README.md

metadata

title: ELAN-Bot
emoji: 🤖
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: mit

ELAN-Bot 🤖

A virtual assistant designed to help users with the ELAN annotation software. The bot can answer questions about ELAN usage and modify EAF (ELAN Annotation Format) files based on user instructions.

Features

• Software Usage Help: Ask questions about ELAN features and functionality
• XML/EAF File Modification: Provide EAF file content with instructions to modify it
• Multilingual Support: Available in English, Spanish, Italian, French, and German
• Vector Search: Uses semantic search to find relevant information from documentation
• Powered by Llama 3.3 70B: Advanced language model for accurate responses

Usage

Simply interact with the chat interface:

• Ask questions: "How can I add a new tier in ELAN?"
• Modify EAF files: Paste your EAF content with instructions at the beginning like:

  instructions: change the participant name from Eleonora to Gianni
  
  <?xml version="1.0" encoding="UTF-8"?>
  <ANNOTATION_DOCUMENT...>
  

Examples

Try these sample questions:

• "How can I add a new tier in ELAN?"
• "¿Cómo puedo exportar anotaciones en formato txt?"
• "Come posso cercare all'interno delle annotazioni?"

Configuration

The app requires a HF_TOKEN environment variable to be set in the Hugging Face Spaces settings for accessing the Llama model.

Technical Details

• Backend: Python with Gradio interface
• Vector Search: Qdrant + SentenceTransformers
• LLM: Meta Llama 3.3 70B Instruct via Hugging Face Inference API
• Text Processing: tiktoken for efficient chunking

Project Structure

elan-bot/
├── app.py                          # Main application entry point
├── requirements.txt                # Python dependencies
├── config/
│   └── settings.py                # Configuration settings
├── prompts/                       # Organized prompt templates
├── services/                      # Core business logic
├── utils/                         # Utility functions
├── ui/                           # Gradio interface components
└── data/                         # Vector database storage