Ensure mapped CURIEs are resolvable

cmungall commented 4 years ago

E.g.

 has phenotype:
    is_a: related to
    description: >-
      holds between a biological entity and a phenotype, where a phenotype is construed broadly as any kind of quality of an organism part, a collection of these qualities, or a change in quality or qualities (e.g. abnormally increased temperature).
    domain: biological entity
    range: phenotypic feature
    notes:
      - check the range
    in_subset:
      - translator_minimal
    close_mappings:
      # RTX contributed term tagged as 'inverse'
      - OMIM:phenotype_of
    exact_mappings:
      - RO:0002200
      - RTXKG1:has_phenotype
      - SNOMED:characterized_by
    broad_mappings:
      # Disease–presents–Symptom; symptom or finding may be a phenotype but not necessarily
      - hetio:PRESENTS_DpS
      - NCIT:disease_may_have_finding
      - OMIM:has_phenotype
    narrow_mappings:
      - NCIT:disease_may_have_molecular_abnormality
      - OBO:doid#has_symptom
      - RO:0004022

Are these all resolvable via the json-ld context?

The NCIT OPs have PURLs like http://purl.obolibrary.org/obo/NCIT_R115 is the doid hash correct?

RichardBruskiewich commented 3 years ago

@cmungall, all of the CURIE namespace prefixes seem to be mapped onto an base URI somewhere in the LinkML prefix resolution (via default_curi_maps, prefixes...) and thus, documented within the generated context.jsonld file. We've had a PR or two to clean up dangling prefix references.

However, not all URI's composed by these prefix base URI mappings are yet guaranteed to be automatically resolvable on the internet as as simple REST calls.

Some of the base URI's (e.g. SEMMEDDB, UMLS) simply point to a sensible web location that "owns" or indirectly documents the CURIE, but may not have (yet, if ever) given a resolvable URI template for the object identifier.

Others (e.g. GAMMA) were injected into the Biolink Model by Translator teams who may or may not have full definitions online for the associated CURIE terms.

Yet other controlled vocularies relevant to Translator and defined offline somewhere could be contemplated for inclusion (e.g. the ICEES and COHD 'feature variables' in clinical cohort studies with exposure data) but lack resolvable namespaces online.

The basic question here is how best to handle these cases within the Biolink Model, within Translator?

RichardBruskiewich commented 3 years ago

The NCIT OPs have PURLs like http://purl.obolibrary.org/obo/NCIT_R115. is the doid hash correct?

This particular URI does resolves to http://www.ontobee.org/ontology/NCIT?iri=http://purl.obolibrary.org/obo/NCIT_R115 (BTW, some browsers like Firefox can't seem to return anything for the resolved URI, while others (e.g. MS Edge) work... haven't figured out why ... not a problem local to us).

RichardBruskiewich commented 3 years ago

cmungall commented 3 years ago

Report to ontobee

On Mon, Jan 11, 2021, 11:21 Richard Bruskiewich @.***> wrote:

The NCIT OPs have PURLs like http://purl.obolibrary.org/obo/NCIT_R115. is the doid hash correct?

This particular URI does resolves to http://www.ontobee.org/ontology/NCIT?iri=http://purl.obolibrary.org/obo/NCIT_R115 (BTW, some browsers like Firefox can't seem to return anything for the resolved URI, while others (e.g. MS Edge) work... haven't figured out why ... not a problem local to us).

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/biolink/biolink-model/issues/508#issuecomment-758167802, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMON7UE2P7P3JKXFWGK3SZNFU5ANCNFSM4TNCLFNQ .

nlharris commented 3 years ago

What is the status of this? Is there work for Richard to do?

biolink / biolink-model

Ensure mapped CURIEs are resolvable #508