnx-1.16.1 onnxruntime-1.18.1 ordered-set-4.1.0 pdf2image-1.17.0 pdfminer.six-20231228 pdfplumber-0.11.2 pikepdf-9.1.0 pillow-heif-0.17.0 portalocker-2.10.1 pypandoc-1.13 pypdf-4.3.1 pypdfium2-4.30.0 pytesseract-0.3.10 python-docx-1.1.2 python-iso639-2024.4.27 python-magic-0.4.27 python-multipart-0.0.9 python-oxmsg-0.0.1 python-pptx-0.6.23 rapidfuzz-3.9.4 requests-toolbelt-1.0.0 timm-1.0.7 unstructured-0.15.0 unstructured-client-0.24.1 unstructured-inference-0.7.36 unstructured.pytesseract-0.3.12\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "PIL",
                  "google",
                  "pydevd_plugins"
                ]
              },
              "id": "0dbc6b41450f4a14b434c49e70ed0999"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# install OCR dependencies for unstructured\n",
        "!sudo apt-get install tesseract-ocr\n",
        "!sudo apt-get install poppler-utils"
      ],
      "metadata": {
        "id": "NhEW-tOywUgt",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "80dd00ff-8481-4d70-b796-01d669382ebd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Reading package lists... Done\n",
            "Building dependency tree... Done\n",
            "Reading state information... Done\n",
            "The following additional packages will be installed:\n",
            "  tesseract-ocr-eng tesseract-ocr-osd\n",
            "The following NEW packages will be installed:\n",
            "  tesseract-ocr tesseract-ocr-eng tesseract-ocr-osd\n",
            "0 upgraded, 3 newly installed, 0 to remove and 45 not upgraded.\n",
            "Need to get 4,816 kB of archives.\n",
            "After this operation, 15.6 MB of additional disk space will be used.\n",
            "Get:1 http://archive.ubuntu.com/ubuntu jammy/universe amd64 tesseract-ocr-eng all 1:4.00~git30-7274cfa-1.1 [1,591 kB]\n",
            "Get:2 http://archive.ubuntu.com/ubuntu jammy/universe amd64 tesseract-ocr-osd all 1:4.00~git30-7274cfa-1.1 [2,990 kB]\n",
            "Get:3 http://archive.ubuntu.com/ubuntu jammy/universe amd64 tesseract-ocr amd64 4.1.1-2.1build1 [236 kB]\n",
            "Fetched 4,816 kB in 1s (4,720 kB/s)\n",
            "debconf: unable to initialize frontend: Dialog\n",
            "debconf: (No usable dialog-like program is installed, so the dialog based frontend cannot be used. at /usr/share/perl5/Debconf/FrontEnd/Dialog.pm line 78, <> line 3.)\n",
            "debconf: falling back to frontend: Readline\n",
            "debconf: unable to initialize frontend: Readline\n",
            "debconf: (This frontend requires a controlling tty.)\n",
            "debconf: falling back to frontend: Teletype\n",
            "dpkg-preconfigure: unable to re-open stdin: \n",
            "Selecting previously unselected package tesseract-ocr-eng.\n",
            "(Reading database ... 123589 files and directories currently installed.)\n",
            "Preparing to unpack .../tesseract-ocr-eng_1%3a4.00~git30-7274cfa-1.1_all.deb ...\n",
            "Unpacking tesseract-ocr-eng (1:4.00~git30-7274cfa-1.1) ...\n",
            "Selecting previously unselected package tesseract-ocr-osd.\n",
            "Preparing to unpack .../tesseract-ocr-osd_1%3a4.00~git30-7274cfa-1.1_all.deb ...\n",
            "Unpacking tesseract-ocr-osd (1:4.00~git30-7274cfa-1.1) ...\n",
            "Selecting previously unselected package tesseract-ocr.\n",
            "Preparing to unpack .../tesseract-ocr_4.1.1-2.1build1_amd64.deb ...\n",
            "Unpacking tesseract-ocr (4.1.1-2.1build1) ...\n",
            "Setting up tesseract-ocr-eng (1:4.00~git30-7274cfa-1.1) ...\n",
            "Setting up tesseract-ocr-osd (1:4.00~git30-7274cfa-1.1) ...\n",
            "Setting up tesseract-ocr (4.1.1-2.1build1) ...\n",
            "Processing triggers for man-db (2.10.2-1) ...\n",
            "Reading package lists... Done\n",
            "Building dependency tree... Done\n",
            "Reading state information... Done\n",
            "The following NEW packages will be installed:\n",
            "  poppler-utils\n",
            "0 upgraded, 1 newly installed, 0 to remove and 45 not upgraded.\n",
            "Need to get 186 kB of archives.\n",
            "After this operation, 696 kB of additional disk space will be used.\n",
            "Get:1 http://archive.ubuntu.com/ubuntu jammy-updates/main amd64 poppler-utils amd64 22.02.0-2ubuntu0.5 [186 kB]\n",
            "Fetched 186 kB in 0s (397 kB/s)\n",
            "debconf: unable to initialize frontend: Dialog\n",
            "debconf: (No usable dialog-like program is installed, so the dialog based frontend cannot be used. at /usr/share/perl5/Debconf/FrontEnd/Dialog.pm line 78, <> line 1.)\n",
            "debconf: falling back to frontend: Readline\n",
            "debconf: unable to initialize frontend: Readline\n",
            "debconf: (This frontend requires a controlling tty.)\n",
            "debconf: falling back to frontend: Teletype\n",
            "dpkg-preconfigure: unable to re-open stdin: \n",
            "Selecting previously unselected package poppler-utils.\n",
            "(Reading database ... 123636 files and directories currently installed.)\n",
            "Preparing to unpack .../poppler-utils_22.02.0-2ubuntu0.5_amd64.deb ...\n",
            "Unpacking poppler-utils (22.02.0-2ubuntu0.5) ...\n",
            "Setting up poppler-utils (22.02.0-2ubuntu0.5) ...\n",
            "Processing triggers for man-db (2.10.2-1) ...\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pymupdf==1.24.4"
      ],
      "metadata": {
        "id": "MWNjOhSbRaOw",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "adc9f207-3873-4ff7-d556-b435babb1fbf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting pymupdf==1.24.4\n",
            "  Downloading PyMuPDF-1.24.4-cp310-none-manylinux2014_x86_64.whl.metadata (3.4 kB)\n",
            "Collecting PyMuPDFb==1.24.3 (from pymupdf==1.24.4)\n",
            "  Downloading PyMuPDFb-1.24.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.4 kB)\n",
            "Downloading PyMuPDF-1.24.4-cp310-none-manylinux2014_x86_64.whl (3.5 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.5/3.5 MB\u001b[0m \u001b[31m40.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading PyMuPDFb-1.24.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (15.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m15.8/15.8 MB\u001b[0m \u001b[31m101.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: PyMuPDFb, pymupdf\n",
            "Successfully installed PyMuPDFb-1.24.3 pymupdf-1.24.4\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### PDF Loaders\n",
        "\n",
        "[Portable Document Format (PDF)](https://en.wikipedia.org/wiki/PDF), standardized as ISO 32000, is a file format developed by Adobe in 1992 to present documents, including text formatting and images, in a manner independent of application software, hardware, and operating systems.\n",
        "\n",
        "LangChain integrates with a host of PDF parsers. Some are simple and relatively low-level; others will support OCR and image-processing, or perform advanced document layout analysis. The right choice will depend on your use-case and through experimentation.\n",
        "\n",
        "Here we will see how to load PDF documents into the LangChain `Document` format\n",
        "\n",
        "We download a research paper to experiment with"
      ],
      "metadata": {
        "id": "tZB2fxI9fKxC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "If the following command fails you can download the paper manually by going to http://arxiv.org/pdf/2103.15348.pdf, save it as `layoutparser_paper.pdf`and upload it on the left in Colab from the upload files option"
      ],
      "metadata": {
        "id": "p2NWiC51KDbm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget -O 'layoutparser_paper.pdf' 'http://arxiv.org/pdf/2103.15348.pdf'"
      ],
      "metadata": {
        "id": "t_zMe1cES7Tb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2b0168af-d454-4bcb-e398-6b323a2291df"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-07-27 06:58:49--  http://arxiv.org/pdf/2103.15348.pdf\n",
            "Resolving arxiv.org (arxiv.org)... 151.101.3.42, 151.101.131.42, 151.101.67.42, ...\n",
            "Connecting to arxiv.org (arxiv.org)|151.101.3.42|:80... connected.\n",
            "HTTP request sent, awaiting response... 301 Moved Permanently\n",
            "Location: http://arxiv.org/pdf/2103.15348 [following]\n",
            "--2024-07-27 06:58:49--  http://arxiv.org/pdf/2103.15348\n",
            "Reusing existing connection to arxiv.org:80.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4686220 (4.5M) [application/pdf]\n",
            "Saving to: ‘layoutparser_paper.pdf’\n",
            "\n",
            "layoutparser_paper. 100%[===================>]   4.47M  14.0MB/s    in 0.3s    \n",
            "\n",
            "2024-07-27 06:58:49 (14.0 MB/s) - ‘layoutparser_paper.pdf’ saved [4686220/4686220]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### PyMuPDFLoader\n",
        "\n",
        "This is the fastest of the PDF parsing options, and contains detailed metadata about the PDF and its pages, as well as returns one document per page. It uses the `pymupdf` library internally."
      ],
      "metadata": {
        "id": "x_T0_7KtfUJj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_community.document_loaders import PyMuPDFLoader\n",
        "\n",
        "loader = PyMuPDFLoader(\"./layoutparser_paper.pdf\")\n",
        "pages = loader.load()"
      ],
      "metadata": {
        "id": "l3KTMV_3XmfL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(pages)"
      ],
      "metadata": {
        "id": "JXUAxQnJhuHO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bc6d42cb-f4e0-4001-a2c2-195f2ffc5c3c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "16"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pages[0]"
      ],
      "metadata": {
        "id": "CZM5poERdRpL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "92186e6a-555e-426e-c0d1-31ed0b66570e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Document(metadata={'source': './layoutparser_paper.pdf', 'file_path': './layoutparser_paper.pdf', 'page': 0, 'total_pages': 16, 'format': 'PDF 1.5', 'title': '', 'author': '', 'subject': '', 'keywords': '', 'creator': 'LaTeX with hyperref', 'producer': 'pdfTeX-1.40.21', 'creationDate': 'D:20210622012710Z', 'modDate': 'D:20210622012710Z', 'trapped': ''}, page_content='LayoutParser: A Uniﬁed Toolkit for Deep\\nLearning Based Document Image Analysis\\nZejiang Shen1 (\\x00), Ruochen Zhang2, Melissa Dell3, Benjamin Charles Germain\\nLee4, Jacob Carlson3, and Weining Li5\\n1 Allen Institute for AI\\nshannons@allenai.org\\n2 Brown University\\nruochen zhang@brown.edu\\n3 Harvard University\\n{melissadell,jacob carlson}@fas.harvard.edu\\n4 University of Washington\\nbcgl@cs.washington.edu\\n5 University of Waterloo\\nw422li@uwaterloo.ca\\nAbstract. Recent advances in document image analysis (DIA) have been\\nprimarily driven by the application of neural networks. Ideally, research\\noutcomes could be easily deployed in production and extended for further\\ninvestigation. However, various factors like loosely organized codebases\\nand sophisticated model conﬁgurations complicate the easy reuse of im-\\nportant innovations by a wide audience. Though there have been on-going\\neﬀorts to improve reusability and simplify deep learning (DL) model\\ndevelopment in disciplines like natural language processing and computer\\nvision, none of them are optimized for challenges in the domain of DIA.\\nThis represents a major gap in the existing toolkit, as DIA is central to\\nacademic research across a wide range of disciplines in the social sciences\\nand humanities. This paper introduces LayoutParser, an open-source\\nlibrary for streamlining the usage of DL in DIA research and applica-\\ntions. The core LayoutParser library comes with a set of simple and\\nintuitive interfaces for applying and customizing DL models for layout de-\\ntection, character recognition, and many other document processing tasks.\\nTo promote extensibility, LayoutParser also incorporates a community\\nplatform for sharing both pre-trained models and full document digiti-\\nzation pipelines. We demonstrate that LayoutParser is helpful for both\\nlightweight and large-scale digitization pipelines in real-word use cases.\\nThe library is publicly available at https://layout-parser.github.io.\\nKeywords: Document Image Analysis · Deep Learning · Layout Analysis\\n· Character Recognition · Open Source library · Toolkit.\\n1\\nIntroduction\\nDeep Learning(DL)-based approaches are the state-of-the-art for a wide range of\\ndocument image analysis (DIA) tasks including document image classiﬁcation [11,\\narXiv:2103.15348v2  [cs.CV]  21 Jun 2021\\n')"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pages[0].metadata"
      ],
      "metadata": {
        "id": "Uhr0Y1C90TH0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "10e4f7ad-ec77-4e6d-ff6b-eb4723b9e75a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'source': './layoutparser_paper.pdf',\n",
              " 'file_path': './layoutparser_paper.pdf',\n",
              " 'page': 0,\n",
              " 'total_pages': 16,\n",
              " 'format': 'PDF 1.5',\n",
              " 'title': '',\n",
              " 'author': '',\n",
              " 'subject': '',\n",
              " 'keywords': '',\n",
              " 'creator': 'LaTeX with hyperref',\n",
              " 'producer': 'pdfTeX-1.40.21',\n",
              " 'creationDate': 'D:20210622012710Z',\n",
              " 'modDate': 'D:20210622012710Z',\n",
              " 'trapped': ''}"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(pages[0].page_content)"
      ],
      "metadata": {
        "id": "BdaA9hkHXmhs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c11f09eb-9162-43dd-84a4-409be133302a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "LayoutParser: A Uniﬁed Toolkit for Deep\n",
            "Learning Based Document Image Analysis\n",
            "Zejiang Shen1 (\u0000), Ruochen Zhang2, Melissa Dell3, Benjamin Charles Germain\n",
            "Lee4, Jacob Carlson3, and Weining Li5\n",
            "1 Allen Institute for AI\n",
            "shannons@allenai.org\n",
            "2 Brown University\n",
            "ruochen zhang@brown.edu\n",
            "3 Harvard University\n",
            "{melissadell,jacob carlson}@fas.harvard.edu\n",
            "4 University of Washington\n",
            "bcgl@cs.washington.edu\n",
            "5 University of Waterloo\n",
            "w422li@uwaterloo.ca\n",
            "Abstract. Recent advances in document image analysis (DIA) have been\n",
            "primarily driven by the application of neural networks. Ideally, research\n",
            "outcomes could be easily deployed in production and extended for further\n",
            "investigation. However, various factors like loosely organized codebases\n",
            "and sophisticated model conﬁgurations complicate the easy reuse of im-\n",
            "portant innovations by a wide audience. Though there have been on-going\n",
            "eﬀorts to improve reusability and simplify deep learning (DL) model\n",
            "development in disciplines like natural language processing and computer\n",
            "vision, none of them are optimized for challenges in the domain of DIA.\n",
            "This represents a major gap in the existing toolkit, as DIA is central to\n",
            "academic research across a wide range of disciplines in the social sciences\n",
            "and humanities. This paper introduces LayoutParser, an open-source\n",
            "library for streamlining the usage of DL in DIA research and applica-\n",
            "tions. The core LayoutParser library comes with a set of simple and\n",
            "intuitive interfaces for applying and customizing DL models for layout de-\n",
            "tection, character recognition, and many other document processing tasks.\n",
            "To promote extensibility, LayoutParser also incorporates a community\n",
            "platform for sharing both pre-trained models and full document digiti-\n",
            "zation pipelines. We demonstrate that LayoutParser is helpful for both\n",
            "lightweight and large-scale digitization pipelines in real-word use cases.\n",
            "The library is publicly available at https://layout-parser.github.io.\n",
            "Keywords: Document Image Analysis · Deep Learning · Layout Analysis\n",
            "· Character Recognition · Open Source library · Toolkit.\n",
            "1\n",
            "Introduction\n",
            "Deep Learning(DL)-based approaches are the state-of-the-art for a wide range of\n",
            "document image analysis (DIA) tasks including document image classiﬁcation [11,\n",
            "arXiv:2103.15348v2  [cs.CV]  21 Jun 2021\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(pages[4].page_content)"
      ],
      "metadata": {
        "id": "eDxTzEe0Le6I",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "807e4ce2-32c6-43dd-8de9-aaadf706df0d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "LayoutParser: A Uniﬁed Toolkit for DL-Based DIA\n",
            "5\n",
            "Table 1: Current layout detection models in the LayoutParser model zoo\n",
            "Dataset\n",
            "Base Model1 Large Model\n",
            "Notes\n",
            "PubLayNet [38]\n",
            "F / M\n",
            "M\n",
            "Layouts of modern scientiﬁc documents\n",
            "PRImA [3]\n",
            "M\n",
            "-\n",
            "Layouts of scanned modern magazines and scientiﬁc reports\n",
            "Newspaper [17]\n",
            "F\n",
            "-\n",
            "Layouts of scanned US newspapers from the 20th century\n",
            "TableBank [18]\n",
            "F\n",
            "F\n",
            "Table region on modern scientiﬁc and business document\n",
            "HJDataset [31]\n",
            "F / M\n",
            "-\n",
            "Layouts of history Japanese documents\n",
            "1 For each dataset, we train several models of diﬀerent sizes for diﬀerent needs (the trade-oﬀbetween accuracy\n",
            "vs. computational cost). For “base model” and “large model”, we refer to using the ResNet 50 or ResNet 101\n",
            "backbones [13], respectively. One can train models of diﬀerent architectures, like Faster R-CNN [28] (F) and Mask\n",
            "R-CNN [12] (M). For example, an F in the Large Model column indicates it has a Faster R-CNN model trained\n",
            "using the ResNet 101 backbone. The platform is maintained and a number of additions will be made to the model\n",
            "zoo in coming months.\n",
            "layout data structures, which are optimized for eﬃciency and versatility. 3) When\n",
            "necessary, users can employ existing or customized OCR models via the uniﬁed\n",
            "API provided in the OCR module. 4) LayoutParser comes with a set of utility\n",
            "functions for the visualization and storage of the layout data. 5) LayoutParser\n",
            "is also highly customizable, via its integration with functions for layout data\n",
            "annotation and model training. We now provide detailed descriptions for each\n",
            "component.\n",
            "3.1\n",
            "Layout Detection Models\n",
            "In LayoutParser, a layout model takes a document image as an input and\n",
            "generates a list of rectangular boxes for the target content regions. Diﬀerent\n",
            "from traditional methods, it relies on deep convolutional neural networks rather\n",
            "than manually curated rules to identify content regions. It is formulated as an\n",
            "object detection problem and state-of-the-art models like Faster R-CNN [28] and\n",
            "Mask R-CNN [12] are used. This yields prediction results of high accuracy and\n",
            "makes it possible to build a concise, generalized interface for layout detection.\n",
            "LayoutParser, built upon Detectron2 [35], provides a minimal API that can\n",
            "perform layout detection with only four lines of code in Python:\n",
            "1 import\n",
            "layoutparser as lp\n",
            "2 image = cv2.imread(\"image_file\") # load\n",
            "images\n",
            "3 model = lp. Detectron2LayoutModel (\n",
            "4\n",
            "\"lp:// PubLayNet/ faster_rcnn_R_50_FPN_3x /config\")\n",
            "5 layout = model.detect(image)\n",
            "LayoutParser provides a wealth of pre-trained model weights using various\n",
            "datasets covering diﬀerent languages, time periods, and document types. Due to\n",
            "domain shift [7], the prediction performance can notably drop when models are ap-\n",
            "plied to target samples that are signiﬁcantly diﬀerent from the training dataset. As\n",
            "document structures and layouts vary greatly in diﬀerent domains, it is important\n",
            "to select models trained on a dataset similar to the test samples. A semantic syntax\n",
            "is used for initializing the model weights in LayoutParser, using both the dataset\n",
            "name and model name lp://<dataset-name>/<model-architecture-name>.\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### UnstructuredPDFLoader\n",
        "\n",
        "[Unstructured.io](https://unstructured-io.github.io/unstructured/) supports a common interface for working with unstructured or semi-structured file formats, such as Markdown or PDF. LangChain's [`UnstructuredPDFLoader`](https://api.python.langchain.com/en/latest/document_loaders/langchain_community.document_loaders.pdf.UnstructuredPDFLoader.html) integrates with Unstructured to parse PDF documents into LangChain [`Document`](https://api.python.langchain.com/en/latest/documents/langchain_core.documents.base.Document.html) objects."
      ],
      "metadata": {
        "id": "3Zvyk9ACL8fx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_community.document_loaders import UnstructuredPDFLoader"
      ],
      "metadata": {
        "id": "Z8qz1HKWMRTf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Load PDF with complex parsing, table detection and chunking by sections"
      ],
      "metadata": {
        "id": "CeB_kHTjR6u5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# takes 3-4 mins on Colab\n",
        "loader = UnstructuredPDFLoader('./layoutparser_paper.pdf',\n",
        "                               strategy='hi_res',\n",
        "                               extract_images_in_pdf=True,\n",
        "                               infer_table_structure=True,\n",
        "                               chunking_strategy=\"by_title\", # section-based chunking\n",
        "                               max_characters=4000, # max size of chunks\n",
        "                               new_after_n_chars=3800, # preferred size of chunks\n",
        "                               combine_text_under_n_chars=2000, # smaller chunks < 2000 chars will be combined into a larger chunk\n",
        "                               mode='elements')\n",
        "data = loader.load()"
      ],
      "metadata": {
        "id": "pdKdca4RMmY8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(data)"
      ],
      "metadata": {
        "id": "0WvtzXCgMots",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ce118316-412b-4d2a-9ad1-efa456902204"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "18"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "[doc.metadata['category'] for doc in data]"
      ],
      "metadata": {
        "id": "JM5fCpJNPPiv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e887db04-f080-46b2-dc9d-c66a80b4a34b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'Table',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'Table',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement',\n",
              " 'CompositeElement']"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data[0]"
      ],
      "metadata": {
        "id": "9f7zcYUxMq0t",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e04d08b5-2c5b-4f0b-fff9-a5ba04f5344a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Document(metadata={'source': './layoutparser_paper.pdf', 'filetype': 'application/pdf', 'languages': ['eng'], 'last_modified': '2023-01-23T09:15:33', 'page_number': 1, 'orig_elements': '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', 'file_directory': '.', 'filename': 'layoutparser_paper.pdf', 'category': 'CompositeElement'}, page_content='1 2 0 2\\n\\nn u J 1 2 ] V C . s c [\\n\\n2 v 8 4 3 5 1 . 3 0 1 2 : v i X r a\\n\\nLayoutParser: A Uniﬁed Toolkit for Deep Learning Based Document Image Analysis\\n\\nZejiang Shen! (4), Ruochen Zhang”, Melissa Dell?, Benjamin Charles Germain Lee*, Jacob Carlson’, and Weining Li®\\n\\n1 Allen Institute for AI shannons@allenai.org 2 Brown University ruochen zhang@brown.edu 3 Harvard University {melissadell,jacob carlson}@fas.harvard.edu 4 University of Washington bcgl@cs.washington.edu 5 University of Waterloo w422li@uwaterloo.ca\\n\\nAbstract. Recent advances in document image analysis (DIA) have been primarily driven by the application of neural networks. Ideally, research outcomes could be easily deployed in production and extended for further investigation. However, various factors like loosely organized codebases and sophisticated model conﬁgurations complicate the easy reuse of im- portant innovations by a wide audience. Though there have been on-going eﬀorts to improve reusability and simplify deep learning (DL) model development in disciplines like natural language processing and computer vision, none of them are optimized for challenges in the domain of DIA. This represents a major gap in the existing toolkit, as DIA is central to academic research across a wide range of disciplines in the social sciences and humanities. This paper introduces LayoutParser, an open-source library for streamlining the usage of DL in DIA research and applica- tions. The core LayoutParser library comes with a set of simple and intuitive interfaces for applying and customizing DL models for layout de- tection, character recognition, and many other document processing tasks. To promote extensibility, LayoutParser also incorporates a community platform for sharing both pre-trained models and full document digiti- zation pipelines. We demonstrate that LayoutParser is helpful for both lightweight and large-scale digitization pipelines in real-word use cases. The library is publicly available at https://layout-parser.github.io.\\n\\nKeywords: Document Image Analysis · Deep Learning · Layout Analysis · Character Recognition · Open Source library · Toolkit.')"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(data[0].page_content)"
      ],
      "metadata": {
        "id": "XdjS77Fh0oKJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "64895ec2-2d90-4c12-9875-b8160d4f5269"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1 2 0 2\n",
            "\n",
            "n u J 1 2 ] V C . s c [\n",
            "\n",
            "2 v 8 4 3 5 1 . 3 0 1 2 : v i X r a\n",
            "\n",
            "LayoutParser: A Uniﬁed Toolkit for Deep Learning Based Document Image Analysis\n",
            "\n",
            "Zejiang Shen! (4), Ruochen Zhang”, Melissa Dell?, Benjamin Charles Germain Lee*, Jacob Carlson’, and Weining Li®\n",
            "\n",
            "1 Allen Institute for AI shannons@allenai.org 2 Brown University ruochen zhang@brown.edu 3 Harvard University {melissadell,jacob carlson}@fas.harvard.edu 4 University of Washington bcgl@cs.washington.edu 5 University of Waterloo w422li@uwaterloo.ca\n",
            "\n",
            "Abstract. Recent advances in document image analysis (DIA) have been primarily driven by the application of neural networks. Ideally, research outcomes could be easily deployed in production and extended for further investigation. However, various factors like loosely organized codebases and sophisticated model conﬁgurations complicate the easy reuse of im- portant innovations by a wide audience. Though there have been on-going eﬀorts to improve reusability and simplify deep learning (DL) model development in disciplines like natural language processing and computer vision, none of them are optimized for challenges in the domain of DIA. This represents a major gap in the existing toolkit, as DIA is central to academic research across a wide range of disciplines in the social sciences and humanities. This paper introduces LayoutParser, an open-source library for streamlining the usage of DL in DIA research and applica- tions. The core LayoutParser library comes with a set of simple and intuitive interfaces for applying and customizing DL models for layout de- tection, character recognition, and many other document processing tasks. To promote extensibility, LayoutParser also incorporates a community platform for sharing both pre-trained models and full document digiti- zation pipelines. We demonstrate that LayoutParser is helpful for both lightweight and large-scale digitization pipelines in real-word use cases. The library is publicly available at https://layout-parser.github.io.\n",
            "\n",
            "Keywords: Document Image Analysis · Deep Learning · Layout Analysis · Character Recognition · Open Source library · Toolkit.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data[5]"
      ],
      "metadata": {
        "id": "six4KwFGPX-S",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "264f3a1d-d692-466d-cbc8-ecb7a2cb48e3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Document(metadata={'source': './layoutparser_paper.pdf', 'last_modified': '2023-01-23T09:15:33', 'text_as_html': \"<table><thead><tr><th>Dataset</th><th>| Base Model'|</th><th>| Notes</th></tr></thead><tbody><tr><td>PubLayNet B8]|</td><td>F/M</td><td>Layouts of modern scientific documents</td></tr><tr><td>PRImA</td><td>M</td><td>Layouts of scanned modern magazines and scientific report</td></tr><tr><td>Newspaper</td><td>F</td><td>Layouts of scanned US newspapers from the 20th century</td></tr><tr><td>TableBank</td><td>F</td><td>Table region on modern scientific and business document</td></tr><tr><td>HJDataset</td><td>F/M</td><td>Layouts of history Japanese documents</td></tr></tbody></table>\", 'filetype': 'application/pdf', 'languages': ['eng'], 'page_number': 5, 'orig_elements': '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', 'file_directory': '.', 'filename': 'layoutparser_paper.pdf', 'category': 'Table'}, page_content='Dataset Base Model1 Large Model Notes PubLayNet [38] PRImA [3] Newspaper [17] TableBank [18] HJDataset [31] F / M M F F F / M M - - F - Layouts of modern scientiﬁc documents Layouts of scanned modern magazines and scientiﬁc reports Layouts of scanned US newspapers from the 20th century Table region on modern scientiﬁc and business document Layouts of history Japanese documents')"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data[5].page_content"
      ],
      "metadata": {
        "id": "AYe9e5K3Pfje",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        },
        "outputId": "e0833c02-7aca-434e-c524-4e0aae985dab"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'Dataset Base Model1 Large Model Notes PubLayNet [38] PRImA [3] Newspaper [17] TableBank [18] HJDataset [31] F / M M F F F / M M - - F - Layouts of modern scientiﬁc documents Layouts of scanned modern magazines and scientiﬁc reports Layouts of scanned US newspapers from the 20th century Table region on modern scientiﬁc and business document Layouts of history Japanese documents'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from IPython.display import HTML\n",
        "\n",
        "HTML(data[5].metadata['text_as_html'])"
      ],
      "metadata": {
        "id": "nFI5BBDjPoHK",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 147
        },
        "outputId": "10fead30-7a01-48e6-84fa-8eb9c3b41c4a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<table><thead><tr><th>Dataset</th><th>| Base Model'|</th><th>| Notes</th></tr></thead><tbody><tr><td>PubLayNet B8]|</td><td>F/M</td><td>Layouts of modern scientific documents</td></tr><tr><td>PRImA</td><td>M</td><td>Layouts of scanned modern magazines and scientific report</td></tr><tr><td>Newspaper</td><td>F</td><td>Layouts of scanned US newspapers from the 20th century</td></tr><tr><td>TableBank</td><td>F</td><td>Table region on modern scientific and business document</td></tr><tr><td>HJDataset</td><td>F/M</td><td>Layouts of history Japanese documents</td></tr></tbody></table>"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "! ls -l ./figures"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mgU18vNXCNEY",
        "outputId": "d03d389f-673f-494d-ff00-a37ae2d39971"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "total 336\n",
            "-rw-r--r-- 1 root root 89611 Jul 27 07:20 figure-10-4.jpg\n",
            "-rw-r--r-- 1 root root 42187 Jul 27 07:20 figure-11-5.jpg\n",
            "-rw-r--r-- 1 root root 62655 Jul 27 07:20 figure-13-6.jpg\n",
            "-rw-r--r-- 1 root root 36497 Jul 27 07:20 figure-4-1.jpg\n",
            "-rw-r--r-- 1 root root 31580 Jul 27 07:20 figure-6-2.jpg\n",
            "-rw-r--r-- 1 root root 72034 Jul 27 07:20 figure-9-3.jpg\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from IPython.display import Image\n",
        "\n",
        "Image('./figures/figure-4-1.jpg')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 345
        },
        "id": "vh_zj91eB3Ff",
        "outputId": "de6d3bde-8929-4710-fda0-d050582659c6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/jpeg": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Microsoft Office Document Loaders\n",
        "\n",
        "The Microsoft Office suite of productivity software includes Microsoft Word, Microsoft Excel, Microsoft PowerPoint, Microsoft Outlook, and Microsoft OneNote. It is available for Microsoft Windows and macOS operating systems. It is also available on Android and iOS.\n",
        "\n",
        "[Unstructured.io](https://docs.unstructured.io/open-source/introduction/overview) provides a variety of document loaders to load MS Office documents. Check them out [here](https://docs.unstructured.io/open-source/core-functionality/partitioning).\n",
        "\n",
        "Here we will leverage LangChain's [`UnstructuredWordDocumentLoader`](https://api.python.langchain.com/en/latest/document_loaders/langchain_community.document_loaders.word_document.UnstructuredWordDocumentLoader.html) to load data from a MS Word document."
      ],
      "metadata": {
        "id": "QQV0X4grW971"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!gdown 1DEz13a7k4yX9yFrWaz3QJqHdfecFYRV-"
      ],
      "metadata": {
        "id": "V37elBlCT6Mn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "26f1309c-c0b6-4b97-e7ee-c1ea4db3b496"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1DEz13a7k4yX9yFrWaz3QJqHdfecFYRV-\n",
            "To: /content/Intel Strategy.docx\n",
            "100% 1.13M/1.13M [00:00<00:00, 9.98MB/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_community.document_loaders import UnstructuredWordDocumentLoader"
      ],
      "metadata": {
        "id": "BqYBSoqHT6TY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Load word doc with complex parsing and section based chunks"
      ],
      "metadata": {
        "id": "ILz-pXIyX8e0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "loader = UnstructuredWordDocumentLoader('./Intel Strategy.docx',\n",
        "                                        strategy='fast',\n",
        "                                        chunking_strategy=\"by_title\",\n",
        "                                        max_characters=4000, # max limit of a document chunk\n",
        "                                        new_after_n_chars=3800, # preferred document chunk size\n",
        "                                        mode='elements')\n",
        "data = loader.load()"
      ],
      "metadata": {
        "id": "7cKYWNjAUo1A"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(data)"
      ],
      "metadata": {
        "id": "CkOn6t0QVOyE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "96612d50-0a7c-463e-ea1d-0ed157e63926"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data[0]"
      ],
      "metadata": {
        "id": "IV4cp6yqVQZ1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "936a181a-a394-4ca4-9100-f3b80f2993d7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Document(metadata={'source': './Intel Strategy.docx', 'emphasized_text_contents': ['The Superpowers', 'Pervasive Connectivity', 'Ubiquitous compute'], 'emphasized_text_tags': ['b', 'b', 'b'], 'file_directory': '.', 'filename': 'Intel Strategy.docx', 'languages': ['eng'], 'last_modified': '2024-07-27T07:21:41', 'orig_elements': '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', 'filetype': 'application/vnd.openxmlformats-officedocument.wordprocessingml.document', 'category': 'CompositeElement'}, page_content='Intel Strategy\\n\\nOver the last few years, Intel, one of the world’s biggest chipmakers, has been transitioning towards a more datacentric approach than PC-centric. This is a welcome move, not only for the company but for innovation in technology as well, says Pat Gelsinger, CEO, Intel. According to him, the changing times as well as strides in innovation have placed Intel in a position where it can leverage the “superpowers” to make the world of computing better sustainable and far superior to the present scenario.\\n\\nThe Superpowers\\n\\nPervasive connectivity, Ubiquitous compute, AI and Cloud-to-Edge Infrastructure -- the four superpowers that will bolster Intel’s footprints into the future, will also play a key role in transforming the world of computing in any device.\\n\\n“Each of these superpowers is impressive on its own, but when they come together, that’s magic. If you’re not applying AI to every one of your business processes, you’re falling behind. We’re seeing this across every industry,” Gelsinger said.\\n\\nPervasive Connectivity: 5G-empowered pervasive connectivity, that intends to connect all, allows customers to gather, store, write, access, and analyze data regardless of device or location. This level of connectivity is essential in creating an improved quality of life, Gelsinger said. He added that Intel was partnering with Taiwan’s Pegatron to produce 5G networking that could be deployed in extreme conditions, too. “Think of it … earthquakes, tornadoes, natural disasters, where the communications infrastructure is knocked out. And imagine that you were a first responder. You’re showing up for a disaster relief situation and you have no communications.” “We’re taking advantage of the advances in 5G availability of wireless spectrum. And you can think about this as a blueprint for next-generation, commercial 5G, the ramp deployments,” Gelsinger said.\\n\\nUbiquitous compute: “Everything has become a computer, essentially any device we touch. Literally compute is now how we experience the world.” Gelsinger said. It is in line with the company’s data-centric approach as well, which include Server and Storage, including CPUs, chipsets, accelerators, memory and storage media in servers and storage systems; Networking and Connectivity, including CPUs, chipsets, accelerators, memory and storage media, and connectivity devices in network appliances and network function virtualization (NFV) systems; Internet of Things, including addressable logic application-specific integrated circuits and standard products, microprocessors, microcontrollers, digital signal processors, memory and storage media and modems in industrial, transportation, automated driving, retail, video surveillance, healthcare, public sector, office automation, gaming and smart home.')"
            ]
          },
          "metadata": {},
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(data[0].page_content)"
      ],
      "metadata": {
        "id": "mW5UhN9BVTaE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "219ce832-785b-4d17-b1ec-763cedafda45"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Intel Strategy\n",
            "\n",
            "Over the last few years, Intel, one of the world’s biggest chipmakers, has been transitioning towards a more datacentric approach than PC-centric. This is a welcome move, not only for the company but for innovation in technology as well, says Pat Gelsinger, CEO, Intel. According to him, the changing times as well as strides in innovation have placed Intel in a position where it can leverage the “superpowers” to make the world of computing better sustainable and far superior to the present scenario.\n",
            "\n",
            "The Superpowers\n",
            "\n",
            "Pervasive connectivity, Ubiquitous compute, AI and Cloud-to-Edge Infrastructure -- the four superpowers that will bolster Intel’s footprints into the future, will also play a key role in transforming the world of computing in any device.\n",
            "\n",
            "“Each of these superpowers is impressive on its own, but when they come together, that’s magic. If you’re not applying AI to every one of your business processes, you’re falling behind. We’re seeing this across every industry,” Gelsinger said.\n",
            "\n",
            "Pervasive Connectivity: 5G-empowered pervasive connectivity, that intends to connect all, allows customers to gather, store, write, access, and analyze data regardless of device or location. This level of connectivity is essential in creating an improved quality of life, Gelsinger said. He added that Intel was partnering with Taiwan’s Pegatron to produce 5G networking that could be deployed in extreme conditions, too. “Think of it … earthquakes, tornadoes, natural disasters, where the communications infrastructure is knocked out. And imagine that you were a first responder. You’re showing up for a disaster relief situation and you have no communications.” “We’re taking advantage of the advances in 5G availability of wireless spectrum. And you can think about this as a blueprint for next-generation, commercial 5G, the ramp deployments,” Gelsinger said.\n",
            "\n",
            "Ubiquitous compute: “Everything has become a computer, essentially any device we touch. Literally compute is now how we experience the world.” Gelsinger said. It is in line with the company’s data-centric approach as well, which include Server and Storage, including CPUs, chipsets, accelerators, memory and storage media in servers and storage systems; Networking and Connectivity, including CPUs, chipsets, accelerators, memory and storage media, and connectivity devices in network appliances and network function virtualization (NFV) systems; Internet of Things, including addressable logic application-specific integrated circuits and standard products, microprocessors, microcontrollers, digital signal processors, memory and storage media and modems in industrial, transportation, automated driving, retail, video surveillance, healthcare, public sector, office automation, gaming and smart home.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Better Prompting for Consistent Results"
      ],
      "metadata": {
        "id": "U3hZ_O32as0x"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Enter Open AI API Key"
      ],
      "metadata": {
        "id": "H9c37cLnSrbg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from getpass import getpass\n",
        "\n",
        "OPENAI_KEY = getpass('Enter Open AI API Key: ')"
      ],
      "metadata": {
        "id": "cv3JzCEx_PAd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5a47a66d-49c8-4893-f71e-409c69116ca5"
      },
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Enter Open AI API Key: ··········\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "\n",
        "os.environ['OPENAI_API_KEY'] = OPENAI_KEY"
      ],
      "metadata": {
        "id": "zT8VTBilFwQB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Load Connection to LLM\n",
        "\n",
        "Here we create a connection to ChatGPT to use later in our chains"
      ],
      "metadata": {
        "id": "2oeckxFBcc0E"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_openai import ChatOpenAI\n",
        "\n",
        "chatgpt = ChatOpenAI(model_name='gpt-3.5-turbo', temperature=0)"
      ],
      "metadata": {
        "id": "vHa9LMOfcOCV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Problematic RAG Prompt"
      ],
      "metadata": {
        "id": "htTiJ8RTgiAG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_core.prompts import ChatPromptTemplate\n",
        "\n",
        "prompt = \"\"\"You are an assistant for question-answering tasks.\n",
        "            Give an answer to the following question with the context provided\n",
        "\n",
        "            Question:\n",
        "            {question}\n",
        "\n",
        "            Context:\n",
        "            {context}\n",
        "\n",
        "            Answer:\n",
        "         \"\"\"\n",
        "\n",
        "prompt_template = ChatPromptTemplate.from_template(prompt)"
      ],
      "metadata": {
        "id": "JcHrMo7tF2dR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "context = \"\"\"\n",
        "The cricket world cup champion of 2023 is Australia\n",
        "\n",
        "The cricket Twenty20 International champion of 2024 is India\n",
        "\"\"\"\n",
        "\n",
        "question = \"Who is the euro 2020 champion?\"\n",
        "\n",
        "qa_rag_chain = (\n",
        "    prompt_template\n",
        "      |\n",
        "    chatgpt\n",
        ")"
      ],
      "metadata": {
        "id": "nfR1L5bKGA6Z"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "response = qa_rag_chain.invoke({'context': context, 'question': question})\n",
        "print(response.content)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NCZARTPXGapG",
        "outputId": "6f0b80c3-b036-4a06-bffe-d0298d940e38"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The euro 2020 champion is Italy.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Better RAG Prompt"
      ],
      "metadata": {
        "id": "uVc9olR5gmP6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_core.prompts import ChatPromptTemplate\n",
        "\n",
        "prompt = \"\"\"You are an assistant for question-answering tasks.\n",
        "            Use the following pieces of retrieved context to answer the question.\n",
        "            If no context is present or if you don't know the answer, just say that you don't know.\n",
        "            Do not make up the answer unless it is there in the provided context.\n",
        "\n",
        "            Question:\n",
        "            {question}\n",
        "\n",
        "            Context:\n",
        "            {context}\n",
        "\n",
        "            Answer:\n",
        "         \"\"\"\n",
        "\n",
        "prompt_template = ChatPromptTemplate.from_template(prompt)"
      ],
      "metadata": {
        "id": "6VhrwTZcHKed"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "qa_rag_chain = (\n",
        "    prompt_template\n",
        "      |\n",
        "    chatgpt\n",
        ")"
      ],
      "metadata": {
        "id": "5_oTcrZAHbdv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "response = qa_rag_chain.invoke({'context': context, 'question': question})\n",
        "print(response.content)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DBT4WW12HZ_j",
        "outputId": "8c37a9f3-145c-43e0-dafc-a32aa4f11a91"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "I don't know.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Agentic RAG with Tools"
      ],
      "metadata": {
        "id": "DeLmkbOIaoWF"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Enter Tavily Search Tool API Key\n",
        "\n",
        "Get a free API key from [here](https://tavily.com/#api)"
      ],
      "metadata": {
        "id": "ucWRRI3QztL2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "TAVILY_API_KEY = getpass('Enter Tavily Search API Key: ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mK-1WLzOrJdb",
        "outputId": "a33b7cd6-b390-4a75-8cd1-c0ecd2daa95d"
      },
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Enter Tavily Search API Key: ··········\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "os.environ['TAVILY_API_KEY'] = TAVILY_API_KEY"
      ],
      "metadata": {
        "id": "R2h8vdQNVNjZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Setup Search Tool"
      ],
      "metadata": {
        "id": "GQ36nHepgrgp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from langchain_community.tools.tavily_search import TavilySearchResults\n",
        "from langchain_core.tools import tool\n",
        "import rich\n",
        "\n",
        "@tool\n",
        "def search_web(query: str) -> list:\n",
        "    \"\"\"Search the web for a query.\"\"\"\n",
        "    tavily_tool = TavilySearchResults(max_results=3,\n",
        "                                      search_depth='advanced',\n",
        "                                      max_tokens=10000)\n",
        "    results = tavily_tool.invoke(query)\n",
        "    return [doc['content'] for doc in results]"
      ],
      "metadata": {
        "id": "ALfCdvXqVO09"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "question"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "N7jzQYr7VesD",
        "outputId": "504af0ea-d30c-4659-f058-3db478b6a593"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'Who is the euro 2020 champion?'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "search_web(question)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bBuA2jXoVdL_",
        "outputId": "8027e5ea-cf3b-41a5-bbf1-1de633eec5f1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[\"Italy are the champions of Europe. On Saturday in the Euro 2020 final, the Italians defeated England in a penalty shootout to become the undisputed kings of the continent. The win is Italy's ...\",\n",
              " \"The UEFA Euro 2020 final was an association football match between Italy and England that took place at Wembley Stadium in London, England, on 11 July 2021 to determine the winner of UEFA Euro 2020.It was the 16th final of the UEFA European Championship, a quadrennial tournament contested by the senior men's national teams of the member associations of UEFA to decide the champions of Europe.\",\n",
              " \"The 2020 UEFA European Football Championship, commonly referred to as UEFA Euro 2020 or simply Euro 2020, was the 16th UEFA European Championship, the quadrennial international men's football championship of Europe organised by the Union of European Football Associations (UEFA). To celebrate the 60th anniversary of the European Championship competition, UEFA president Michel Platini declared ...\"]"
            ]
          },
          "metadata": {},
          "execution_count": 93
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Bind Tools to LLM"
      ],
      "metadata": {
        "id": "k4Khdw6Agvrn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "tools = [search_web]\n",
        "chatgpt = ChatOpenAI(model_name='gpt-4o-mini', temperature=0)\n",
        "chatgpt_with_tools = chatgpt.bind_tools(tools)"
      ],
      "metadata": {
        "id": "4hCzYX4lVguE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Better RAG Prompt with Tool Calling"
      ],
      "metadata": {
        "id": "m2TlF4AwgyJT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "prompt = \"\"\"You are an assistant for question-answering tasks.\n",
        "            Use the following pieces of retrieved context to answer the question.\n",
        "            If no context is present or if you don't know the answer,\n",
        "            check and see if you can use the tools available to you to get the answer.\n",
        "\n",
        "            Question:\n",
        "            {question}\n",
        "\n",
        "            Context:\n",
        "            {context}\n",
        "\n",
        "            Answer:\n",
        "         \"\"\"\n",
        "\n",
        "prompt_template = ChatPromptTemplate.from_template(prompt)\n",
        "\n",
        "qa_rag_chain = (\n",
        "    prompt_template\n",
        "      |\n",
        "    chatgpt_with_tools\n",
        ")"
      ],
      "metadata": {
        "id": "X-5xXrygWyLL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "context = \"\"\"\n",
        "The cricket world cup champion of 2023 is Australia\n",
        "\n",
        "The cricket Twenty20 International champion of 2024 is India\n",
        "\"\"\"\n",
        "\n",
        "question = \"Who is the euro 2020 champion?\"\n",
        "\n",
        "qa_rag_chain.invoke({'context': context, 'question': question})"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pYUMXThpW8Bs",
        "outputId": "6ea92733-1472-46cc-ad1b-fe1bf6d5d88d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_QN6XQaX2He2mknR9kKvnXz49', 'function': {'arguments': '{\"query\":\"Euro 2020 champion\"}', 'name': 'search_web'}, 'type': 'function'}]}, response_metadata={'token_usage': {'completion_tokens': 18, 'prompt_tokens': 146, 'total_tokens': 164}, 'model_name': 'gpt-4o-mini-2024-07-18', 'system_fingerprint': 'fp_611b667b19', 'finish_reason': 'tool_calls', 'logprobs': None}, id='run-28ca4691-01cd-4859-972d-88c3cccbd1ea-0', tool_calls=[{'name': 'search_web', 'args': {'query': 'Euro 2020 champion'}, 'id': 'call_QN6XQaX2He2mknR9kKvnXz49', 'type': 'tool_call'}], usage_metadata={'input_tokens': 146, 'output_tokens': 18, 'total_tokens': 164})"
            ]
          },
          "metadata": {},
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "question = \"Who is the T20 2024 champion?\"\n",
        "\n",
        "qa_rag_chain.invoke({'context': context, 'question': question})"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HvmXXv_YXQgc",
        "outputId": "3a634cdb-39ed-46ab-c929-816f22583dd1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "AIMessage(content='The T20 2024 champion is India.', response_metadata={'token_usage': {'completion_tokens': 11, 'prompt_tokens': 147, 'total_tokens': 158}, 'model_name': 'gpt-4o-mini-2024-07-18', 'system_fingerprint': 'fp_ba606877f9', 'finish_reason': 'stop', 'logprobs': None}, id='run-fbf2b8da-b724-40ea-9af1-16484b0ae0b1-0', usage_metadata={'input_tokens': 147, 'output_tokens': 11, 'total_tokens': 158})"
            ]
          },
          "metadata": {},
          "execution_count": 100
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Simple Agentic RAG with Tool Calls"
      ],
      "metadata": {
        "id": "mGB14NL5g94n"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def agentic_rag(question, context):\n",
        "  tool_call_map = {'search_web' : search_web}\n",
        "  response = qa_rag_chain.invoke({'context': context, 'question': question})\n",
        "\n",
        "  # if response content is present then we have our answer\n",
        "  if response.content:\n",
        "    print('Answer is in retrieved context')\n",
        "    answer = response.content\n",
        "\n",
        "  # if no response content present then call search tool\n",
        "  elif response.tool_calls:\n",
        "    print('Answer not in context, trying to use tools')\n",
        "    tool_call = response.tool_calls[0]\n",
        "    selected_tool = tool_call_map[tool_call[\"name\"].lower()]\n",
        "    print(f\"Calling tool: {tool_call['name']}\")\n",
        "    tool_output = selected_tool.invoke(tool_call[\"args\"])\n",
        "    context = '\\n\\n'.join(tool_output)\n",
        "    response = qa_rag_chain.invoke({'context': context, 'question': question})\n",
        "    answer = response.content\n",
        "\n",
        "  # no answer found from web search also\n",
        "  else:\n",
        "    answer = 'No answer found'\n",
        "\n",
        "  print(answer)"
      ],
      "metadata": {
        "id": "VzsRkX2AXGgl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "context = \"\"\"\n",
        "The cricket world cup champion of 2023 is Australia\n",
        "\n",
        "The cricket Twenty20 International champion of 2024 is India\n",
        "\"\"\"\n",
        "\n",
        "question = \"Who is the T20 2024 champion?\"\n",
        "\n",
        "agentic_rag(question, context)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KBm8mJSpXjNd",
        "outputId": "e423a026-e72c-4442-8c70-4a5de553454c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Answer is in retrieved context\n",
            "The T20 2024 champion is India.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "question = \"Who is the euro 2024 champion?\"\n",
        "\n",
        "agentic_rag(question, context)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U-Zdn-quXjhY",
        "outputId": "e016abc7-7958-4b53-f17d-7c10f32da117"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Answer not in context, trying to use tools\n",
            "Calling tool: search_web\n",
            "Spain is the Euro 2024 champion, having won their fourth European Championship by defeating England 2-1 in the final held in Berlin.\n"
          ]
        }
      ]
    }
  ]
}