Description of problem: The statistics API is currently taking into account copied translations and awarding credit to the original translator. In order to keep a more accurate track of a translator's contributions, copied translations should not be counted as 'Translated' strings. They should either be not counted, or counted as a different type of contribution.
This is caused because Copy Trans copies the original translator into the translator field. Estimated implementation: - Need to record more metadata: - copied from copy version (100%) - copied with version merge (100%) - copied by copy trans ("project auto-fill") (100%) - copied with TM merge (100% can be copied as Translated) - Does metadata apply to just the revision, or stay on the text flow for its lifetime? - Just the history entry has it. Don't want later edits to be uncredited, since the point is to be able to see the amount of actual work done by a translator. - Wipe the metadata on the HTextFlowTarget when an edit is made. Changing from fuzzy to translated (without new text) will make a new revision. This is fine because you have contributed by reviewing it. Technical note: wiping metadata when creating a revision is done for revision comments at the moment, so use that for reference. Stats that should be influenced by not recording copies as translations: - User profile page (All contributions by a single user across all projects.) - Contribution API stats (Contributions to a particular project per-user.)
Pull request: https://github.com/zanata/zanata-server/pull/888
Verified at 70e87501e4b0bdbc8c03be446e384184dfa61285
Pull request: release branch pull request: https://github.com/zanata/zanata-server/pull/900