Commits

Author Commit Message Labels Comments Date
Matt Chaput
Field boost was not applied to inlined weights. Fixed bug in UnionMatcher.weight().
Matt Chaput
Fixed MultiSegmentWriter to properly close sub-writers' files. Fixes issue #113.
Matt Chaput
Converted tests to nose.
Matt Chaput
Fixed syntax and logic errors in Sorter.
Matt Chaput
Added file manifest template to get setuptools to include docs, tests, and other files in the source distribution. Fixes issue #111.
Matt Chaput
Bumped version number to 1.7.5 for MultiPool bugfix.
Matt Chaput
Fixed data loss bug where the buffer was not flushed by MultiPool.finish().
Matt Chaput
Fixed bugs with Google App Engine compatibility.
Matt Chaput
Added check for old version of StemFilter class in __setstate__. Issue #112.
Matt Chaput
Bumped version number to 1.7.3.
Matt Chaput
Added an updated version of the code from Whoosh AppEngine for storing a FileIndex's files in the app engine datastore as blobs.
Matt Chaput
The "content" field's analyzer was stripping the punctuation, so I wasn't test what I thought I was. Fixes issue #109.
Matt Chaput
Fixed import of methodcaller, which was only added in Python 2.6. Fixes issue #110.
Matt Chaput
Added Searcher.more_like() and Hit.more_like_this() methods to make finding similar documents based on key words easier.
Branches
morelikethis
Matt Chaput
Added a minor parsing test.
Matt Chaput
Fixed field name parsing. Bumped version to 1.7.2. The Field token was converting non-schema fields to text before the field alias plugin could see them.
Matt Chaput
Bumped version to 1.7.1 for bugfix release. See issue #108.
Matt Chaput
Defined Query.__ne__ as the opposite of Query.__eq__. Fixed RangeMixin.__eq__. Bizarre Python thing I learned today: __ne__() does not automatically return the opposite of __eq__, you actually have to define it. What I don't understand is why the fact that __ne__ wasn't defined masked this problem on Windows and Mac OS X but not on Ubuntu. There might be a Python platform bug here. Anyway, I fixed the problem that was masking the error on Mac and Windows (lack of __ne__), a…
Matt Chaput
API docs were missing the Hits object.
Matt Chaput
Fixed problems with sorting API docs.
Matt Chaput
Added an API doc for whoosh.sorting.
Matt Chaput
Changed version to 1.7.0, updated release notes. Added back NullFragmenter as an alias for WholeFragmenter. Added another test for result lengths. Fixed docstring problems.
Matt Chaput
Moved sorting code into whoosh.sorting, restored ability to do complex sorts.
Matt Chaput
Fixed method name. Sigh.
Matt Chaput
Ugh, forgot to remove import.
Matt Chaput
OK, the last commit was too complex, reverted and just added the boosting matcher in the Phrase.matcher() method. See #107.
Matt Chaput
Fixed Phrase query's ability to pass its boost on to generated span query. Fixes #107. Added BoostQuery class to wrap aggregate queries in a single boost.
Matt Chaput
Limit the size of the job queue to prevent MultiPool from filling up the temp directory with job files.
Matt Chaput
Changed meaning of StemFilter/StemmingAnalyzer cachesize parameter so None means no caching. Added info about unbounded stem caching to batch indexing docs.
Matt Chaput
Redefined Query.accept(), .copy(), and .replace() as applications of .apply().
  1. Prev
  2. Next