data

Basic sample classes for Lightning IR.

This module defines the basic samples classes for Lightning IR. A sample is single entry in a dataset and can be grouped into batches for processing.

Classes

DocSample(doc_id, doc)	A sample of document data containing a document and its id.
IndexBatch(doc_ids, docs)	A batch of index that combines multiple DocSample instances
QuerySample(query_id, query[, qrels])	A sample of query data containing a query and its id.
RankBatch(queries, docs[, query_ids, ...])	A batch of ranking data combining multiple RankSample instances
RankSample(query_id, query, doc_ids, docs[, ...])	A sample of ranking data containing a query, a ranked list of documents, and optionally targets and qrels.
SearchBatch(query_ids, queries[, doc_ids, qrels])	A batch of search data that combines multiple QuerySample instances.
TrainBatch(queries, docs[, query_ids, ...])	A batch of ranking data that combines multiple RankSample instances