High-Performance RAG Solution with Pgvectorscale and Python

This project demonstrates setting up and using pgvectorscale with Docker and Python, leveraging Gemini's gemini-1.5-flash model for embeddings. It combines advanced retrieval techniques with intelligent answer generation based on the retrieved context, using PostgreSQL as a vector database.

Pgvectorscale Documentation

For more information about using PostgreSQL as a vector database in AI applications with Timescale, check out these resources:

• GitHub Repository: pgvectorscale
• Blog Post: PostgreSQL and Pgvector: Now Faster Than Pinecone, 75% Cheaper, and 100% Open Source
• Blog Post: RAG Is More Than Just Vector Search
• Blog Post: A Python Library for Using PostgreSQL as a Vector Database in AI Applications

Why PostgreSQL?

Using PostgreSQL with pgvectorscale as your vector database offers several key advantages:

• PostgreSQL is a robust, open-source database with a rich ecosystem of tools, drivers, and connectors.
• Manage both relational and vector data within a single database, reducing operational complexity.
• Pgvectorscale enhances pgvector with faster search capabilities, higher recall, and efficient time-based filtering.

Prerequisites

• Docker
• Python 3.7+
• Gemini API key
• PostgreSQL GUI client

Steps

1. Set up Docker environment
2. Connect to the database using a PostgreSQL GUI client
3. In docker db, run command CREATE EXTENSION IF NOT EXISTS vectorscale CASCADE;
4. Insert document chunks as vectors using embeddings
5. Perform similarity search

Usage

1. Create a copy of example.env and rename it to .env
2. Fill in your Gemini API key in .env
3. Run the Docker container
4. Install the required Python packages using pip install -r requirements.txt
5. Execute insert_vectors.py to populate the database
6. Use similarity_search.py to perform similarity searches

Using ANN Search Indexes

Timescale Vector offers indexing options to accelerate similarity queries, particularly beneficial for large vector datasets:

1. Supported indexes:

   • timescale_vector_index (default): A DiskANN-inspired graph index
   • pgvector's HNSW: Hierarchical Navigable Small World graph index
   • pgvector's IVFFLAT: Inverted file index

2. The DiskANN-inspired index provides improved performance. Refer to the Timescale Vector explainer blog for detailed information and benchmarks.

For optimal query performance, creating an index on the embedding column is recommended, especially for large vector datasets.

Cosine Similarity in Vector Search

Cosine similarity measures the cosine of the angle between two vectors in a multi-dimensional space. It's a measure of orientation rather than magnitude.

• Range: -1 to 1 (for normalized vectors)
• 1: Vectors point in the same direction (most similar)
• 0: Vectors are orthogonal (unrelated)
• -1: Vectors point in opposite directions (most dissimilar)

Cosine Distance

In pgvector, the <=> operator computes cosine distance, which is 1 - cosine similarity.

• Range: 0 to 2
• 0: Identical vectors (most similar)
• 1: Orthogonal vectors
• 2: Opposite vectors (most dissimilar)

Interpreting Results

• Lower distance values indicate higher similarity.
• A distance of 0 would mean an exact match.
• Distances closer to 0 indicate high similarity.
• Distances around 1 suggest little to no similarity.
• Distances approaching 2 indicate opposite meanings.