chenejac
commented
3 years ago Dominik Feldschnieders said:
I found the problem and we discussed it in the i18n slack. The problem is that the autocomplete works with the Solr index and the current Solr index is completely language-unaware.
So a solution would be, for every match, get the right label or filter the inappropriate labels in a SPARQL-Query. We think that would greatly slow down the autocomplete. Another solution would be to leave this issue for the next sprint/release, when we internationalize the whole search index.
Any thoughts on this?
Nicolas B Dickn (Migrated from VIVO-1879) said:
When adding a Geographic type entry, autocomplete should offer content according to the linguistic context selected.
In the dataset used for this test, the ontology only included English country names, so it was to be expected that using autocomplete in fr_CA context wouldn't pull "Suède". However, neither was is it expected to offer "Sweden", which it did.
I suspected the dataset might have caused the problem, so I created a "Suède" entry while in the fr_CA context, expecting it would have the correct linguistic label and wouldn't get pulled in the en_US context - but it did nonetheless.