Commits

Author Commit Message Labels Comments Date
Rusty Klophaus
Merge branch 'pt5636184-query-parser' of github.com:rklophaus/riak_search into pt5636184-query-parser
Rusty Klophaus
Simplify preplanning process to make it easier to rewrite query. Add checking around skipped terms, and throw exception if the user searches on a stopword. Add tests around proximity searching.
Rusty Klophaus
Update range searches to account for negative integers and to switch out-of-order searches.
Rusty Klophaus
Update AnalysisHandler to account for AnalyzerFactories that don't use the PositionIncrementAttribute. Update date_test and verify that date searching works.
Rusty Klophaus
Fix bug where searching for the same word twice caused error.
Rusty Klophaus
Fix scoring and node balancing for proximity and phrase searches.
Rusty Klophaus
Merge branch 'master' of git@github.com:rklophaus/riak_search into pt5636184-query-parser
Rusty Klophaus
Continue proximity and phrase search work.
Rusty Klophaus
Update #intersection and #union operators to restore load balancing functionality across nodes in the cluster depending on how many results we expect to read from the node.
Rusty Klophaus
Begin updating Riak Search to use the new Lucene Parser project, including terms, booleans, wildcards, ranges, and scoring. (Still working on node balancing, phrase searching, and proximity searching.)
Rusty Klophaus
Add Lucene Parser project, parses Lucene-syntax queries and emits a query graph.
Rusty Klophaus
Simplify preplanning process to make it easier to rewrite query. Add checking around skipped terms, and throw exception if the user searches on a stopword. Add tests around proximity searching.
Rusty Klophaus
Update range searches to account for negative integers and to switch out-of-order searches.
Rusty Klophaus
Update AnalysisHandler to account for AnalyzerFactories that don't use the PositionIncrementAttribute. Update date_test and verify that date searching works.
Rusty Klophaus
Fix bug where searching for the same word twice caused error.
Rusty Klophaus
Fix scoring and node balancing for proximity and phrase searches.
Rusty Klophaus
Continue proximity and phrase search work.
Rusty Klophaus
Update #intersection and #union operators to restore load balancing functionality across nodes in the cluster depending on how many results we expect to read from the node.
Rusty Klophaus
Begin updating Riak Search to use the new Lucene Parser project, including terms, booleans, wildcards, ranges, and scoring. (Still working on node balancing, phrase searching, and proximity searching.)
Rusty Klophaus
Add Lucene Parser project, parses Lucene-syntax queries and emits a query graph.
Bryan Fink
expose existence of 'riak-admin remove' and protect its args
Daniel Reverri
Add remove command
Bryan Fink
update search ets backend for new interface (bz://795) Mostly: - all of the multi_* became primary ops, with the single-arg ops going away - range is no longer a pre-plan operation, but instead an execution-time process (so it operates the same as stream, but over a range instead of a specific IFT) Passes tests/run_all.sh just as well as merge_index at this point
Bryan Fink
include stop terms in position calculation (bz://732) Erlang analyzers may now return the atom 'skip' in their term lists. This atom marks terms that should not be indexed, but should be counted in the term position calculation. Java analyzers should set the PositionIncrementAttribute of their TokenStream correctly. Skipped terms are encoded as additional zeros in the qilr_analyzer stream, and converted to 'skip' atoms by qilr_analyzer:parse_result…
Rusty Klophaus
Remove spaces (test reviews).
Ryan Tilder
We were missing a magic variable For some reason riak_search's installation path is riaksearch. Go figure.
Ryan Tilder
Missing %files entry for the initscript
Ryan Tilder
One last change
Ryan Tilder
Revert an accident Included a change to the top level Makefile that didn't help
Ryan Tilder
Pull in the various updates for - vs _ There can be only one!
  1. Prev
  2. Next