Random notes from an initial investigation

• EntityParserOutputDataUpdater should probably not know about StatementDataUpdater and SiteLinkDataUpdater.
• ReferencedEntitiesDataUpdater
  • split into 3 classes: one actual ReferencedEntitiesDataUpdater, one service that collects entity ids from statements, one service that collects entity ids from sitelink badges
  • could either take a list of entity ids from per-entity type collector classes, or take a per-entity service that takes the entity and scrapes its entity ids

More concrete plan:

• rename EntityParserOutputDataUpdater to EntityParserOutputDataUpdaterCollection (naming suggestions welcome)
  • it should only be used to register EntityParserOutputDataUpdaters, pass them the entity they should operate on, and call updateParserOutput on them
  • processStatements and processSitelinks should go into concrete EntityParserOutputDataUpdater implementations

• make EntityParserOutputDataUpdater an interface that operates on entities and updates the ParserOutput
  • has methods processEntity and updateParserOutput

• existing StatementDataUpdaters should either be changed to implement EntityParserOutputDataUpdater instead, or they need an adapter (is that better?)
  • note: this is worse performance-wise. Instead of iterating over all statements once and passing individual statements to StatementDataUpdaters we would need to iterate over all statements once per StatementDataUpdater
  • everything concerning SiteLinks and Statements is removed from ReferencedEntitiesDataUpdater - it should only deal with EntityIds and add their respective links to ParserOutput
    • ReferencedEntitiesDataUpdater gets a per entity type collection of EntityIdCollectors (naming suggestions welcome) injected
    • the SiteLink bit of it becomes a SiteLinkBadgeItemIdCollector
    • the Statement bit becomes StatementReferencedEntityIdCollector

Change 442864 had a related patch set uploaded (by Jakob; owner: Jakob):
[mediawiki/extensions/Wikibase@master] Refactor ParserOutputDataUpdater

https://gerrit.wikimedia.org/r/442864

Change 443104 had a related patch set uploaded (by Jakob; owner: Jakob):
[mediawiki/extensions/Wikibase@master] Make EntityReferenceExtractors configurable per entity type

https://gerrit.wikimedia.org/r/443104

Change 442864 merged by jenkins-bot:
[mediawiki/extensions/Wikibase@master] Refactor ParserOutputDataUpdater

https://gerrit.wikimedia.org/r/442864

Change 443451 had a related patch set uploaded (by Jakob; owner: Jakob):
[mediawiki/extensions/Wikibase@master] Clean up around EntityParserOutputDataUpdater related code

https://gerrit.wikimedia.org/r/443451

Change 443451 merged by jenkins-bot:
[mediawiki/extensions/Wikibase@master] Clean up around EntityParserOutputDataUpdater related code

https://gerrit.wikimedia.org/r/443451

Change 443104 merged by jenkins-bot:
[mediawiki/extensions/Wikibase@master] Make EntityReferenceExtractors configurable per entity type

https://gerrit.wikimedia.org/r/443104