Vector based similarity search index for QLever database
Improved search for scalable open-source graph database
This project extends QLever, an extremely efficient and scalable open-source graph database, by implementing a generic vector-based similarity search index. By integrating this feature alongside existing support for full-text and geo-spatial search, the project creates a unified engine that efficiently combines structured graph queries with semantic vector search. This makes massive Linked Open Data datasets readily available for AI-driven Retrieval Augmented Generation (RAG), including datasets such as Wikidata, UniProt, and OpenStreetMap.
- The project's own website: https://github.com/ad-freiburg/qlever
Run by RS WORKS EE
This project was funded through the NGI0 Commons Fund, a fund established by NLnet with financial support from the European Commission's Next Generation Internet programme, under the aegis of DG Communications Networks, Content and Technology under grant agreement No 101135429. Additional funding is made available by the Swiss State Secretariat for Education, Research and Innovation (SERI).
