Commits

Author Commit Message Labels Comments Date
Daniel Lindsley
Added failing test for deleting by term.
Daniel Lindsley
Added failing tests for query parsing.
Matt Chaput
Added code for shift_step=0 case. Renamed NumericRange._or_query(). Added tests. Minor cleanups.
Matt Chaput
Finished work on tiered numeric ranges.
Matt Chaput
Checking in work on tiered numeric indexing. Made NUMERIC and BOOLEAN fields deal with *. Moved QueryParserError into qparser.default. Don't propagate errors in self-parsing.
Matt Chaput
Fix bug where I didn't update Facets.from_field() to new API. Fixes issue #54.
Matt Chaput
Fixed brain fart in NUMERIC Decimal support. Fixes issue #56.
Matt Chaput
Check whether the posting reader for a term is active before calling id() on it. Issue #55.
Matt Chaput
Added type check to NUMERIC initializer. Fixed test failures.
Matt Chaput
Added Decimal support to NUMERIC field type via the decimal_places keyword arg.
Matt Chaput
Created whoosh.support.numeric to prepare for tiered numeric indexing. Minor fixes.
Matt Chaput
Bumped version number.
Matt Chaput
Fixed a bug in collect() when matcher.replace() returned a NullMatcher
Matt Chaput
Merging branches.
Matt Chaput
Adding benchmark products to .hgignore, missed a doc in last commit
Matt Chaput
Checking in half-finished docs, minor changes.
Matt Chaput
Fixed a bug in FilePostingReader. The all_ids() method doesn't set self.blockinfo as it goes, but _read_ids() and _read_weights() were looking in self.blockinfo for the compressed lengths, so I had to add the compressed lengths as arguments to the read functions.
Matt Chaput
Tried to clarify BatchWriter docstring.
Matt Chaput
Switched posting header to use a 4-byte delta to next block instead of an 8-byte absolute pointer, freeing up bytes for metadata. Fixed ResultsPage tests. Added test for SpanCondition.
Matt Chaput
Added signed varints to StructFile. Minor docstring fixups.
Matt Chaput
Added spans.SpanCondition query type.
Matt Chaput
Check whether the matcher is active in SpanWrappingMatcher._find_next().
Matt Chaput
Fixed query mode usage of NgramTokenizer. Added simple accent-folding charset. Unit tests.
Matt Chaput
Fixed wrong number of arguments in AsyncWriter.delete_document(). Fixes issue #49 .
Matt Chaput
Raise ValueError if user calls Searcher.search_page() with too high pagenum. Added ResultsPage.is_last_page() method.
Matt Chaput
Added logic to protect SpanWrappingMatcher against trying to get the spans when the underlying matcher is inactive.
Matt Chaput
Added more reasonable error messages for methods on non-indexed fields.
Matt Chaput
Fixed error in variations() function affecting words eding in "e". Made simplify() return normalized results.
Matt Chaput
Bumped version and changed index format to account for changes due to index compression work.
Matt Chaput
Term index and vector files now use 16-bit codes instead of full field names. Changed architecture of term index and term vector reader/writers.
  1. Prev
  2. Next