Document Synchronizer

A powerful tool for maintaining multi-language documentation consistency. This project provides utilities to ensure your documentation in different languages stays synchronized and up-to-date.

This tool is an agent-based multilingual document synchronization system built on the MetaGPT framework. It leverages LLM-powered agents to intelligently analyze, compare, and translate documentation across multiple languages, ensuring consistency while maintaining the technical accuracy and nuance of specialized content.

Features

• Document Structure Checking: Identifies missing files across language directories
• Content Comparison: Detects differences between document versions in different languages
• Automatic Translation: Generates translations for missing documents
• Translation Improvement: Updates existing translations to match the source content
• Multi-Agent System: Advanced option with separate roles for checking and translating
• Dry-Run Mode: Preview changes without modifying files
• Comprehensive Logging: Console and file-based logging

Installation

Prerequisites

• Python 3.9+ (required by MetaGPT)
• MetaGPT library
• Internet connection for API access

Note: This code was developed and tested using Python 3.11.

Setup

1. Clone this repository:

   git clone https://github.com/gly11/Document-Synchronizer.git
   cd Document-Synchronizer

2. Install MetaGPT following the official installation guide

3. Configure API access according to MetaGPT documentation

4. Install additional dependencies:

   pip install colorama

   Note: colorama is used for better log display and terminal output formatting

Usage

Basic Document Maintainer

The doc_maintainer.py script provides a straightforward approach to document maintenance:

python doc_maintainer.py --path ./docs --langs en,zh,es --primary en

Multi-Agent Document Maintainer

The multi_agent_doc_maintainer.py script offers an advanced multi-agent approach:

python multi_agent_doc_maintainer.py --path ./docs --langs en,zh,es --primary en

Command Line Options

Both scripts support the following options:

Option	Description	Default
-p, --path	Base path for documentation	./examples
-l, --langs	Comma-separated language directories	en,zh
-m, --primary	Primary language for comparisons	en
-v, --verbose	Display verbose output	False
-d, --dry-run	Dry run mode, don't modify files	False

How It Works

1. Scanning: The tool scans all language directories for markdown files
2. Structure Analysis: Identifies which files are missing in each language
3. Content Comparison: For files that exist in multiple languages, compares content for differences
4. Translation: Generates translations for missing files or outdated content
5. Reporting: Provides detailed statistics about the documentation state

Example Workflow

# Check documentation structure and preview changes
python doc_maintainer.py --path ./project-docs --langs en,zh,ja,fr --primary en --dry-run

# Apply changes to synchronize all documentation
python doc_maintainer.py --path ./project-docs --langs en,zh,ja,fr --primary en

Choosing Between Scripts

• doc_maintainer.py - Simpler implementation, good for most use cases
• multi_agent_doc_maintainer.py - More advanced with separate agent roles, better for complex documentation

Logs

Both tools generate logs in their respective log files:

• doc_maintainer.log
• multi_agent_doc_maintainer.log

These provide detailed information about the operations performed during execution.

License

MIT License