Python SDK for VectorDB
Currently, it can embed documents using Langchain document loader and huggingface embeddings then search through. See here
The above file is having this issue while searching
ERROR:__main__:Error during search: [WinError 10061] No connection could be made because the target machine actively refused it
Fixing it asap!!
- Use Type Annotations and Data Validations
- Test and fix the langchain
CustomVectorDBVectorStore - Implement
as_retriever()feature. - Search threshold for more relevant results
- Logging
- API Keys Middleware
- Handle edge cases
- Refine the examples
- Implement pagination for search results
- Delete document endpoint
- Allow to add similar documents in different collections
- Metadata filtering inside a collection docs
- Test Async client and examples. See here
Implement CustomVectorStore for compatibility with langchain, langgraph etc. (Currently, not working)