LLM Translator (English to Chinese)

LLM翻译（英汉）

A web-based application that uses LLMs (Large Language Models) to translate and summarize documents from English to Chinese. Supports multiple file formats including PDF, DOCX, TXT.

一个基于网络的应用程序，使用LLMs （大型语言模型）将文档从英语翻译和总结为中文。支持多种文件格式，包括 PDF、DOCX、TXT 。

Features

• Document translation from English to Chinese
• Document summarization
• Support for multiple file formats (PDF, DOCX, TXT)
• PDF format is for academic papers specifically
• Web-based interface using PyWebIO
• Uses Ollama for local LLM processing

Prerequisites

• Python 3.8+
• Ollama installed and running locally
• Required Python packages (see requirements.txt)

Installation

1. Clone the repository:

git clone https://github.com/theohlong/LLM_Translator_en2cn.git
cd LLM_Translator_en2cn

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows use: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Install and start Ollama: Follow instructions at Ollama's official documentation

Usage

0. Make sure the Ollama app is installed and in running state

1. Start the application:

cd src
python app.py

2. Open your web browser and navigate to:

http://localhost:8501

3. Choose the processing mode:

   • EXTRACT: Summarize each passage
   • TRANSLATE: Translate each passage

4. Upload your document and wait for processing

Configuration

The application uses several configuration files:

• src/config.py: Main application configuration
• Environment variables (create a .env file):

  OLLAMA_HOST=http://localhost:11434
  MAX_FILE_SIZE=100M
  DEBUG=True
  

Project Structure

LLM_Translator_en2cn/
├── src/
│   ├── models/         # LLM and text processing models
│   ├── services/       # Core services (translation, summarization)
│   ├── utils/          # Utility functions
│   ├── app.py         # Main application
│   └── config.py      # Configuration
├── tests/             # Test files
├── requirements.txt   # Project dependencies
└── README.md         # This file

License

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

Acknowledgments

• Ollama for providing the local LLM capability
• PyWebIO for the web interface framework
• This project was inspired by and incorporates code from Andrew Ng's translation agent (https://github.com/andrewyng/translation-agent)
• PDF processing utilizes the Allen Institute's Paper-to-HTML conversion tool (https://papertohtml.org/)