Vector Store: LanceDB

LanceDB is an open-source, serverless vector database for AI that stores data locally or on cloud object stores (like AWS S3). It requires no infrastructure and provides sub-millisecond search latencies.

1. Configuration Parameters

Specify LanceDB options inside vector_store_params:

Parameter	Type	Default	Description
`vector_store_type`	`string`	`"LanceDBVectorStore"`	Identifier of the LanceDB vector store class.
`uri`	`string`	`"/tmp/lancedb"`	Absolute or relative directory path where the database files are saved.
`table_name`	`string`	`"aih_embeddings"`	Table name within LanceDB for storing your data.

2. Configuration Example

from aih import ChatBot
 
lancedb_config = {
    "vector_store_type": "LanceDBVectorStore",
    "uri": "./local_data/lancedb_store",
    "table_name": "project_notes"
}
 
chatbot = ChatBot.txt_chat(
    input_files=["./notes.txt"],
    vector_store_params=lancedb_config
)

Azure Cognitive Search Large Language Models