Permalink
Commits on Aug 3, 2011
  1. bumping version to 0.4.3

    aaw committed Aug 3, 2011
  2. 1.8.7 compatibility

    tyok committed Aug 3, 2011
Commits on Jul 28, 2011
  1. bumping version to v0.4.2

    aaw committed Jul 28, 2011
  2. Merge pull request #16 from dblock/mongoid_indexes

    Delaying creation of mongodb indexes in-sync with MongoId
    aaw committed Jul 28, 2011
  3. Updated readme.

    dblock committed Jul 28, 2011
Commits on Jul 27, 2011
  1. bumping to v0.4.1

    aaw committed Jul 27, 2011
  2. Merge pull request #15 from dblock/master

    Bug: sort order matters for indexes created via multiple models
    aaw committed Jul 27, 2011
  3. Using Mongoid.logger for logging.

    dblock committed Jul 27, 2011
  4. Fixed order of filters.

    dblock committed Jul 27, 2011
  5. Try again with => .

    dblock committed Jul 27, 2011
Commits on Jul 19, 2011
  1. bumping version to 0.4.0

    aaw committed Jul 19, 2011
  2. fixing a bug in ensure_indexes

    aaw committed Jul 19, 2011
  3. clean-up

    aaw committed Jul 19, 2011
  4. fixing specs

    aaw committed Jul 19, 2011
Commits on Jul 18, 2011
  1. Removing all use of map/reduce, reworking the logic for full-text sea…

    …rch using
    
    straight queries. I think this is the best way to go for now. We need to
    somehow bound the number of candidates we consider after breaking the query
    into n-grams, but that would mean a query like
    
      {ngram: {$in: ['mon','ong','ngo','god','odb']}}
    
    (or a similar query using $or) along with a sort-and-limit by the score field
    (e.g.,get the 1000 records with the highest scores that satisfy that query.)
    But MongoDB currently won't use an index on {ngram: 1, score: -1} to run that
    query, instead preferring to do a full table scan for the ngram matches.
    
    So we're going to treat the ngram query in a similar fashion to how we might
    treat an AND query: get frequency counts of all of the ngrams and use those
    to direct some direct queries into the index collection, using the score
    field to return only the top-scoring results.
    
    Most tests pass, still some work to do.
    aaw committed Jul 18, 2011
Commits on Jul 15, 2011
  1. updating README

    aaw committed Jul 15, 2011
  2. little bit of clean-up

    aaw committed Jul 15, 2011
Commits on Jul 14, 2011
  1. putting a hard limit on the number of candidates returned from map/re…

    …duce, sorting those by score
    aaw committed Jul 14, 2011
Commits on Jul 12, 2011
  1. Merge pull request #12 from dblock/master

    I like my rspec to be colorful
    aaw committed Jul 12, 2011
  2. Added some color.

    dblock committed Jul 12, 2011
  3. Merge pull request #11 from steverandy/master

    Use constantize to support class name with module for example (Module::ClassConstantName).
    aaw committed Jul 12, 2011
  4. Use constantize to support class name with module for example (Module…

    …::ClassConstantName).
    steverandy committed Jul 12, 2011
Commits on Jul 7, 2011
  1. updating version info for 0.3.7

    aaw committed Jul 7, 2011
  2. Merge pull request #10 from dblock/update-ngram-index-bulk

    Rebuild indexes in bulk, support for incremental and non-incremental rebuild.
    aaw committed Jul 7, 2011
  3. Removed incremental parts.

    dblock committed Jul 7, 2011
Commits on May 27, 2011
  1. bumping version to 0.3.6

    aaw committed May 27, 2011