PDF Translation Tool

A professional PDF translation tool based on PDFMathTranslate-next and BabelDOC that converts PDF documents to bilingual (dual-language) versions while preserving original formatting including formulas, tables, and graphics.

Features

• Format Preservation: Maintains original layout, formulas, tables, and graphics
• Bilingual Output: Generates both dual-language (bilingual) and single-language (translation only) PDFs
• Multiple Translation Engines: Supports various translation providers including ZhipuAI, VolcEngine, SiliconFlow, and OpenAI
• Term Extraction: Automatically extracts and generates glossary files
• Batch Processing: Supports batch translation of multiple PDF files
• Customizable: Supports custom system prompts and translation parameters
• Page Range Selection: Translate specific pages or ranges
• High Performance: Configurable QPS (queries per second) for faster translation

Quick Start

Installation

# Clone the repository
git clone https://github.com/boyingliu01/pdf-translation.git
cd pdf-translation

# Install dependencies
pip install -r requirements.txt

Configuration

Copy and edit a configuration template based on your preferred translation provider:

# Using ZhipuAI (recommended, free tier available)
cp config/config.zhipu.json config/config.json
# Edit config/config.json and replace "your-zhipuai-api-key-here" with your actual API key

# Using VolcEngine
cp config/config.volcengine.json config/config.json
# Edit config/config.json and replace "your-volcengine-api-key-here" with your actual API key

# Using SiliconFlow
cp config/config.siliconflow.json config/config.json
# Edit config/config.json and replace "your-siliconflow-api-key-here" with your actual API key

# Using OpenAI
cp config/config.openai.json config/config.json
# Edit config/config.json and replace "your-api-key-here" with your actual API key

API Key Registration:

• ZhipuAI: https://open.bigmodel.cn/ - Free tier available, recommended
• VolcEngine: https://console.volcengine.com/ark - Requires subscription
• SiliconFlow: https://siliconflow.cn/ - Various models available
• OpenAI: https://platform.openai.com/ - Pay-as-you-go

Translation

# Basic translation
python translate_pdf.py -i path/to/document.pdf -o path/to/output

# With language specification
python translate_pdf.py -i document.pdf -o output --lang-in en --lang-out zh

# Specific pages only
python translate_pdf.py -i document.pdf -o output --pages "1-5"

# Without dual-language output
python translate_pdf.py -i document.pdf -o output --no-dual

Usage

Command Line Arguments

Argument	Short	Description	Default
--input	-i	Input PDF file path (required)	-
--output	-o	Output directory	PDF parent directory
--config	-c	Configuration file path	config/config.json
--lang-in	-li	Source language code	en
--lang-out	-lo	Target language code	zh
--no-dual	-	Skip dual-language PDF output	false
--no-mono	-	Skip single-language PDF output	false
--watermark	-	Watermark mode: watermarked/no_watermark/both	watermarked
--pages	-	Page range: 1,2,1-,-3,3-5	All pages
--max-pages-per-part	-	Max pages per batch for large documents	-
--enhance-compatibility	-	Enable compatibility enhancements	false
--create-config	-	Create example config file	-

Configuration Options

Option	Description	Default
translation_engine	Translation engine type	openai
openai_api_key	API key for the translation service	-
openai_base_url	API base URL	-
openai_model	Model name	glm-4-flash (ZhipuAI)
qps	Queries per second (higher = faster)	4
min_text_length	Minimum text length for translation	5
debug	Enable debug mode	false
custom_system_prompt	Custom system prompt	null
enable_term_extraction	Enable term extraction/glossary	false
page_markdown	Enable page markdown	false

Translation Engines

ZhipuAI (Recommended)

Uses the GLM-4-Flash model from ZhipuAI. Offers a generous free tier with high translation quality.

Registration: https://open.bigmodel.cn/

VolcEngine

Uses Doubao models from VolcEngine. Requires a subscription plan.

Registration: https://console.volcengine.com/ark

SiliconFlow

Supports various models including DeepSeek.

Registration: https://siliconflow.cn/

OpenAI

Uses native OpenAI models like GPT-4o-mini.

Registration: https://platform.openai.com/

Output Files

Translation generates the following files:

• <filename>.zh-CN.dual.pdf - Bilingual PDF (recommended)
• <filename>.zh-CN.mono.pdf - Single-language PDF (translation only)
• <filename>.zh-CN.dual.no-watermark.pdf - Unwatermarked dual (if watermark mode is set to both or no_watermark)
• <filename>.zh-CN.mono.no-watermark.pdf - Unwatermarked mono (if watermark mode is set to both or no_watermark)
• <filename>.zh-CN.glossary.csv - Glossary file (if term extraction enabled)

Performance Tips

1. Increase QPS: Set a higher qps value in the configuration file for faster translation
2. Use faster models: Models like glm-4-flash or doubao-pro-4k offer good speed-quality balance
3. Batch processing: Use --max-pages-per-part for very large documents
4. Page range selection: Translate specific pages first to test quality before full translation

Documentation

For more detailed information, see the docs/ directory:

• QUICKSTART.md - Quick start guide
• USAGE_GUIDE.md - Detailed usage guide
• ENGINES_GUIDE.md - Translation engine configuration guide

FAQ

Q: The translation speed is slow. What can I do?

A: Increase the qps value in your configuration file, use a faster model (like glm-4-flash), or use the --max-pages-per-part option for batch processing.

Q: How can I improve translation quality?

A: Use a higher-quality model (like gpt-4o), add a custom system prompt, or enable term extraction to provide context.

Q: Which languages are supported?

A: The tool supports translation between all major language pairs, including Chinese-English, English-Japanese, English-French, and more.

Q: Can I use my own translation API?

A: Yes, as long as your API is OpenAI-compatible, you can configure it by setting the appropriate openai_base_url and openai_model in the configuration file.

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

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

Acknowledgments

This project is built upon:

• PDFMathTranslate-next - Core translation library
• BabelDOC - PDF parsing and layout analysis

Contact

For questions or support, please open an issue on GitHub.