skills.rest
aliyun
aliyunOFFICIAL
Explore all skills by this author →
WebsiteGitHub

hologres-knowledge-base

Build Hologres knowledge bases with vector and full-text indexes.

17|Updated Apr 22, 2026
Search Tags:#rag#search#knowledge#vector#knowledge-base#fulltext#hologres
One-click install
npx skills add https://github.com/aliyun/hologres-ai-plugins --skill hologres-knowledge-base
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hologres-knowledge-base
Source: https://github.com/aliyun/hologres-ai-plugins/tree/main/agent-skills/skills/hologres-knowledge-base
Command: npx skills add https://github.com/aliyun/hologres-ai-plugins --skill hologres-knowledge-base

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Build enterprise search and RAG knowledge bases on Hologres using vector indices, full-text inverted indexes, and embed-to-index workflows.

Core Features & Use Cases

  • Create knowledge bases as a single column-store table combining text content, embeddings, and scalar metadata to support vector and BM25-style search in one query.
  • Ingest documents with client-side or server-side embeddings, then build HGraph vector indexes and full-text indexes for hybrid search and Q&A workflows.
  • Use cases include building internal knowledge bases, document repositories, and knowledge pipelines that leverage both semantic search and keyword search for accurate retrieval.

Quick Start

Create a knowledge base that stores document chunks, embeds them, and enables hybrid search across text and vectors.

Frequently Asked Questions about hologres-knowledge-base

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a RAG knowledge base with vector and full-text search on Hologres?▼

To build a RAG knowledge base on Hologres, create a single column-store table for text, embeddings, and metadata, then build HGraph vector indices and full-text inverted indexes to enable hybrid search in one query.

What is hybrid search and how does it work for enterprise document retrieval?▼

Hybrid search combines semantic vector search with BM25-style keyword search within a single Hologres query, retrieving accurate document chunks by leveraging both HGraph and full-text inverted indexes.

Do I need Hologres V4+ to create vector indices and full-text indexes?▼

Yes, building RAG knowledge bases with HGraph vector indices and full-text inverted indexes requires Hologres V4+ features to support the embed-to-index workflow and hybrid search.

Can I use client-side embeddings with holo-search-sdk instead of server-side ai_gen()?▼

Yes, you can ingest documents using either server-side embedding via the ai_gen() function or client-side embedding with holo-search-sdk to populate your Hologres knowledge base.

What is the best way to store document chunks and embeddings for a Q&A pipeline?▼

The best way is storing document chunks, embeddings, and scalar metadata in a single Hologres column-store table, enabling both vector and text search for Q&A retrieval pipelines.

Why is my Hologres knowledge base not returning accurate semantic search results?▼

Inaccurate semantic search may result from missing HGraph vector indices or improper document chunk ingestion, so ensure embeddings are built correctly and hybrid search queries both text and vectors.

Related Skills

Explore More Data & Analytics Skills→
ironsheepironsheep
4

ingest-source

Orchestrate source document ingestion into the P2 knowledge base.

Community
Advanced
AI-Foundry-CoreAI-Foundry-Core
4

rag-implementation

Retrieve relevant documents to ground LLM outputs with citations.

Official
Advanced
manifoldlogicmanifoldlogic

maproom-search

Convert code chunks into numerical vectors for semantic codebase search.

Official
Advanced
MSamiulHasnatMSamiulHasnat
1

AgentDB Vector Search

Index embeddings and perform context-aware similarity queries for document retrieval.

Community
Advanced