# Pre-historic Knowledge Assistant Web App

A web application for the RAG-based knowledge assistant.

## Features

- Multiple LLM provider support (Azure OpenAI, OpenAI, Ollama, vLLM, custom endpoints)
- Flexible embedding configuration
- Web interface with real-time responses
- REST API endpoints
- Health check endpoint

## Setup

1. Install dependencies:
   ```bash
   pip install -r requirements.txt
   ```

2. Create `.env` file from `.env.example`:
   ```bash
   cp .env.example .env
   # Edit .env with your configuration
   ```

3. Run the application:
   ```bash
   python app.py
   # or
   ./run.sh
   ```

## Configuration

### LLM Providers

The application supports multiple LLM providers:

- **Azure OpenAI**: Set `LLM_PROVIDER=azure_openai` and configure Azure credentials
- **OpenAI**: Set `LLM_PROVIDER=openai` and provide API key
- **Ollama**: Set `LLM_PROVIDER=ollama` and configure host URL
- **vLLM**: Set `LLM_PROVIDER=vllm` and configure vLLM host
- **Custom**: Set `LLM_PROVIDER=custom` for any OpenAI-compatible endpoint

### Environment Variables

See `.env.example` for all available configuration options.

## API Endpoints

- `GET /`: Web interface
- `POST /ask`: Process question via web form
- `POST /api/ask`: REST API endpoint for questions
- `GET /api/health`: Health check endpoint

## Development

To run in development mode with auto-reload:

```bash
python app.py
```

The application will be available at `http://localhost:8000`.