Closed danielhuppmann closed 1 year ago
I'm all for consistency so CodeList.validate()
sounds like a good idea me. I'll rename it accordingly and remove the pyam version pin.
Not sure it's really necessary to go for a full new release of nomenclature yet. The process
function does not yet take a RequiredDataValidator
yet. I can open a PR for that including docs and then we can create a new release.
Ok, sounds good, thanks!
Just wanted to create two small PRs one for the pyam version and one for renaming CodeList.invalid_items
to CodeList.validate
.
Turns out the second one won't work since CodeList
inherits from pyam.BaseModel
and that already defines a validate
method. I could use validate_items
instead or just keep it as is since we're not going to get it all aligned then. Alternatively we could also rename the validate
method in DataStructureDefinition
.
Good point... I guess validate_items()
sounds better than invalid_items()
... I would not change the DataStructureDefinition any more unless strictly necessary.
Closing as release v0.8 is now live.
With a new release of pyam v1.7.0 just out including the new
require_data()
method, it's time to fix the pyam-pin and then create a new release!One thing to think about before the release: whether to stick with the method name
CodeList.invalid_items()
(see here) or call that methodvalidate()
for consistency?So
DataStructureDefinition.validate()
validates against an entire IamDataFrame, andCodeList.validate()
validates against a list of items...What do you think @phackstock?