{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# marker pdf conversion\n",
        "\n",
        "\n",
        "- https://github.com/VikParuchuri/marker/tree/master\n",
        "\n",
        "\n",
        "---"
      ],
      "metadata": {
        "id": "PzogeOQ6rbvL"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lIYdn1woOS1n"
      },
      "outputs": [],
      "source": [
        "!git clone https://github.com/VikParuchuri/marker.git\n",
        "%cd marker"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install -U -q pip ninja poetry"
      ],
      "metadata": {
        "id": "YkbBuXMudX5g"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%%bash\n",
        "set -e\n",
        "sudo apt-get update\n",
        "sudo apt-get install -y software-properties-common\n",
        "sudo apt-get install -y build-essential python3 python3-pip git\n",
        "sudo apt-get install -y python3-dev libssl-dev libc6-dev && sudo apt install zip unzip tmux nano p7zip-full git -y\n",
        "clear\n",
        "echo \"done\""
      ],
      "metadata": {
        "id": "JzW38yGYhH3W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# %%bash\n",
        "!sudo apt update && sudo apt upgrade -y > /dev/null\n",
        "!bash scripts/install/tesseract_5_install.sh\n",
        "# !bash scripts/install/ghostscript_install.sh && clear\n",
        "!sudo apt install ghostscript && clear\n",
        "!cat scripts/install/apt-requirements.txt | xargs sudo apt install -y --allow-unauthenticated > LOG_installs_apt.log"
      ],
      "metadata": {
        "id": "cDcjkP3hd_Kh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%%writefile convert_pdf_dir.sh\n",
        "#@title convert_pdf_dir.sh\n",
        "#!/bin/bash\n",
        "\n",
        "# Default source directory is the current directory if not specified\n",
        "SOURCE_DIR=${1:-$(pwd)}\n",
        "# Default target directory is parent of source directory + marker-pdf2text\n",
        "TARGET_DIR=${2:-$(dirname \"$(realpath $SOURCE_DIR)\")/marker-pdf2text}\n",
        "\n",
        "# Function to convert a single PDF file\n",
        "convert_pdf() {\n",
        "    local source_pdf=$1\n",
        "    local target_md=$2\n",
        "    mkdir -p \"$(dirname \"$target_md\")\" # Ensure the target directory exists\n",
        "    poetry run python convert_single.py \"$source_pdf\" \"$target_md\"\n",
        "}\n",
        "\n",
        "export -f convert_pdf\n",
        "\n",
        "# Find all PDF files and convert them\n",
        "find \"$SOURCE_DIR\" -type f -name '*.pdf' -exec bash -c 'convert_pdf \"$0\" \"${1/pdf/md}\"' {} \"$TARGET_DIR/$(realpath --relative-to=\"$SOURCE_DIR\" \"{}\")\" \\;\n",
        "\n",
        "echo \"Conversion complete. All markdown files have been saved to $TARGET_DIR.\""
      ],
      "metadata": {
        "cellView": "form",
        "id": "aRi6CBR_ekIw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!find / -name tessdata\n",
        "!echo \"TESSDATA_PREFIX=/usr/share/tesseract-ocr/5/tessdata\" > local.env\n",
        "!sudo rm *.lock\n",
        "!poetry install -q"
      ],
      "metadata": {
        "id": "TlQ6wV_Lfawh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "299eca48-79ba-41a3-95c1-7b5ffc657b14"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1A\u001b[0J  \u001b[34;1m-\u001b[39;22m \u001b[39mInstalling \u001b[39m\u001b[36mtorch\u001b[39m\u001b[39m (\u001b[39m\u001b[39;1m2.3.0\u001b[39;22m\u001b[39m)\u001b[39m: \u001b[34mInstalling...\u001b[39m\n",
            "\u001b[1A\u001b[0J  \u001b[32;1m-\u001b[39;22m \u001b[39mInstalling \u001b[39m\u001b[36mtorch\u001b[39m\u001b[39m (\u001b[39m\u001b[32m2.3.0\u001b[39m\u001b[39m)\u001b[39m\n",
            "\n",
            "\u001b[34mWriting lock file\u001b[39m\n",
            "\n",
            "\u001b[39;1mInstalling\u001b[39;22m the current project: \u001b[36mmarker-pdf\u001b[39m (\u001b[39;1m0.1.3\u001b[39;22m)\u001b[1G\u001b[2K\u001b[39;1mInstalling\u001b[39;22m the current project: \u001b[36mmarker-pdf\u001b[39m (\u001b[32m0.1.3\u001b[39m)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "jNOFjXVJSz2L"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## from dropbox\n"
      ],
      "metadata": {
        "id": "yn0qVN93V0v6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "URL = \"https://www.dropbox.com/scl/fo/kbbliz7or5ptwc1c7vgb9/AF0b76tmbY3BuxHDnMzNO1E?rlkey=x5q3r7o9wdob85xjazjanloe3&dl=1\" # @param {type:\"string\"}\n",
        "!wget -q -O data.zip $URL\n",
        "!unzip data.zip -d /content/pdf-textbooks\n",
        "!rm data.zip && clear"
      ],
      "metadata": {
        "id": "YwnblF4_iZuu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# !bash convert_pdf_dir.sh /content/pdf-textbooks"
      ],
      "metadata": {
        "id": "MmkB5ix-imgH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!DEFAULT_LANG=\"en\" TORCH_DEVICE=cuda poetry run python convert.py\\\n",
        "    /content/pdf-textbooks/python-book \\\n",
        "    /content/marker-md2text/python-book \\\n",
        "    --workers 6 --min_length 100"
      ],
      "metadata": {
        "id": "GWZURJALmDtV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "out_file=\"/content/archived-all-marker.7z\"\n",
        "!find /content/marker-md2text -type f -name \"*.json\" -print -delete # fuck off\n",
        "!7z a $out_file /content/marker-md2text"
      ],
      "metadata": {
        "id": "c5xndco3mJdq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## napierone"
      ],
      "metadata": {
        "id": "qybBz9Oqp2pH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "target_dir = '/content/napierone-epub'\n",
        "!wget -q http://napierone.com.s3-eu-north-1.amazonaws.com/NapierOne/Data/EPUB/EPUB-total.zip\n",
        "!unzip -o -q EPUB-total.zip -d $target_dir\n",
        "!ls $target_dir | wc -l\n",
        "!rm *.zip"
      ],
      "metadata": {
        "id": "jhB5lOA6n8f5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!DEFAULT_LANG=\"en\" TORCH_DEVICE=cuda poetry run python convert.py\\\n",
        "    $target_dir \\\n",
        "    /content/marker-napierone-file2text/EPUB \\\n",
        "    --workers 6 --min_length 300"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Poi65wYSpryu",
        "outputId": "8a20c15e-ae37-4be3-85e8-be3dea88afb8"
      },
      "execution_count": null,
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "2024-05-08 22:37:05,427\tINFO worker.py:1749 -- Started a local Ray instance.\n",
            "Loaded texify model to cuda with torch.float16 dtype\n",
            "Converting 5000 pdfs in chunk 1/1 with 6 processes, and storing in /content/marker-napierone-file2text/EPUB\n",
            "100% 4999/5000 [14:01:52<00:08,  8.66s/it]"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "out_file=\"/content/archived-marker-napierone-epub.7z\"\n",
        "!find /content/marker-napierone-file2text -type f -name \"*.json\" -print -delete # fuck off\n",
        "!7z a $out_file /content/marker-napierone-file2text"
      ],
      "metadata": {
        "id": "sU8QpnNStjFn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import files\n",
        "files.download(out_file)"
      ],
      "metadata": {
        "id": "iyzMPMmtqAzI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "CiYoo9ZEtrEz"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "L4",
      "machine_shape": "hm"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}