Package com.fxpal.search.reverted.filters

Specifies interfaces and classes that can be used to construct filter chains for generating basis queries to build a reverted index.

See:
          Description

Interface Summary
BasisQuerySource This interface specifies methods that classes that generate or filter basis queries must implement.
 

Class Summary
BasisQueryAdapter This class wraps an Iterator or Iterable of strings and creates BasisQuery instances of the specified type.
BasisQueryRecorder This class writes basis queries to a file as they are generated by the source.
BasisQueryWrapper Base class for composing filters for basis queries.
Filter Specifies a means to filter a source, emitting fewer terms than the source generates.
RevertedQueryBasisQuerySource Iterator that takes a RankedList of results from a reverted index search and transforms each one into a BasisQuery.
StreamBasisQuerySource Reads queries from a file or some other stream.
 

Package com.fxpal.search.reverted.filters Description

Specifies interfaces and classes that can be used to construct filter chains for generating basis queries to build a reverted index.

To build a reverted index, a source of basis queries is required. This can be a query log file, the terms from an inverted index, or some other source. The classes in this package define a composition/filter pattern that can be used to filter or transform the terms from a source to generate queries that will be run by a RevertedDocumentIterator to create reverted documents. For example, LuceneIndexTermIterator.html#createStandardIterator() implements the following filter chain:

BasisQuerySource terms = new LuceneIndexTermIterator(index, basisQueryFields); terms = new LuceneDocumentFrequencyFilter(terms, index, minDF); terms = new BasisQueryRecorder(terms, basisQueryFile);

More information on reverted indexing is in the overview. For complete details on the rationale behind reverted indexing, see our paper. If you publish research results in the preparation of which you used our code, we would appreciate a citation to the above paper. You can also include a reference to where others can download it.

Author:
Gene Golovchinsky

(c) 2011 FX Palo Alto Laboratory, Inc.