add descriptions and other annotations to extensions

[turbomam]

No description provided.

[mslarae13]

I think we can make a separate issue for this.. but is it really foodfood? Why is food twice?

[turbomam]

Good question. That is not a typo or bug. In this case, it is helpful to look at the submitter-provided title of this class: 'food-food production facility'. Pre-LinkML, an attempt was made to approximate hierarchy in the Extensions, as if there is a category of "food" Extensions, of which 'food production facility' is a child Extension.

Following standard naming for LinkML classes, the title 'food-food production facility' has been transformed into a term name of 'FoodFoodProductionFacility'.

I don't have any objection t changing that, but it should be done systematically.

[only1chunts]

@turbomam is correct, it was an attempt at hierarachy, the FDA were involved in designing a set of 4 extensions, so they grouped them all under a "Food" heading.

• MIxS-food-animal and animal feed
• MIxS-food-farm environment
• MIxS-food-food production facility
• MIxS-food-human foods

I'm not against renaming them, but there maybe parties that might be opposed to it, so I have opened a discussion ticket for it #779

[only1chunts]

I'm not sure what this file is, or how it relates to the updating of extension descriptions, but I trust its meant to be here?

[turbomam]

See notes on assets/mixs-schemasheets-concise.tsv

[only1chunts]

I'm not sure what this file is, or how it relates to the updating of extension descriptions, but I trust its meant to be here?

[turbomam]

This file, and the following status logs, are is generated by the the assets/mixs-schemasheets-concise.tsv target in project.Makefile

• mixs-schemasheets-concise-log.txt
• mixs-schemasheets-concise-report.txt

It is a tabular representation of almost everything in the MIxS schema. (It doesn't' include enumerations of permissible values).

• definitions of classes
• definitions of terms/slots
• assertions of how terms should be used within classes

[only1chunts]

I'm not sure what this file is, or how it relates to the updating of extension descriptions, but I trust its meant to be here?

[turbomam]

see comments about assets/isolate_slots.py

[only1chunts]

unable to view this as github says the diff is too large!

[turbomam]

See notes on assets/mixs-schemasheets-concise.tsv

[only1chunts]

unable to voiew this as github says the diff is too large.

[turbomam]

I had added a project.Makefile target for regenerating it that illustrates how this file is generated.

LinkML provides two mechanisms for applying a regular expression constraint to the values of a term/slot.

• the pattern meta-slot, in which you have to provide the exact regular expression that you consider valid for that slot
• the structured_pattern meta-slot, which allows you to compose the regular expression from previously-defined building blocks in the schemas's settings section.

To see how the structured_patterns are converted to regular expressions, you need to run the schema through some kind of materializing generator.

[only1chunts]

I'm not sure what this file is, or how it relates to the updating of extension descriptions, but I trust its meant to be here?

[turbomam]

This script extracts the global definitions of MIxS terms/slots from assets/mixs-schemasheets-concise.tsv. "Global" mean that the output does not indicate the cases in which one Extension class may. It also does a little more convenience filtering:

• removes the schemasheet header rows that start with a > character
• removes rows about classes
• removes rows about slots used by the purely infrastructural class MixsCompliantData, which nis used for validating tables of data that is believed to follow MIxS guidelines
• removes any columns if they don't contain any data values

It creates assets/mixs-schemasheets-concise-global-slots.tsv

I have commented out some lines that can insert the filtered table back into a relational database. This is especially useful for NMDC but I am glad to teach other people how to use it.

[only1chunts]

I'm not sure what this file is, or how it relates to the updating of extension descriptions, but I trust its meant to be here?

[turbomam]

I have updated project.Makefile to demonstrate how this file is built.

The result is a file that we discussed a few weeks ago: a tabular representation of textual annotations on the MIxS classes (Checklists, Extensions and Combinations)

[turbomam]

https://github.com/GenomicsStandardsConsortium/mixs/blob/646-classes-checklists-and-extensions-require-improved-descriptions/assets/class_summary_results.tsv

[turbomam]

That can be regenerated as a convenient overview any time we make changes to the YAML source of truth. It also demonstrates a format we could use to add more textual annotations to the classes or terms, as an alternative to free-form GH issues or word processing documents.

[only1chunts]

A couple of minor typos to correct, and a couple of files with changes that I'm not able to review as it says there are too many diffs to show?! not sure how to do review any other way.

[turbomam]

Good questions @only1chunts

I have replied, so you could re-review and choose accept or request changes as you see fit.

[turbomam]

Discussed in CIG meeting 2024-03-26

Thanks for the feedback. I will stick to smaller PRs next time!

Followup actions:

• integrate contributor and reference knowledge from
• work on separate release artifact directory (just a renaming of project/?) What files are required, in terms of structure and content?
• come to an agreement on structured formats for sharing attributes of entities (like Extensions). Something other than Google (word processing) Docs