# ragtime-ai

**Repository Path**: bali-dalian/ragtime-ai

## Basic Information

- **Project Name**: ragtime-ai
- **Description**: demo for rag
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-08-11
- **Last Updated**: 2025-08-12

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 📖 **Ragtime-AI: A RAG (Retrieval-Augmented Generation) System**

This is a Retrieval-Augmented Generation (RAG) system built with:
- **FastAPI**
- **HuggingFace embedding models** (Qwen3-embedding-0.6B)
- **Weaviate** as the vector database
- **Deepseek** as LLM

🔖 **Features**
- File uploading
- Document indexing to vector DB
- Semantic search
- Question answering (Chat with uploaded docs)


## 📜 Requirements
- Python 3.13+ (Recommended)
  - Install the packages:
    ```
    pip install --no-cache-dir -r requirements.txt
    ```
- Weaviate 1.25.1+
  - Run the following command to start a Weaviate instance using Docker (Default port 8080):
    ```
    cd weaviate
    docker-compose up -d
    ```
  - If you want to customize your Weaviate configuration, refer to: https://docs.weaviate.io/deploy/installation-guides/docker-installation

- Set environment variables
  - Create a `.env` file in `/ragtime_ai/config/` and provide your Deepseek API key:
    ```
    DEEPSEEK_API_KEY=sk-
    ```



## 🚀 Quick Start
- Start the FastAPI endpoints:
  ```bash
  python app.py
  ```
  - The server runs at http://localhost:5000
  - You can open the API docs in your browser: http://localhost:5000/docs
- Launch the UI:
  ```bash
  streamlit run ui.py
  ```



## 🗄️ **Data Persistence**

* Uploaded files are stored on host: `./file/` (not tracked in Git)
* HuggingFace model cache: `./.cache/huggingface` (not tracked in Git)
* Weaviate data: `./weaviate_data/` (not tracked in Git)

These are mounted as Docker volumes so that data persists even when containers are stopped.



## 📄 **Configuration**

* Application config: `ragtime_ai/config/config.yaml`
* Environment variables: `ragtime_ai/config/.env` (not tracked in Git)



## 🐳 Run with Docker Compose
- Change the config file path in `ragtime_ai/config.py`:
  - Find this line:
    ```
    config_path = os.path.join(current_dir, "config", "config.yaml")
    ```
  - Change it to:
    ```
    config_path = os.path.join(current_dir, "config", "config_docker.yaml")
    ```

- Run:

  ```bash
  docker-compose up -d --build
  ```

  This will:

  - Build and start the FastAPI server (`ragtime-fastapi`)
  - Start the Weaviate vector database (`ragtime-weaviate`)


### 🌐 Access the FastAPI server

* Open the API docs in your browser:
  👉 [http://localhost:8000/docs](http://localhost:8000/docs)

### 🧪 Test Weaviate (Optional)

You can check if Weaviate is ready:

```bash
curl http://localhost:8081/v1/.well-known/ready
```


### 🧹 Stop and clean up

```bash
docker-compose down
```

If you also want to remove volumes:

```bash
docker-compose down -v
```
## 📜 **Notes**

* `.env` is gitignored. Please create your own and add required environment variables.
* `.dockerignore` excludes unnecessary files (like `file/` and `.cache/`) from the image.
* The `file/` and `weaviate_data/` directories are created automatically if not present.





## 👨‍💻 Author

> Rachel asdxl1996@163.com