Fix kinematic vs kinematic_coverage check

[Radonirinaunimi]

Currently, when kinematic_coverage does not contain 3 variables and len(kinematics.variable)<=3, extra_ variables are added to comply with vp.

nnpdf/validphys2/src/validphys/commondataparser.py

Line 399 in 8b9b29f

nkincov = self.kinematic_coverage + [f"extra_{i}" for i in range(diff_to_3)]

This, however, could be inconsistent with the following check:

nnpdf/validphys2/src/validphys/commondataparser.py

Lines 435 to 439 in 8b9b29f

	for var in self.kinematic_coverage:
	if var not in self.kinematics.variables:
	raise ValidationError(
	f"Variable {var} is in `kinematic_coverage` but not included in `kinematics` for {self.name}"
	)

This is a minimal fix by basically skipping extra_ whenever kinematics and kinematic_coverage are compared.

[scarlehoff]

thanks!
I guess you were the first one trying to use a dataset with only two variables :__)

We need to get rid of this limitation of "3 and only 3" though...

[Radonirinaunimi]

thanks! I guess you were the first one trying to use a dataset with only two variables :__)

We need to get rid of this limitation of "3 and only 3" though...

Yes, I fully agree! In the meantime, I just added it to the holy list #1974 xd