llama-index-cmem

llama-index tools eccenca Corporate Memory Integration

Development

• Run task to see all major development tasks.
• Use pre-commit to avoid errors before commit.
• This repository was created with this copier template.

Usage

The llama-index-cmem package allows using eccenca Corporate Memory together with LlamaIndex.

Components

• SPARQLReader: Use to load documents from SPARQL query to use in ingestion pipeline.
• SPARQLRetriever: Execute SPARQL query to retrieve context from SPARQL endpoint.
• NLSPARQLRetriever: Text-to-SPARQL retriever to retrieve context from SPARQL endpoint using natural language.
• CMEMQueryCatalogRetriever: Execute SPARQL query from CMEM query catalog to retrieve context from SPARQL endpoint.

There are several example notebooks available:

• NLSPARQLRetriever and CMEMQueryCatalogRetriever.
• ChatEngine: Chatting with an RDF knowledge graph data.
• QueryEngine: Single Q&A engine with RDF knowledge graph data.
• PGVector: Generate embeddings from RDF knowledge graph and query with natural language (via vector similarity).