Vespa Product Updates, August 2019: BM25 Rank Feature, Searchable Parent References, Tensor Summary Features, and Metrics Export

Vespa, an open-source big data processing and serving engine mainly developed by Yahoo engineers, announced several updates for August 2019, focusing on enhancing its functionality for various applications like Yahoo News and Verizon Media Ad Platform. The updates introduce features such as the BM25 Rank Feature, which implements the Okapi BM25 ranking function for text document ranking, and a Searchable Reference Attribute that allows searching using the document id of a parent document-type instance. Additionally, the Tensor Summary Features now support returning tensors in summaries for easier rank tuning, and a new node metric interface facilitates exporting metrics with the capability of aliasing metric names for integration with monitoring tools like CloudWatch and Prometheus. These enhancements reflect Vespa's commitment to incorporating community feedback and contributions to improve its big data solutions.