pgvector-perl

Perl client and examples for using the pgvector PostgreSQL extension, enabling vector data types and similarity queries from Perl applications.

• Category: SDKs & Libraries
• Language: Perl
• Ecosystem: PostgreSQL, pgvector
• License: MIT
• Source: https://github.com/pgvector/pgvector-perl

Overview

pgvector-perl provides example code and patterns for integrating the pgvector PostgreSQL extension with Perl via DBD::Pg. It demonstrates how to store and query vector embeddings, perform similarity search, and implement hybrid and sparse retrieval workflows from Perl scripts.

Features

• pgvector integration for Perl

  • Shows how to use the vector column type provided by the pgvector extension in PostgreSQL.
  • Demonstrates creating tables with vector(dim) columns.
  • Example: CREATE TABLE items (id bigserial PRIMARY KEY, embedding vector(3));

• DBD::Pg support

  • Designed to work with the DBD::Pg Perl database driver.
  • Uses standard DBI/DBD::Pg patterns for connecting to PostgreSQL, preparing statements, and executing queries.

• Extension setup

  • Example code for enabling the pgvector extension in a PostgreSQL database:
    • CREATE EXTENSION IF NOT EXISTS vector;

• Vector insertion examples

  • Demonstrates how to construct vector literals and insert them as parameters via prepared statements.
  • Example patterns for inserting multiple vector rows into a vector column.

• Similarity search and retrieval use cases
  The repository contains end-to-end examples of embedding generation and querying for common retrieval tasks:

  • Embeddings search with OpenAI
    • Example script (examples/openai/example.pl) that generates embeddings with OpenAI and stores/queries them in PostgreSQL using pgvector.
  • Binary embeddings with Cohere
    • Example script (examples/cohere/example.pl) for using Cohere’s binary embeddings with pgvector.
  • Hybrid search with Ollama
    • Example script (examples/hybrid/example.pl) illustrating hybrid search using Reciprocal Rank Fusion (RRF) to combine vector and other signals.
  • Sparse search with Text Embeddings Inference
    • Example script (examples/sparse/example.pl) showing sparse retrieval workflows and how to integrate scores into PostgreSQL queries.

• Standalone example script

  • example.pl at the repository root demonstrates a minimal setup:
    • Connecting to PostgreSQL.
    • Enabling the vector extension.
    • Creating a table with a vector column.
    • Inserting embeddings.
    • Running basic similarity queries (via SQL using pgvector operators/functions).

• Automation / CI configuration

  • GitHub Actions workflow configuration under .github/workflows (e.g., for tests or checks) to ensure examples remain functional.

Usage

• Install and configure PostgreSQL with the pgvector extension.
• Use DBD::Pg from Perl to connect to your PostgreSQL database.
• Enable the extension in your database (CREATE EXTENSION IF NOT EXISTS vector;).
• Define tables with vector(dim) columns and use the provided examples as templates for inserting embeddings and performing similarity, hybrid, or sparse search queries.

Pricing

• Open-source under the MIT license; no pricing or plans are specified in the repository.