[UNDER CONSTRUCTION]

Marovi Translation

A tool for translating PDF documents to Spanish in Markdown format, specifically tailored for NeurIPS papers.

Note

This project is currently undergoing a revamp to create a beta version for translating any PDF locally. Core features under development include:

• Incorporating GROBID for PDF parsing
• Integrating the Google Translate API (the current googletrans version uses an outdated API)
• Adding ChatGPT for agentic translation
• Refactoring and introducing new features to prepare the CLI for pip

Please reach out if you are interested in contributing or working on any of these features.

The video presentation for the first version can be found here https://youtu.be/r2CcKOnehs4

Installation

Prerequisites

• Anaconda or Miniconda (recommended for managing dependencies and environments)
• Python 3.12.0

Setup

1. Clone the Repository:

   git clone [email protected]:felipefelixarias/MaroviTranslation.git
   cd MaroviTranslation

2. Create and Activate Conda Environment:

   conda create --name your_env_name python=3.12.0
   conda activate your_env_name

3. Install Dependencies:

   pip install .

   This will install all the dependencies listed in requirements.txt.

Usage

1. PDF to Markdown Conversion: A working example is available in MaroviTranslation/tutorial/NeurIPSExample.py. Run it with the following command.

    python MaroviTranslation/tutorial/NeurIPSExample.py

   The markdowns will be saved in MaroviTranslation/outputs/ More generally, use the NeurIPSPDFToSpanishMarkdown class to convert a PDF file into a Spanish Markdown file. Example usage:

    from MaroviTranslation.converters.NeurIPS import NeurIPSPDFToSpanishMarkdown
    from MaroviTranslation.translation.core import Translator
    from MaroviTranslation.translation.GoogleTranslator import GoogleTranslator
   
    # Initialize translator and converter
    translator = Translator()
    translator.set_translator(GoogleTranslator())
    converter = NeurIPSPDFToSpanishMarkdown("path_to_pdf", "path_to_output_folder", translator)
   
    # Parse PDF, create image map, and generate Markdown
    converter.parse_pdf()
    converter.create_image_map()
    converter.generate_markdown()

   This will generate Markdown files with translated content and images in the specified output directory.

Project Structure

• MaroviTranslation/
  • converters/: Modules for parsing, translation, and markdown generation.
    • NeurIPS.py: Class for converting NeurIPS PDF to Spanish markdowns.
  • markdown/: Modules for handling Markdown generation.
    • core.py: Core functionalities for Markdown manipulation.
  • outputs/: Output directory for generated Markdown files and images.
  • parsing/: Modules for parsing PDF files.
    • core.py: Core pasrsing class.
    • NeurIPSParser.py: Parser specific for NeurIPS papers.
  • pdfs/: Directory to place PDF files for conversion.
  • translation/: Translation modules.
    • core.py: Core translation class.
    • GoogleTranslator.py: Google Translator API wrapper.
  • tutorial/ Tutorial script.
    • NeurIPSExample.py: Script with example usage.
• requirements.txt: List of project dependencies.
• setup.py: Setup script for installing the project.
• README.md: Documentation for the project (this file).