Problem:
Currently, the search index text contains the complete json, including the hidden fields and all the labels. So searching for schematext or shexc would bring up every Schema.
acceptance criteria
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Improve the search index text for Schemas | mediawiki/extensions/WikibaseSchema | master | +41 -0 |
Change 501534 had a related patch set uploaded (by Michael Große; owner: Michael Große):
[mediawiki/extensions/WikibaseSchema@master] Improve the search index text for Schemas
Change 501534 merged by jenkins-bot:
[mediawiki/extensions/WikibaseSchema@master] Improve the search index text for Schemas
Note for verification: This might need a rebuild of the search index and it could be that it only works on newly edited pages correctly.
Strange discovery:
A search for Änderung isn't showing a snippet: https://wikidata-shex.wmflabs.org/w/index.php?search=%C3%84nderung&title=Special%3ASearch&profile=advanced&fulltext=1&ns640=1
A search for Testen is showing a snippet: https://wikidata-shex.wmflabs.org/w/index.php?search=Testen&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1&ns640=1
Do we have some kind of Unicode problem here? 🤨
Looks like regular Wikitext pages have the same problem: https://wikidata-shex.wmflabs.org/w/index.php?search=Änderung&ns1=1
I assume on Wikidata it’ll behave differently due to CirrusSearch anyways 🤷