Closed aclum closed 2 months ago
analysis_type
is "recommended" in the NMDC schema definition
Then, in the submission schema, modifications_long.tsv updates it to "required" for all interface classes that use DhMultiviewCommonColumnsMixin
(which is all the interfaces)
Except instead of replacing "recommended" with "required", both stick around.
As another example, the same thing is happening with dna_concentration
in JgiMgInterface
. dna_concentration
is "recommended" in the NMDC schema definition, and then updated to be "required" in modifications_long.tsv. The result in the submission schema LinkML source is this:
dna_concentration:
name: dna_concentration
title: DNA concentration in ng/ul
comments:
- Units must be in ng/uL. Enter the numerical part only. Must be calculated using
a fluorometric method. Acceptable values are 0-2000.
examples:
- value: '100'
from_schema: https://w3id.org/nmdc/nmdc
see_also:
- nmdc:nucleic_acid_concentration
rank: 5
owner: Biosample
domain_of:
- Biosample
- ProcessedSample
- JgiMgInterface
slot_group: jgi_metagenomics_section
range: float
required: true
recommended: true
minimum_value: 0
maximum_value: 2000
tl;dr I don't know how to fix this but it's happening in a bunch of places in addition to WaterInterface
Thanks for looking. Should we Zoom and look at it together?
@sierra-moxon @turbomam @bmeluch Will you have time to work on this this sprint? Do you think it'll take a couple sprints to identify the issue & resolve it?
I looked some more at modifications_long.tsv.
As far as I can tell, the modification action "replace_attribute" should really be named "replace_value". It's updating the value of an attribute but the modifications_and_validation
function in sheets-and-friends doesn't actually change the attribute that's there.
People who know about sheets-and-friends, does that sound right? Could we add another possible value for "action" that makes modifications_and_validation
work with two attribute names, rather than an attribute name and a value?
From the metadata meeting today, 24 Apr 2024, we decided that the most straightforward fix for this is to update modifications_long.tsv so that wherever it adds "required = true" to a slot with "recommended = true" in nmdc-schema, it also updates "recommended = false". Both attributes will still show up in the submission-schema documentation page for slots with this modification, but they will no longer be redundantly both true.
I don't think this is fixed... recommended:true is still appearing
@bmeluch @turbomam @sujaypatil96
https://microbiomedata.github.io/nmdc-schema/analysis_type/ https://microbiomedata.github.io/submission-schema/analysis_type/
oh no! It looks like the Github Pages documentation site hasn't been updated in a year, if I'm reading this right? Maybe it just hasn't captured recent fixes?
This looks right on the water interface which is I think all we can hope for. This was changed in modifications_long.tsv not in nmdc-schema itself. analysis_type: name: analysis_type description: Select all the data types associated or available for this biosample title: analysis/data type examples:
Nvm, @aclum is right. This is a correction at the interface level, not the slot level, and the interface pages now read correctly. @mslarae13 we're good!
Reviewed with Bea. I was looking at https://microbiomedata.github.io/submission-schema/analysis_type/ & needed to look at the individual (type)Interface
pages like https://microbiomedata.github.io/submission-schema/BiofilmInterface/
Separately, we could start a discussion around why analysis_type is only recommended in nmdc-schema. Not required for this issue
See https://microbiomedata.github.io/submission-schema/WaterInterface/ If I look at the linkml source for slot analysis_type for this class has both required and recommended are both set to true. analysis_type: name: analysis_type description: Select all the data types associated or available for this biosample title: analysis/data type examples: