Skip to content

Latest commit

 

History

History
85 lines (69 loc) · 3.64 KB

helper.md

File metadata and controls

85 lines (69 loc) · 3.64 KB
Raw

helper.py

Table of Contents

Script Overview

This script is designed to initialize and return a PGVector instance that interacts with a PostgreSQL database to store and retrieve vectorized document embeddings.

Import Libraries

  • logging: Used for logging script actions and errors.
  • psycopg2: For interacting with PostgreSQL databases.
  • BedrockEmbeddings: LangChain community embeddings instance for handling document embeddings. This project uses the Amazon Titan Text Embeddings V2 model to generate embeddings.
  • PGVector: PostgreSQL-based vector store for storing and retrieving vectorized documents.

Helper Functions

  • get_vectorstore: Initializes and returns a PGVector instance connected to the PostgreSQL database. It handles connection setup and error logging.

Execution Flow

  1. Database Connection: A connection string is built based on the provided database credentials.
  2. Vector Store Initialization: A PGVector instance is initialized using the connection string and collection name.
  3. Logging: The script logs successful initialization and error messages if any occur.

Detailed Function Descriptions

Function: get_vectorstore 

def get_vectorstore(
    collection_name: str, 
    embeddings: BedrockEmbeddings, 
    dbname: str, 
    user: str, 
    password: str, 
    host: str, 
    port: int
) -> Optional[PGVector]:
    try:
        connection_string = (
            f"postgresql+psycopg://{user}:{password}@{host}:{port}/{dbname}"
        )

        logger.info("Initializing the VectorStore")
        vectorstore = PGVector(
            embeddings=embeddings,
            collection_name=collection_name,
            connection=connection_string,
            use_jsonb=True
        )

        logger.info("VectorStore initialized")
        return vectorstore, connection_string

    except Exception as e:
        logger.error(f"Error initializing vector store: {e}")
        return None

Purpose

Initializes and returns a PGVector instance that connects to a PostgreSQL database and prepares a vector store for storing embedded document data.

Process Flow

  1. Database Connection Setup: Creates a connection string using the provided database credentials and logs successful connections or errors.
  2. Vector Store Initialization: Constructs a PGVector instance using the connection string, collection name, and embeddings. If successful, returns the initialized PGVector instance along with the connection string.
  3. Error Handling: Captures and logs any errors that occur during vector store initialization.

Inputs and Outputs

  • Inputs:

    • collection_name: The name of the collection in the vector store.
    • embeddings: The BedrockEmbeddings instance for creating embeddings.
    • dbname: Database name.
    • user: Database user.
    • password: Database password.
    • host: Host for the PostgreSQL database.
    • port: Port for the PostgreSQL database.

  • Outputs:

    • Returns the initialized PGVector instance and the connection string if successful.
    • Returns None if an error occurred during setup.

🔼 Back to top