|RolandUnger added a comment.|
Thanks very much for your answers.
As I told now parsing takes now between 8 to 10 s, a few weeks ago it was only 4 s. I have no explanations for this phenomenon of the 100 % increase of the computing time. Is there anybody who knows the reason?
I estimated the timings from the profiling data in the html script comparing vCard templates with and without Wikidata usage. If I add Scribunto_LuaSandboxCallback::getEntity and Scribunto_LuaSandboxCallback::addStatementUsage times then I get about 5000 ms as I estimated it above. But I thought that most of the time was used for Scribunto_LuaSandboxCallback::getEntity not for accessing the claims table.
Our intention is to use Wikidata as much as possible. This is necessary for multiple use in a multi-language project like Wikivoyage. For instance, we are fetching almost all data from data sets like https://www.wikidata.org/wiki/Q42016015. So we have of course many (sometimes up to 20) accesses to the claims table. As it was told to me by the Wikidata team this is less expensive than using mw.wikibase.getBestStatements( id, p ) for each property. Maybe there was a change of the anEntity:getBestStatements( p ) function causing that considerable increase in the computing time.
Of course, we are using frame:extensionTag() once per template call. It is done in Module:Map generating a <maplink> tag. There is no other way to do this.
Cc: Anomie, Reedy, thiemowmde, Lydia_Pintscher, MaxSem, RolandUnger, Lahi, Gq86, Darkminds3113, GoranSMilovanovic, QZanden, Marostegui, LawExplorer, Vali.matei, Minhnv-2809, SundanceRaphael, Volker_E, Luke081515, Wikidata-bugs, aude, GWicke, Dinoguy1000, jayvdb, MrStradivarius, Jackmcbarn, Mbch331, Jay8g, Krenair
_______________________________________________ Wikidata-bugs mailing list Wikidataemail@example.com https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs