Details
-
Bug
-
Resolution: Fixed
-
None
-
High
-
Translation status assignment rewritten for terms tagged by TermTagger
-
Empty show more show less
Description
problem
The editor_Plugins_TermTagger_RecalcTransFound class is heavily using the term DB table, especially on import.
Caching could reduce the DB usage here.
solution
Regarding terminology there is already a similar mechanism - where the terminology.tbx file is created per each task, validity is ensured over the tbxHash stored in LEK_task_meta per each task.
If a term collection is updated, the tbxHash of all affected tasks are invalidated.
For import it makes sense to cache all the term data collected for that task (by collectionIds and languages).
Use Zend Cache Mechanisms directly in the recalc class.
Consider to move the termModel::loadByMid, termModel::getAllTermsOfGroup and the termModel::findHomonym methods out of termModel and place them into a separate class.
Then the caching for the import purpose could go into that class.
Advantages: making termmodel smaller. Controlmechanism if using caching or not, could be controlled by global const ZFEXTENDED_IS_WORKER_THREAD
since on single segment save this is not defined.
Attachments
Issue Links
- relates to
-
TRANSLATE-2935 Avoid term-check false positive in case of homonyms and display homonyms in source and target
- Done