Commits

Author Commit Message Labels Comments Date
Matt Chaput
Changed hash file and codec to only write serially (no seek backs while writing). * Hash file now writes data, then hash tables, then hash directory, then extras. * Hash readers must not be passed file length so they can seek relative to the end. * Postings now use negative block length to indicate last block instead of seeking back and writing number of blocks. * Moved old hash file code into legacy whoosh2 codec. Simplified and commented filetables.py and whoosh3.py along th…
Matt Chaput
Stop writing overlapping tokens one after another. See issue #280.
Matt Chaput
Removed import of unicode_literals and wrapped non-ASCII strings with u(). See issue #275.
Matt Chaput
Add a guard against skipping past the end of the array. See issue #278.
Matt Chaput
We have to use key_to_name to get a comparable string, since key_for might return a memoryview in Python 3.x.
Matt Chaput
Merging Thomas's markup fixes.
Matt Chaput
Added create() and destroy() methods to Storage. Added docstrings.
Matt Chaput
Removing workspace.xml from version control.
Matt Chaput
Added simple implementation of version number parsing for possible future use.
Matt Chaput
Updated configuration files.
Matt Chaput
Fixed bug where TranslatingColumn didn't implement sort_key. Implemented collector methods for counting results and getting docnums. Updated Results object to use new collector methods instead of re-running search.
Matt Chaput
Fixed whoosh3 codec filenames. Added sortable option to KEYWORD field type. Minor formatting changes.
Matt Chaput
Added a unit test for using a numeric field as the unique key for update_document. I'm pretty sure this already exists somewhere else, but I couldn't find it.
Thomas Waldmann
fix markup errors in searching docs
Matt Chaput
Fixed pickling of LanguageAnalyzer.
Matt Chaput
Changed matchers to store scores as doubles, so stored scores match "live" ones. Changed Preload matcher to not store 0s for documents outside the matching range.
Matt Chaput
Changed W3 codec to save columns in separate files. This avoids the common case of a compound file inside a compound file.
Matt Chaput
Added PreloadedUnionMatcher to speed up Or queries.
Matt Chaput
Fixed copy/paste error.
Matt Chaput
Renamed ComplexPhrasePlugin to SequencePlugin. Added optional slop factor. Added optional prefix to fuzzy syntax. Added optional slop factor to phrase syntax. Added docs for sequence and fuzzy queries. Added unit tests.
Matt Chaput
Added Sequence query type. Added ComplexPhrasePlugin to create sequence queries.
Matt Chaput
Changed copyright message to year range. Apparently that's the done thing.
Matt Chaput
Added scored=False argument to search() to allow fast, unordered results.
Matt Chaput
Added explicit __ne__ methods.
Matt Chaput
Added explicit __ne__ methods.
Branches
2.4x
Matt Chaput
Fixed syntax error from bad merge.
Branches
2.4x
Matt Chaput
Compatibility fixes for Python 2.5.
Matt Chaput
Changes to support FunctionPlugin.
Matt Chaput
Made ArrayUnionMatcher more efficient when skipping.
Matt Chaput
Fixed off-by-one error in FuzzyTermPlugin.
  1. Prev
  2. Next