Commits

Show all
Author Commit Message Labels Comments Date
mchaput
Changed API to use field names everywhere except the lowest filedb levels.
Branches
fieldnames
mchaput
Fixed errors in unit tests.
Branches
pre1
mchaput
Changed to compress email bodies.
Branches
pre1
mchaput
Initial work on flexible schema. Added explicit weight argument to posting-handling signatures. Term info now stores total weight instead of total frequency. Added numerickeys argument to stored_fields. Fixed open file leaks. Fixed file descriptor leak in file lock. Changed AsyncWriter API, added BatchWriter. Fixed tests. Added misc. similarity functions to whoosh.classify.
Branches
pre1
mchaput
Fixed bugs in more obscure matchers. Added superficial tests for matcher types.
Branches
pre1
Matt Chaput
Changed uses of "self.index.schema" to "self.schema".
Branches
pre1
Matt Chaput
Fixed bugs to get tests passing. Commented out ramdb tests for now.
Branches
pre1
Matt Chaput
Finished up block quality optimizations.
Branches
pre1
Matt Chaput
Removed leftover "main" testing code.
Branches
pre1
Matt Chaput
Fixed bugs in BitSet.
Branches
pre1
Matt Chaput
Replaced filepostings with filepostings2. Started updating old code for new APIs. Added matching.MultiMatcher. Added Searcher.postings(). Added union() and intersection() methods to BitVector and BitSet.
Branches
pre1
Matt Chaput
Put back boost arguments, added copy() method to matchers, removed reset() method. Moved IDF implementation into Weighting object. Added BitSet object.
Branches
pre1
Matt Chaput
Checking in progress on quality-based search optimizations.
Branches
pre1
Matt Chaput
Checking in experimental quality-based optimizations.
Branches
pre1
Matt Chaput
Checking in beginnings of new matcher architecture.
Branches
pre1
Matt Chaput
Removed rogue assert.
Branches
pre1
Matt Chaput
Changed byte-order < back to !.
Branches
pre1
Matt Chaput
Long list of changes to indexing and searching. Changed field length file to a list of approximated byte values that are read into memory. Pools must now accept any keyword argument in their __init__ method. Pools spool lengths to disk files and then collect them. Pools now use marshal instead of custom encode/decode functions. StructFile.read/write_array() now use array methods and byteswap if necessary. Changed scoring and Searcher interfaces…
Branches
pre1
Matt Chaput
Updates tests to current API.
Branches
pre1
Matt Chaput
Added missing add_field_length() method to pools. Added missing functionality to ramdb.
Branches
pre1
Matt Chaput
Changed types of on-disk structs to little ("<") from network order ("!").
Branches
pre1
Matt Chaput
Merged with 48f7e10c0077
Branches
pre1
Matt Chaput
Added LengthWriter/LengthReader for storing field lengths by field.
Branches
pre1
Matt Chaput
Now store field lengths in a custom file instead of StructHash. OrderedHashReader now does binary searches on disk instead of in memory. Bug fixes in query module. Added Reader.max_field_length() method. Removed field_length method from Index interface. Renamed Fake* classes to List* in postings. Added util.now() function.
Branches
pre1
Matt Chaput
Added options to initializers in analysis module for maxsize parameter of StopFilter.
Branches
pre1
Matt Chaput
Added maxsize filter to StopFilter.
Branches
pre1
Matt Chaput
Added default argument to doc_field_length().
Branches
pre1
Matt Chaput
Fixed pool field lengths. Changed pools to store temp files in a directory.
Branches
pre1
Matt Chaput
Added code to MultiPool to delete temp files. Added Schema.clean() method. Sorted and expanded default stop list. Renamed HtmlFormatter.clear() to clean().
Branches
pre1
Matt Chaput
Fixed pools.MultiPool implementation to not deadlock, hopefully.
Branches
pre1
  1. Prev
  2. Next