Closed jakubklimek closed 8 years ago
The visualizer goes through all the languages (currently cs, en, and no language). Firstly, it shifts the currently selected language to first position. Then it loops through the list of languages and tries to find any label in the given ordering of languages.
Are all the labels present in both en and cs?
Could you provide a source of the data? I tried yesterdays TTL and LOD.cz and it's not there.
skos:prefLabel "10-14 let"@cs ,
"10-14 years"@en ,
"10-14 years"@cs ;
Ow... that is the temporary fix... ok, will clear that and try again
There's a bug in extraction as well, but this won't make it better...
OK, fixed the data (deleted the temporary fix), however, it is still happening.
Moreover, weird thing is happening when switching languages in menu:
Czech:
English:
Notice that only 0 and 25-29 changes, and it changes to skos:notation
, not a label
It's queried like this:
SELECT DISTINCT ?l ?spl ?sn ?sna ?st
WHERE {
OPTIONAL { <$uri> rdfs:label ?l . }
OPTIONAL { <$uri> skos:prefLabel ?spl . }
OPTIONAL { <$uri> skos:notation ?sn . }
OPTIONAL { <$uri> schema:name ?sna . }
OPTIONAL { <$uri> schema:title ?st . }
}
And labels are taken in this ordering: sn, l, spl, sna, st and then set to a map, so I think that the priority is actually reversed.
You can choose the expected priority, I will replace it.
This does not make sense because it behaves differently for each concept, even though each of the age categories, for example, has both skos:prefLabel
in Czech and English and skos:notation
.
As to the priority, it should be ?spl, ?sna, ?l, ?sn, ?dct, ?st where ?dct is dcterms:title
Extractor rewritten. It previously took just the first result from the result set, which was basically the randomization. Now it assembles the whole labels table and chooses the most suitable one.
The priorities are not correct. For regions from ruian, ?sn is displayed instead of ?sna.
Those are labels from dereferencing. Maybe the issue is there?
Related to https://github.com/ldvm/LDVMi/issues/93?