search

SSHOC / vocabularies

0 stars 0 forks source link

Use always english labels at first when ingesting vocabularies #19

Closed dpancic closed 1 year ago

dpancic commented 3 years ago

In GitLab by @KlausIllmayer on Jun 9, 2021, 15:58

Currently, when ingesting a vocabulary that has multiple languages as literals for skos:prefLabel (for the Concepts) and rdfs:label/rdfs:comment (for the ConceptScheme) the backend will only take the first found literal regardless of the language-tag, e.g. skos:prefLabel "Deutsch"@de, "German"@en, "allemand"@fr; will take only "Deutsch" for the concept in MP.

Instead it should take english (if there is a language tag @en) and only if there is no english literal found, it should take the first one.

This issue has some priority because it may compromise the ingest of items (when searching for the label) and it affects the output in the frontend.

It involves: skos:prefLabel for skos:Concept and for skos:ConceptScheme rdfs:label as well as rdfs:comment (but should be good practice for all language string literals.

In general, we may need to think about if we like to support multi-lingual vocabularies.

Opinions? @vronk @tparkola @vronk @laureD19

dpancic commented 3 years ago

In GitLab by @tparkola on Jun 10, 2021, 07:43

@KlausIllmayer , attach a file with multilingual vocabulary please.

dpancic commented 3 years ago

In GitLab by @KlausIllmayer on Jun 10, 2021, 10:00

The vocabulary where I discovered this issue is this one: https://gitlab.gwdg.de/sshoc/vocabularies/-/blob/master/iso-639-3/iso-639-3.ttl - please download it from there directly.

I've ingested the vocabulary as it was on the development and if you look there at /api/vocabularies/iso-639-3 you will see the label of the vocabulary in German: "ISO 639-3 Sprachcodes", because it was there on first position. I changed the ttl so that English is on the first position and ingested it on the stage and there you will see for /api/vocabularies/iso-639-3 that the label is now in English: "ISO 639-3 Language Codes". The same is true for the labels of the concepts.

dpancic commented 3 years ago

In GitLab by @KlausIllmayer on Jun 16, 2021, 15:55

mentioned in issue sshoc-marketplace#66

dpancic commented 3 years ago

In GitLab by @KlausIllmayer on Jun 21, 2021, 19:02

mentioned in issue data-ingestion#3

dpancic commented 3 years ago

In GitLab by @tparkola on Jun 23, 2021, 23:20

dpancic commented 3 years ago

In GitLab by @KlausIllmayer on Jul 2, 2021, 16:32

just a brief feedback: works as expected - thanks!

and a sidenote: if doing an update (PUT) of a vocabulary, changes will be only visible after calling /api/concept-reindex