Mass-curate OLS redirects, e.g. nmrCV

[sneumann]

Hi,

I found an issue on https://registry.identifiers.org/registry/nmr
and suspect there could be something systematic, and if so,
this could be mass-curated for other affected identifiers.org prefixes.

For example, https://identifiers.org/NMR:1000003
redirects to
https://www.ebi.ac.uk/ols4/ontologies/nmrcv/terms?short_form=NMR:1000003
which in turn gives a "not found" redirecting to
https://www.ebi.ac.uk/ols4/api/v2/ontologies/nmrcv/classes?shortForm=NMR:1000003&includeObsoleteEntities=true

I can't tell where the underlying issue is (the identifiers.org entry or the OLS4 side ?!),
the correct URL after all redirection would be
https://www.ebi.ac.uk/ols4/ontologies/nmrcv/classes/http%253A%252F%252FnmrML.org%252FnmrCV%2523NMR%253A1000003

and maybe @henrietteharmse could chime in.

Thanks,
Yours,
Steffen

[renatocjn]

Hi Steffen,
When OLS moved to OLS4, there were issues with some ontologies. I opened this ticket at the time but never got around to fixing the remaining ontologies because I was waiting for any further evolutions and other tasks came up. Thank you for reminding me. I see some ontologies still result in not found, any help from the OLS team would be appreciated.

[renatocjn]

Current list of namespaces with OLS resources: bao, biomodels.kisao, biomodels.teddy, bto, cco, chebi, cheminf, cl, clo, doid, eco, edam, efo, envo, eo, fma, foodon, go, hp, ido, ma, mamo, maxo, mcro, mi, mod, mp, ms, ncit, nmr, obcs, obi, omit, opl, orphanet.ordo, pato, po, pr, probonto, pw, ro, sbo, sio, so, stato, taxonomy, uberon, uo, vario

[henrietteharmse]

@sneumann The issue is with the short_form specified. Short_forms have this format: NMR_1000003. That is, it should use an _ rather than a :.

Hence, this works: https://www.ebi.ac.uk/ols4/ontologies/nmrcv/classes?short_form=NMR_1000003.

If you prefer to use curies, you could use: https://www.ebi.ac.uk/ols4/ontologies/nmrcv/classes?curie=NMR:1000003.

[renatocjn]

Thank you for the information @henrietteharmse !

We, unfortunately, can't use underscores since identifiers.org only allows for the colon or the slash as separating characters.

Is the curie query parameter available for all ontologies?

[henrietteharmse]

Yes, the curie parameter is available for all ontologies!

[renatocjn]

@henrietteharmse quick question on edam. It seems that the identification scheme used is a bit different from how it is showing in bioportal. In bio portal it uses the current regex setup in identifiers.org (^(data|topic|operation|format)\_\d{4}$), but the in OLS it seems to follow a curie format with a numerical number.

For miriam for example: bioportal uses data_1664, but OLS has several IDs for miriam. Could you advise on how to update our registry? For now I deactivated it since the url pattern is not valid.

[henrietteharmse]

Yes, there is indeed several terms in the EDAM ontology that matches on Miriam. See EDAM ontology source.

Use IRI

The fail-safe way to get this consistently from OLS for EDAM is to use an IRI. I.e. in this case the IRI you are interested in is
http://edamontology.org/data_1164.

To access the OLS UI, use the following where the IRI is double URL encoded:
https://www.ebi.ac.uk/ols4/ontologies/edam/classes/{double urlencoded IRI}

==>
https://www.ebi.ac.uk/ols4/ontologies/edam/classes/http%253A%252F%252Fedamontology.org%252Fdata_1164

Or, if you are using the API use:
https://www.ebi.ac.uk/ols4/api/ontologies/edam/terms/{double urlencode IRI}

https://www.ebi.ac.uk/ols4/api/ontologies/edam/terms/http%253A%252F%252Fedamontology.org%252Fdata_1164

Use Curie

If you want to use curies, you can use the following:

For UI: https://www.ebi.ac.uk/ols4/ontologies/edam/classes?curie=EDAM:1164
For API: https://www.ebi.ac.uk/ols4/api/ontologies/edam/terms?curie=EDAM:1164

However, with curies there are some ambiguity.

[renatocjn]

Awesome, thank you @henrietteharmse, that greatly helped.

I went throught he list from my original ticket from OLS, and the only problem I'm having is that some ontologies aren't available on OLS aparently. Here's a list of the ones I'm having trouble finding:

• https://www.ebi.ac.uk/ols/ontologies/vario
• https://www.ebi.ac.uk/ols/ontologies/eo
  Unless I'm looking at the wrong place.

[henrietteharmse]

VARIO and EO are no longer maintained and marked as inactive by the OBO Foundry. See #453. Therefore they are not on OLS as they are not recommended for use any longer.

[renatocjn]

Awesome, thank you for all the help @henrietteharmse

@sneumann may I close the issue?

[sneumann]

https://identifiers.org/NMR:1000003 now works as expected, thanks for fixing ! Yours, Steffen