Advanced Qdrant Query Interface with Streamlit

This Streamlit application provides an advanced interface for querying a Qdrant vector database. It allows users to perform semantic searches with optional filters using either OpenAI embeddings with future plans to incorporate Qdrant's built-in FastEmbed model.

Features

• Embedding Options: Choose between OpenAI embeddings for query vectorization. Qdrant's FastEmbed will be implemented in the near future
• Environment Variable Support: Utilize environment variables for API keys and URLs.
• Dynamic Filter Interface: Create complex filters (must, must_not, should) based on metadata fields.
• Collection Management: Select from available collections in your Qdrant database.
• Result Exploration: View and expand search results with payload details.

Demo

Installation

1. Clone the Repository

   git clone https://github.com/JosefButts/qdrant-ui.git
   cd qdrant-ui

2. Create a Virtual Environment

   python3 -m venv venv
   source venv/bin/activate  # On Windows use 'venv\Scripts\activate'

3. Install Dependencies

   pip install -r requirements.txt

Configuration

Environment Variables

Set the following environment variables if you prefer to use them:

• OPENAI_API_KEY: Your OpenAI API key (required if using OpenAI embeddings).
• QDRANT_URL: The URL of your Qdrant server.
• QDRANT_KEY: Your Qdrant API key (if required).

You can set them in your terminal:

export OPENAI_API_KEY='your-openai-api-key'
export QDRANT_URL='your-qdrant-url'
export QDRANT_KEY='your-qdrant-api-key'

Streamlit Sidebar Configuration

Alternatively, you can enter your API keys and URLs directly in the Streamlit app via the sidebar configuration.

Usage

1. Run the Application

   streamlit run src/app.py

2. Configure API Settings

   • Use the sidebar to select the embedding provider (OpenAI or FastEmbed).
   • Choose whether to use environment variables or custom settings.
   • Enter your API keys and URLs if not using environment variables.

3. Query Qdrant

   • Enter your query text.
   • Select the collection you want to search.
   • Set the result limit.
   • Create filters using the dynamic filter interface:
     • Must Filters: Conditions that must be met.
     • Must Not Filters: Conditions that must not be met.
     • Should Filters: Conditions that should be met (optional).
   • Click on "Query Qdrant" to execute the search.

4. View Results

   • Expand each result to view the payload and other details.
   • If no results are found, adjust your query or filters.

Requirements

• Python 3.7 or higher
• Streamlit
• LangChain Community Embeddings
• Qdrant Client
• OpenAI API key (if using OpenAI embeddings)
• Access to a Qdrant server with your data indexed

Dependencies

All required Python packages are listed in requirements.txt.

Notes

• Metadata Keys: The filter interface dynamically loads metadata keys from your selected collection.

License

This project is licensed under the MIT License.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any enhancements or bug fixes.

Acknowledgments

• Streamlit for the interactive web interface.
• Qdrant for the vector database and search capabilities.
• OpenAI for their powerful embedding models.
• LangChain for community embeddings support.

Contact

For any questions or suggestions, please contact buttsjosef@gmail.com.