Change ontology term identifier

[david4096]

Ontology terms had an ambigious 'id' field and description. This PR renames the field to term_id to remove the confusion over the role of the field, as well as improves the comment for its proper use. A field 'term_uri' has been added that allows one to specify the full URI for a term if it is available.

Closes #621 #165

[mbaudis]

+1 Clear improvement.

[sarahhunt]

+1

[mbaudis]

@david4096 (pls. don't forget the docs)

[david4096]

Thanks @mbaudis updated with some minor tweaks to the ontology documentation.

[david4096]

Related to discussion in #165

[mcourtot]

+1 with small modification (with input from @cmungall): "if no CURIE are available." -> "if no URI available".

[david4096]

Thanks @mcourtot! Updated!

[mbaudis]

Please check termId => term_id (if I understand the proto snake_case correctly); same for termUri => term_uri. I usually do things like that with serarch/replace over the whole tree; but don't tell anybody, please.

If fixed -> merge.

[david4096]

Thanks @mbaudis good to go!

[jmcmurry]

In the interest of reducing ambiguity, it may be worthwhile to have this field be termLabel:

[david4096]

Perhaps we can go forward with term_id and address the ambiguity in a later PR. We have improved over the field named id, however, there may be further room for improvement.

[david4096]

Thanks everyone for your input @jmcmurry @mcourtot @drseb ! I'll be happy to integrate any further work on this data model as consensus is reached.

For now, I'd like to see if we can move ahead and take further discussion to another issue. This PR was designed to improve the situation from id to term_id and improve the comments, which it has done. To continue the discussion check out the below links:

Ontology Source Identification
Multiple Ontology Term URIs
OntologyTerm.id definition

[jmcmurry]

That sounds good to me; thanks @david4096

[cmungall]

And -10 to any date scheme that is not ISO-8601!

[mbaudis]

@cmungall @drseb ... but reminding me of phenopackets/phenopacket-format#75

[mbaudis]

@mcourtot @cmungall @david4096 This should now be updated to reflect the Vancouver discussions.

[mcourtot]

Hm - I tried and push a commit to the pr/694 branch to reflect the Vancouver changes but don't see them here. Hadn't tried and commit over a PR before, so maybe i'm looking at the wrong place?

[david4096]

@mbaudis @mcourtot I've incorporated your changes from #726 .

Since we have removed the URLs, I believe we need a way to manage them, or at least specify a place where they might be looked up. What do you think @cmungall ?

[mcourtot]

Thanks @david4096! We're planning on using OLS, http://www.ebi.ac.uk/ols/index to bind prefixes to their URIs. @simonjupp, @tburdett, do you have a specific link?

[david4096]

Implemented in the server, client, and compliance.