mistral-pdf-to-markdown 1.2.0

A CLI tool to convert PDF files to Markdown using the Mistral AI OCR API.

Mistral PDF to Markdown Converter

A simple command-line tool to convert PDF and EPUB files into Markdown format using the Mistral AI OCR API. This tool also extracts embedded images and saves them in a subdirectory relative to the output markdown file.

Installation

You can install the package directly from PyPI using pip:

pip install mistral-pdf-to-markdown

Global Installation (Recommended for CLI Usage)

If you want to use the pdf2md command from anywhere in your system without activating a specific virtual environment, the recommended way is to use pipx:

1. Install pipx (if you don't have it already). Follow the official pipx installation guide. A common method is:

   python3 -m pip install --user pipx
   python3 -m pipx ensurepath
   

   (Restart your terminal after running ensurepath)

2. Install the package using pipx:

   pipx install mistral-pdf-to-markdown
   

This installs the package in an isolated environment but makes the pdf2md command globally available.

Installation from Source

Alternatively, if you want to install from the source:

1. Clone the repository:

   git clone https://github.com/arcangelo7/mistral-pdf-to-markdown.git
   cd mistral-pdf-to-markdown
   

2. Install dependencies using Poetry:

   poetry install
   

Additional Requirements for EPUB Support

To convert EPUB files, you need to install pandoc. See the official installation guide for your operating system.

Usage

1. Set your Mistral API Key: You can set your API key as an environment variable:

   export MISTRAL_API_KEY='your_api_key_here'
   

   Alternatively, you can create a .env file in the project root directory with the following content:

   MISTRAL_API_KEY=your_api_key_here
   

   You can also pass the API key directly using the --api-key option.

2. Run the conversion:

   Convert a Single PDF or EPUB File

   The convert command processes a single PDF or EPUB file.

   poetry run pdf2md convert <path/to/your/document.pdf> [options]
   

   Or, if you have activated the virtual environment (poetry shell):

   pdf2md convert <path/to/your/document.pdf> [options]
   

   Options for Single File Conversion:

   • --output or -o: Specify the path for the output Markdown file. If not provided, it defaults to the same name as the input file but with a .md extension (e.g., document.md).
   • --api-key: Provide the Mistral API key directly.

   Convert Multiple PDF and EPUB Files from a Directory

   The convert-dir command processes all PDF and EPUB files in a specified directory.

   poetry run pdf2md convert-dir <path/to/directory/with/files> [options]
   

   Or, if you have activated the virtual environment (poetry shell):

   pdf2md convert-dir <path/to/directory/with/files> [options]
   

   Options for Directory Conversion:

   • --output-dir or -o: Specify the directory where output Markdown files will be saved. If not provided, it defaults to the same directory as the input files.
   • --api-key: Provide the Mistral API key directly.
   • --max-workers or -w: Maximum number of concurrent conversions (default: 2). Increase this value to process multiple files in parallel for faster conversion.

Image Handling:

The script will attempt to extract images embedded in the document.

• Images are saved in a subdirectory named <output_filename_stem>_images (e.g., if the output is report.md, images will be in report_images/).
• The generated Markdown file will contain relative links pointing to the images in this subdirectory.

Examples:

# Convert a single PDF file (output: ./my_report.md)
poetry run pdf2md convert ./my_report.pdf

# Convert with custom output path
poetry run pdf2md convert ./my_report.pdf -o ./output/report.md

# Convert all files in a directory with 4 concurrent workers
poetry run pdf2md convert-dir ./documents/ -o ./markdown_output/ -w 4

An example output generated from example.pdf (included in the repository) can be found in example.md, with its corresponding images located in the example_images/ directory.

License

This project is licensed under the ISC License - see the LICENSE file for details.