Manticore Search 29.9.0: Chunked Auto-Embeddings and mmap Access for Columnar Attributes

Manticore Search 29.9.0 has been released, introducing significant improvements for search indexing and vector search. Key features include support for chunked and multi-vector auto-embeddings, along with input constraints for embedding models. A major technical update is the default implementation of mmap access for columnar attributes, which enhances data processing performance. Additionally, the developers have added support for UTF-8 identifiers and fixed several issues related to hybrid search, KNN, bulk ingestion, and grouping mechanisms. This release focuses on improving the stability and efficiency of the search engine in high-load scenarios.
This is a summary. Read the full article at the original source:
HabrRelated stories
In a provocative post, Eddie Jaoude argues that the Python library Pandas, long considered the industry standard for data manipulation, has become out…
PlanetScale has officially announced the launch of Neki, a new tool designed to simplify and enhance database management. As the company continues to…
The second part of the series on PostgreSQL temporary tables focuses on performance issues arising from their intensive use. Despite moving temporary…



