Hyperspace Search

Hyperspace search allows to filter and assign score for candidates using classic and vector search. Only candidates that passed the filter stage will be assigned a score.

In addition, Hyperspace allows aggregations of candidate fields. Aggregations can be performed on all candidates, and not just those that passed filtering.

Native Python Syntax and DSL Syntax

Hyperspace allows to build classic, vector and hybrid search queries in either domain specific syntax (.json structure) or native python syntax.

Multiple Score Methods

Hyperspace supports a variety of score mechanisms. These include TF-IDF and BM25 based score, weights and boosts for classic search, and similarity metrics such as Euclidean and Hamming distance for vector search.

No Limitation on Number of Metadata Fields

Hyperspace efficient memory management allows to include an extremely large number of keyword and value fields in each query, allowing practically unlimited number of fields.

Hyperspace fully supports multi model vector search, allowing to use multiple vectors in each search query, and to use the results of each vector search in order to filter other vector searches.

Extremely Large Vectors

Hyperspace allows to perform vector search with extremely large vectors, with thousands of elements per vector.

Hyperspace allows to create sophisticated filtering and scoring logic based on both vector and classic search results.

Hyperspace hybrid search that combines classic and vector search in a simple manner, in a native Python syntax.

Last updated