When should I use this Claude Code skill?

Use this skill when building AI applications that require high-precision retrieval, especially when your queries include specific names, product codes, or domain-specific jargon.

What is hybrid search in the context of RAG?

Hybrid search combines semantic vector search, which understands meaning, with traditional keyword matching (like BM25), which handles exact terms, to provide more accurate retrieval results.

What is Reciprocal Rank Fusion (RRF)?

RRF is a robust algorithm used to merge multiple ranked search results into a single list without needing to normalize the underlying scores of different search methods.

Does this support PostgreSQL?

Yes, it includes specialized templates for PostgreSQL using the pgvector extension for embeddings and built-in tsvector for full-text search.

Hybrid Search Implementation

Name: Hybrid Search Implementation
Author: EricGrill

byEricGrill

•

数据科学与机器学习

Combines vector similarity and keyword-based search to enhance retrieval accuracy in RAG systems and search engines.

关于

This skill provides standardized patterns and templates for implementing hybrid search, bridging the gap between semantic understanding and exact keyword matching. It offers ready-to-use implementations for common fusion methods like Reciprocal Rank Fusion (RRF) and linear weighting, along with database-specific configurations for PostgreSQL (pgvector) and Elasticsearch. Use this skill to improve the recall of Retrieval-Augmented Generation (RAG) applications, especially when handling domain-specific vocabulary, technical identifiers, or queries where pure vector search fails to capture exact matches.

主要功能

Normalization and weighting logic for multi-method retrieval
2 GitHub stars
Elasticsearch hybrid search configurations and RRF sub-searches
Reciprocal Rank Fusion (RRF) and linear scoring implementations
Cross-encoder reranking templates for high-precision results
PostgreSQL patterns using pgvector and Full-Text Search (FTS)

使用场景

Improving search engine relevance for mixed semantic and keyword queries
Building high-performance Retrieval-Augmented Generation (RAG) pipelines
Optimizing document retrieval for domain-specific or technical vocabulary

关于

主要功能

Normalization and weighting logic for multi-method retrieval
2 GitHub stars
Elasticsearch hybrid search configurations and RRF sub-searches
Reciprocal Rank Fusion (RRF) and linear scoring implementations
Cross-encoder reranking templates for high-precision results
PostgreSQL patterns using pgvector and Full-Text Search (FTS)

使用场景

Improving search engine relevance for mixed semantic and keyword queries
Building high-performance Retrieval-Augmented Generation (RAG) pipelines
Optimizing document retrieval for domain-specific or technical vocabulary