EDAM validation using ontodev

[joncison]

From talking to Simon Jupp in Austria last week, we should explore https://github.com/ontodev/robot for EDAM housekeeping including validation, diff, slim generation etc.

Validation is provided by the report function - we'd need to provide a list of SPARQL queries for the specific checks we'd want (@simonjupp - could you pls. send a link to some sample queries - which you showed me last week)

There's also a file structure-independent ontology diff and lots of other stuff

It could augment or perhaps eventually replace edamxpathvalidator

cc @hmenager

[albangaignard]

👍 for SPARQL based quality control !

[hmenager]

Excellent. Thanks a lot for the pointer @simonjupp. edamxpathvalidator has been super useful, but as its author I'd be willing to replace it with something more generic and powerful any day. + it's probably the best alibi I can get to do some Java and SPARQL work!

[simonjupp]

All EFO checks are expressed as SPARQL here https://github.com/EBISPOT/efo/tree/master/src/sparql. Couple these with robot verify http://robot.obolibrary.org/verify.

Also check out the ontology starter kit that provide the inspiration much of the tools for all this https://github.com/INCATools/ontology-development-kit

[albangaignard]

I'll be very happy to help in "SPARQLing" with EDAM. Also, validation queries could be integrated into a jupyter notebook to extract / visualize metrics. A demo notebook is available here : https://github.com/albangaignard/fresh-toolbox .

[joncison]

@hmenager I propose something like

• [ ] rename https://github.com/edamontology/edamxpathvalidator/ to https://github.com/edamontology/edamvalidator/
• [ ] migrate this issue there (easy using GitHub issue mover utility)
• [ ] set up the new validator with a single simple validation to begin - but run this in parallel with edamxpathvalidator, until ...
• [ ] create existing validations in the new system. Retire use of edamxpathvalidator
• [ ] think hard / enumerate additional validations we need, adding one issue/validation

Does that sound about right?

Awesome if you can help @albangaignard ! I want to publish EDAM 2.0 in a good journal later in the year, and doing a systematic job on the validation would certainly be a ticket for that article!

[joncison]

Comprehensive validation would be to encapsulate:

• Guidelines from the Developers Guide for
  • adding concepts
  • hierarchy structure
  • deprecating concepts
  • Rules of thumb for EDAM development from the Editors Guide

[joncison]

it's this issue @albangaignard - pls. post the date of the Nantes meeting here.

see also https://github.com/edamontology/edamontology/issues/366 and https://github.com/edamontology/edamontology/issues/292 which are somewhat related.

[albangaignard]

Meeting in Nantes with @hmenager the 14th of february 2020. Goals :

• Identify validation rules
• Implement them in SHACL and/or SPARQL
• Automate the validation process (CI)

[joncison]

cool ... 6h+ journey (each way!) for me, so I will join remotely, if you can please set up a call-in facility ... cc @matuskalas (are you free?)

[joncison]

Hi @albangaignard and @hmenager

For our call today, I propose we follow the route @simonjupp @ EBI did for EFO ontology, i.e. use https://github.com/ontodev/robot. More specifically:

• robot report utility (http://robot.obolibrary.org/report) which operates with a useful list of general queries
• robot verify for our bespoke queries (http://robot.obolibrary.org/verify)

In summary (distilling the above thread):

• [ ] Decide on name for new utility. I propose edamverify
• [ ] Create repo https://github.com/edamontology/edamverify/ with README.md, and folders: src/ (script), queries/ (sparql queries), and /reports (query output)
• [ ] Install https://github.com/ontodev/robot so that it's invocable via Travis CI
• [ ] For general testing, add Travis CI test (invoked on commit of EDAM_dev.owl), using robot verify. Trigger test and check things are working OK.
• [ ] For initial bespoke testing, create a dummy query. See examples in https://github.com/EBISPOT/efo/tree/master/src/sparql
• [ ] Add Travis CI test (invoked on commit of EDAM_dev.owl), using the dummy query and robot verify utility (http://robot.obolibrary.org/verify). Trigger test and check things are working OK.
• [ ] Enumerate comprehensive list of QC checks, each with a name (==sparql file name), short description and GitHub issue (documented in table in queries/README.md) - capturing things from:
  • Existing checks from https://github.com/edamontology/edamxpathvalidator
  • Guidelines from the Developers Guide for
    • adding concepts
    • hierarchy structure
    • deprecating concepts
  • Rules of thumb for EDAM development from the Editors Guide

If we get time, we can also think about EDAM "vital statistics" generation.

Does that sound like a plan? Ping me here when you guys are settled.