search

acdh-oeaw / pd-labs

Scripts and data for exploring and curating Parthenos Discovery
MIT License
0 stars 0 forks source link

CLARIN WebServices are typed as Volatile Dataset #1

Open vronk opened 4 years ago

vronk commented 4 years ago

Looking for example at the WebLicht Webservice Orchestrator, it has items (e.g. TextCorpus2Lexicon, in original context, typed as pe:PE24_Volatile_Dataset, even though they are actually web services described using the profile clarin.eu:cr1:p_1320657629644.

This is due to an error in the automatic generation of CMDI mappings, which default to Volatile Dataset. There actually is a special handling for web service in place, which should kick-in based on an explicit manual list of profiles: p_1295178776924,p_1299509410083,p_1311927752306,p_1320657629644,p_1360230992146,p_1423750293168,clarin.eu:cr1:p_1469541567396,p_1505397653795

We need to investigate, why this special handling wasn't applied.

How many items are affected: According to the entry in PD, there are 66 records based on the WeblichtWebService profile. According to curation module there 524 records from 10 providers. So it is also to be inspected, why there is such a big difference in the number of records.

wowasa commented 4 years ago

automatic generation of cmdi mapping couldn't work for some profiles because the configuration is wrong (you may not add full profile if »clarin.eu:cr1:p1469541567396« to the CONDITIONS property because colons are used as condition(s)-profiles separator but only the p[0-9]+ part)