feat: Support for vector search with Qdrant

[Anush008]

Description

This PR adds an alternative vector search backend to PixelRAG.

Qdrant is an open-source vector search engine built for high performance and large-scale workloads. Users can opt into Qdrant to take advantage of quantization, disk-backed vectors and more.

The Qdrant client dependency is installed on demand(pip install 'pixelrag[qdrant]). FAISS remains the default backend, so existing users will see no change in behavior.

The PR introduces a class VectorBackend(Protocol) abstraction. It enables any provider like FAISS and Qdrant to power PixelRAG's vector search.

Setup

You can run Qdrant with

docker run -p 6333:6333 qdrant/qdrant

Then follow the instructions in the README.md. The Qdrant dashboard will be accessible at http://localhost:6333/dashboard.

Testing

I've tested the implementation against a local Qdrant instance.

[vercel]

@Anush008 is attempting to deploy a commit to the andylizf's projects Team on Vercel.

A member of the Team first needs to authorize it.