ds_zagres_ai/ai_service/README.md

# OpenWebUI Backend Service

This is a backend service for OpenWebUI that provides OpenWebUI-compatible API endpoints for chat functionality and model switching.

## Features

- OpenWebUI-compatible API endpoints
- Ollama API proxy
- Chat functionality with model switching
- Support for multiple LLM models (gemma3, llama3.3, llama3.1, mistral, deepseek)

## Project Structure

```
ai_service/
├── models/           # Model and chat services
│   ├── model_service.py
│   ├── chat_service.py
│   └── model_parameters.py
├── embeddings/       # Document processing for RAG
│   └── document_service.py
├── openwebui_api.py  # OpenWebUI-compatible API endpoints
├── config.py         # Configuration settings
├── api.py            # FastAPI application
└── deploy.sh         # Deployment script
```

## Setup Instructions

1. Create a virtual environment:
   ```
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   ```

2. Install dependencies:
   ```
   pip install -r requirements.txt
   ```

3. Copy `.env.example` to `.env` and update the values:
   ```
   cp .env.example .env
   # Edit the .env file with appropriate values
   ```

4. Run the service:
   ```
   python ../run_ai_service.py
   ```

## API Documentation

Once the service is running, you can access the API documentation at:
- Swagger UI: http://localhost:5252/docs
- ReDoc: http://localhost:5252/redoc

## Deployment

To deploy the service:

1. Make the deployment script executable:
   ```
   chmod +x deploy.sh
   ```

2. Run the deployment script:
   ```
   ./deploy.sh
   ```

This will start the service on port 5252 using uvicorn with nohup.

## API Endpoints

### Health Check

- `GET /health` - Check if the service is running

### OpenWebUI-Compatible Endpoints

- `GET /api/models` - Get available models in OpenWebUI format
- `POST /api/chat/completions` - OpenAI-compatible chat completions endpoint

### Ollama API Proxy

- `POST /ollama/api/generate` - Proxy to Ollama's generate endpoint

### Original API Endpoints

- `POST /chats` - Create a new chat
- `GET /chats/user/{user_id}` - Get all chats for a user
- `GET /chats/{chat_id}` - Get a chat by ID
- `POST /chats/{chat_id}/messages` - Send a message to a chat
- `POST /chats/{chat_id}/members/{user_id}` - Add a user to a team chat
- `DELETE /chats/{chat_id}/members/{user_id}` - Remove a user from a team chat
- `DELETE /chats/{chat_id}` - Delete a chat