decision about section & column ordering semantics

[turbomam]

I.e. as an alternative to using annotations (dh_sect_ord, dh:section_name and dh:column_number)

• dh:section_name annotation on a column -> slot_group
• dh:column_number annotation on a column-> rank
• dh_sect_ord annotation on a section class -> rank parent slot

Especially don't use the parent class nmdc_dh_section in whatever mechanism we come up with

This issue should deprecate several older issues mentioning things like subsets #2, #58

• How is Damion doing this in https://github.com/cidgoh/DataHarmonizer/tree/linkml-datastructure? Look for linkml*py scripts or communicate with email or Zoom. Make an issue with our suggestions?
  • two methods for making the JS portion of DH interface:
    • currently using data.tsv -> data.js with make_data.py in this repo. We do Google Sheets -> LinkML -> data.tsv before that)
    • Damion has his own LinkML -> JS branch, which has its own section (parent slots/is_a) and column ordering methods

@turbomam's idea:

group columns into sections with https://linkml.io/linkml-model/docs/slot_group/#slot-slot_group

remember that parent slots for each section will have to be created (as opposed to our current sections_as_classes tab from the sheet 'nmdc-dh-sheets')

remember that the slot titles can use spaces, punctuation and mixed case but the slot names shouldn't order the columns with https://linkml.io/linkml-model/docs/rank/

[turbomam]

All of these modelling changes will require changes to our LinkML -> DH code, or a switch to Damion's LinkML -> DH code if we like it

[turbomam]

But we shouldn't make changes to the Google Sheet that's currently in use!

• [x] clone the existing nmdc-dh-sheets into a new nmdc-dh-sheets-sandbox
• [x] change these in Makefile?
  • [x] nmdc_schemasheet_key = 1RACmVPhqpfm2ELm152CzmiEy2sDmULmbN9G0qXK8NDs
  • [x] credentials_file = local/nmdc-dh-sheets-0b754bedc29d.json

May need to reconfigure cogs and generate and other server account file? Or grant permissions on a new Google Sheet to the existing authentication file?

[turbomam]

see also #56

[turbomam]

Tab sections_as_classes from nmdc-dh-sheets defines several DH sections, but many of those are not used (X)

title	Used in
Sample ID	new_terms
GOLD ecosystem path	X
EMSL	new_terms
JGI-Metagenomics	new_terms
JGI-Metatranscriptomics	new_terms
Metadata- MIxS Modified Required	X
Metadata- MIxS Required	X
Metadata- MIxS Modified Required Where Applicable	X
Metadata- MIxS Required Where Applicable	X
Metadata- MIxS Modified Optional	X
Metadata- MIxS Optional	X

What sections really do appear in the soil_emsl_jgi_mg interface?

Just now ran

• make squeaky_clean
• make all

In branch main, against nmdc-dh-sheets (nmdc_schemasheet_key=1RACmVPhqpfm2ELm152CzmiEy2sDmULmbN9G0qXK8NDs)

DH sections are found in docs/template/soil_emsl_jgi_mg/data.tsv, with parent class=''

Several of the sections aren't defined anywhere. That's not getting caught now as we are defining sections with annotations. Defining sections will be essential when we switch to slot_group

sections	notes	Defined in	Used in
Sample ID		sections_as_classes	new_terms
EMSL		sections_as_classes	new_terms
JGI-Metagenomics	Same for JGI-Metatranscriptomics (just not used in soil_emsl_jgi_mg DH template)	sections_as_classes	new_terms
MIxS		X	sections_columns_orders
MIxS (modified)	sections_columns_orders needs to be refactored wider, with changes to sheets_and_friends/mod_by_path.py:mod_by_path	X	sections_columns_orders
MIxS Inspired		X	new_terms