Commits

Show all
Author Commit Message Labels Comments Date
Matt Chaput
Bumped version number to 2.2.1.
Tags
2.2.1
Matt Chaput
Fixed up documentation on scoring a bit. Added scoring.FunctionWeighting.
Matt Chaput
Added tag 2.2 for changeset 9b9108fd23fd
Matt Chaput
Bumped version number to 2.2. Updated release notes.
Tags
2.2
Matt Chaput
Replaced NullMatcher with a singleton. Although instantiation of NullMatchers isn't exactly on a hot path (it can only happen during replace()), I always meant to get around to this because it makes sense, and will add some tiny speed improvement, so why not? Fixes issue #181. Fixed braindead test in test_results.
Matt Chaput
Removed accidentally committed debug print.
Matt Chaput
Fixed method name typo in TF_IDFScorer. Fixes issue #182.
Matt Chaput
Fixed scorer handling of sub-searchers. Fixes issue #183.
Matt Chaput
Added information about allow_overlap and StoredFieldFacet.
Matt Chaput
Added docs fixes from Wouter Bolsterlee. Fixes issue #184.
Matt Chaput
Speed improvements for overlapping facets. OverlappingFieldCategorizer now uses term vectors if available. StoredFieldFacet/Categorizer now support allow_overlap by splitting the stored value. Added vector keyword argument to KEYWORD field type. Raise KeyError instead of generic Exception when a given document does not have a term vector for a given field. Added stress testing code for a faceting a large tag collection. See issue #185.
Matt Chaput
Minor redundant logic fix. See issue #181.
Matt Chaput
More PEP 8 line length fixes.
Matt Chaput
Was calling a.max_quality() even though a might have just been replaced with NullMatcher(). Fixes issue #181. Best solution might have been to just add implement max_quality() on NullMatcher, which I've added here, but I also tried to make the logic clearer/more robust in IntersectionMatcher.replace().
Matt Chaput
Matt Chaput
Reformatted source code for better PEP 8 compliance.
Matt Chaput
Fixed undefined variable error. Fixes issue #178.
Matt Chaput
Added experimental in-memory pool. Switched b(chr(n)) to pack_byte(n).
Matt Chaput
Relaxed type checking of string values passed to DATETIME.
Matt Chaput
Lucene includes the Reuters21578 corpus with their source distribution, so I guess I can too.
Thomas Waldmann
transformed all *.py files to lf lineends, remove trailing blanks, normalize EOF
Thomas Waldmann
DATETIME.to_text - extend type support to text_type (unicode/str) If whoosh parses a query for a DATETIME field, it makes quite some effort to make sense out of numeric strings, so something like 2011 or 20111231235959999999 is parsed as one would expect it. In this changeset, I modified method .to_text (which is used at indexing time) to also try to make sense out of numeric strings (in about the same way). It calls floor() so that 2011 means the same as 2011…
Matt Chaput
SegmentWriter's implementation of Writer.add_field() wasn't passing keyword args up. Fixes issue #170.
Matt Chaput
Fixed up TeeFilter docstring, added warning about (lack of) performance.
Matt Chaput
Simplified DateRangeFacet. Besides decomplicating it, this allows it to support relativedelta objects as gaps.
Matt Chaput
Facets.add_field() wasn't passing the allow_overlap keyword argument on to FieldFacet(). Added test for query facet with overlap. See issue #169.
Matt Chaput
Added TeeFilter to API docs.
Matt Chaput
Added recipes for word length filtering and case-sensitive searches.
Matt Chaput
Added TeeFilter.
Matt Chaput
Expanded * import.
  1. Prev
  2. Next