Requests for improved performance of LinkML tools in MIxS

[turbomam]

MIxS is a large schema. There are hundreds of classes (including the checklist/extension combinations) and about 1000 MIxS terms/LinkML slots. Some of the tools that run in our build process are slow, especially

• [ ] validating example data whit linkml-run-examples
• [x] generating the documentation pages (temporarily solved by eliminating classes-using-terms reports)

[turbomam]

@turbomam to provide metrics

[cpauvert]

Hi @turbomam, thanks for the ongoing efforts on the mixs, really appreciated! I also notice that the validation was slow when using the example (with linkml v1.6.3).

• the schema: https://github.com/GenomicsStandardsConsortium/mixs/blob/main/src/mixs/schema/mixs.yaml
• the example: https://github.com/GenomicsStandardsConsortium/mixs/blob/main/examples/MixsCompliantData-MimsSoil-example.yaml

linkml-validate -s mixs.yaml MixsCompliantData-MimsSoil-example.yaml

with compliant example

real    2m21,002s
user    2m20,868s
sys 0m0,112s

with incorrect metadata

I changed https://github.com/GenomicsStandardsConsortium/mixs/blob/18794c4997aeb189d442d965beac00da073e8664/examples/MixsCompliantData-MimsSoil-example.yaml#L5 to an invalid text : "Column 1"

real    2m19,171s
user    2m19,085s
sys 0m0,072s

This was tested on a Xubuntu 20.04 LTS 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz 32,00 GB (2x 16GB DDR4

Hope that helps! Best,