Add issue tracker item to SSSOM datamodel

[matentzn]

Fixes #78

• docs/ have been added/updated if necessary
• make test has been run locally
• tests have been added/updated (if applicable)
• CHANGELOG.md has been updated.

[sierra-moxon]

@matentzn - do you define the type "uri" in your schema outside of linkml_types? I don't think we have plain "uri" as a type. You can define it more rigorously as just a "uri" in your model (with all the regex you want to restrict that).

[sbello]

Is the intent that this would be a new field in the file where you could attach an issue URL to an individual mapping (e.g. a single row in the tsv file) or is this intended to be a new item in the header to point to the location of an issue tracker for the mapping project? I could see where either would be useful but it wasn't clear to me from the description.

I would also hope that if the field is for each mapping that this field is optional since we don't usually have issues for each mapping.

[matentzn]

@sierra-moxon thank you for your feedback, very much appreciated - what is the correct linkml type for URLs (rather than resources)?

@sbello thank you very much for your feedback. No SSSOM field aside subject, predicate, object and justification will ever be required. This is totally optional!

[matentzn]

What do you think about the name of the element, see first comment above?

[sbello]

Despite a desire to have a short label, I think 'issue' is too ambiguous. Possibly 'tracker_item', a bit shorter and I think still reasonably clear.

[gaurav]

Some other options:

• tracker
• tracked_at
• issue_tracker

And, in a pinch, seeAlso.

[sierra-moxon]

found the schema doc for LinkML types: https://linkml.io/linkml-model/linkml_model/model/schema/types.yaml -- it looks like we've got the types you're looking for!

[matentzn]

Thank you @sierra-moxon, @sbello and @gaurav

I think we should opt for the more verbose "issue_tracker_item", just because the issue tracker URLs are anyways super long, so a long element name wont really make much difference, and I kind of like the fact that it's not too ambiguous.

If you think I am wrong, let me know!

@sierra-moxon thanks for sharing the type spec. Do I see this correctly then that uri as a type is fine (if I import the linkml datatypes, which I do)?

[matentzn]

I will leave this open until Monday night (31st July), then I will merge so I can add this to the OAEI report. Thanks @gouttegd for your suggestions and review!