Determine datatypes in results model of metadata service

[jmkeil]

Currently the metadata service returns sometimes String and sometimes Array for the following attributes. The datatype must be strictly determined to enable automated processing.

parameter	example terminology providing String	example terminology providing Array
date	CHEBI	FLOPO
versionIRI	ENVO	QUDT
saved-by	CHEBI	FLOPO
default-namespace	ENVO	FLOPO
type	ENVO	QUDT
comment	NCBITAXON	ENVO

Further, for some terminologies attributes are provided that are not listed in the API documentation. I recommend to guarantee the use Array for each attribute. This allows the automated processing even of unexpected properties.

By the way: In case of QUDT, due to the multiple included files, the result is a total mess.

[DavidFichtmueller]

Dear Jan Martin,
thank you for your feedback. However the data that is displayed on the /metadata page of a terminology comes from the terminology itself, it is not curated/modified by us. For the terminologies that were created for the GFBio project, we try to have the data harmonized, but all of the examples you stated above are from external sources that we just imported and where we do not want to modify this external content. The data shown at the main API point of a terminology (e.g. https://terminologies.gfbio.org/api/terminologies/QUDT ) is the curated metadata which we provide and harmonize. If you need any common information to be included in this interface as well, let us know and we might update the metadata for the respective terminologies. I hope this helps.

[naouelkaram]

Hi David, actually Jan Martin is right, we should have the same JSON output format also for internal metadata. I will change the output to array for all values.
Concerning QUDT the output is effectively a mess, thanks for the report.