haive.core.engine.vectorstore¶

Vector store module for the Haive framework.

This module provides a comprehensive interface for working with vector databases in the Haive framework. It includes configuration models, utility functions, and abstractions for creating and interacting with vector stores through a unified API.

Key components: - VectorStoreConfig: Main configuration class for vector stores - VectorStoreProvider: Enumeration of supported vector store providers - Utility functions for creating vector stores and retrievers

The vector store system supports various backends (FAISS, Chroma, Pinecone, etc.) and provides a consistent interface for embedding, storing, and retrieving documents using vector similarity.

Examples

>>> from haive.core.engine.vectorstore import (
...     VectorStoreConfig,
...     VectorStoreProvider,
...     create_vs_from_documents
... )
>>> from langchain_core.documents import Document
>>>
>>> # Create documents
>>> documents = [
...     Document(page_content="Apple iPhone 13 with A15 Bionic chip"),
...     Document(page_content="Samsung Galaxy S21 with Exynos processor")
... ]
>>>
>>> # Create a vector store directly
>>> vectorstore = create_vs_from_documents(
...     documents,
...     vector_store_provider=VectorStoreProvider.FAISS
... )
>>>
>>> # Search for similar documents
>>> results = vectorstore.similarity_search("smartphone with fast processor")

haive.core.engine.vectorstore¶

Submodules¶