Details
-
Improvement
-
Status: Closed
-
Minor
-
Resolution: Fixed
-
2.2
-
None
-
New, Patch Available
Description
Few fixes and improvements for the search quality benchmark package:
- flush report and logger at the end (otherwise long submission reports might miss last lines).
- add run-tag-name to submission report (API change).
- add control over max-#queries to run (useful at debugging a quality evaluation setup).
- move control over max-docs-to-retrieve from benchmark constructor to a setter method (API change).
- add computation of Mean Reciprocal Rank (MRR) in QualityStats.
- QualityStats fixed to not fail if there are no results to average.
- Add a TREC queries reader adequate for the 1MQ track (track started 2007).
All tests pass, will commit this in 1-2 days if there is no objection.