Vectorstore Database Manager

Info

This is a Flask server providing a RAG (Retrieval Augmented Generation) vector store backend. It manages multiple vectorstores, supports semantic search queries, and offers a web interface for testing and inspection.

Features: Semantic similarity search, chunk viewing, store statistics, query testing, and PDF data inspection.

Available Stores

API

Description

This module supplies a Python interface to the vector store API. It wraps each API call in a respective Python function.

Usage

from store_api.vectorstore_api.cli_cls import VsAPICli

# create instance of the API
api = VsAPICli(address='url_to_api')

# ping the server
ping_result : str = api.ping()

# fetch available stores
stores : list[str] = api.fetch_available_stores()

# query a specific store
query = 'query_string'
store_name = 'store_name'
num_chunks = 5  # optional, number of chunks to return
answer : list[dict] = api.query(
    message=query,
    store=store_name,
    score=True,
    num_of_chunks=num_chunks
)

# Answer structure: List of dictionaries with format:
# {"page_content": ,
#  "metadata": {:, ..., "score":}}