Add windIO schema validation to CI

[kenloen]

It would be great to validate the windIO definition against the latest windIO turbine schema in CI to insure that it is following the newest windIO version.
I think it would be best if the validator could be strict and hence not allow unknown fields to ensure that the definition do not have any typos or data that is not part of the schema.

[gbarter]

Fully supportive of expanding the CI in general, however I don't think a strict validation is a good idea here. Models that use windIO will still need to add in model-specific geometry inputs. WISDEM does this for the hub and drivetrain, where windIO has not settled on a parametric geometric description, just the elastic model. In the reference turbine repositories, the yaml file should be sufficient for all of the models that are officially supported (HAWC2, OpenFAST, WISDEM, etc). We cannot achieve that and have a strict validator.

[kenloen]

I do understand that making a strict validation will mean that some data will need to go out for now as it is not part of the schema.
I think it is an issue to add data to the "windIO" file that is not part of the schema to ensure that some tool can run. It will make it hard to draw the line of which tools should be allow to add entries in order for the tool to run. I would prefer that this data would be moved over to the WISDEM input and then you can merge it into the windIO dict as a prepossessing step in WISDEM. Then, when we settle on a schema for the hub and drivetrain it can be moved back in.

It would be great to start the discussion of defining the hub and drivetrain in windIO by creating an issue on windIO.

I think this is pointing to a bigger discussion on what is the purpose of windIO, both in the short term and in the long term.

[fzahle]

we should be able to validate the turbine yaml with the updated version, right?

[gbarter]

We were discussing this and we are leaning to handling the validation within WISDEM, with a slightly expanded schema beyond the WindIO base (we have found the jsonmerge package to be helpful for this). I do see the value to a strict validator now, contrary to my comment 13 months ago. Individual model developers can choose to use that or roll their own approach.