feat[next] Enable embedded field view in ffront_tests #1361
Conversation
| @@ -705,13 +706,33 @@ def __call__( | |||
| ) | |||
There was a problem hiding this comment.
| ) | |
| self.as_program(arg_types, kwarg_types)( | |
| *args, out, offset_provider=offset_provider, **kwargs | |
| ) | |
| return out |
Just noticed this bug.
There was a problem hiding this comment.
Should field_operator as program return the out param? Maybe yes...
There was a problem hiding this comment.
I would say no as programs don't return anything.
There was a problem hiding this comment.
than it was correct, right?
There was a problem hiding this comment.
No, the program does not return something, but the field operator which we are calling does.
There was a problem hiding this comment.
no, we return self.as_program()() i.e. whatever program does which is currently nothing therefore still correct and in case (for whatever reason) program() returns something, we will do the right thing implicitly.
tests/next_tests/integration_tests/feature_tests/ffront_tests/test_math_builtin_execution.py
Show resolved
Hide resolved
src/gt4py/next/ffront/fbuiltins.py
Outdated
| ) -> common.Field | core_defs.ScalarT: | ||
| assert core_defs.is_scalar_type(lhs) | ||
| assert core_defs.is_scalar_type(rhs) | ||
| return BINARY_MATH_NUMBER_BUILTIN_TO_PYTHON_SCALAR_FUNCTION[name](lhs, rhs) # type: ignore[operator] # Cannot call function of unknown type |
There was a problem hiding this comment.
| return BINARY_MATH_NUMBER_BUILTIN_TO_PYTHON_SCALAR_FUNCTION[name](lhs, rhs) # type: ignore[operator] # Cannot call function of unknown type | |
| return getattr(np, name)(lhs, rhs) |
Works just fine for scalars and the dict (with the clumsy name) is not needed in that case.
There was a problem hiding this comment.
no, that's not what it does. it maps to python builtins, but thanks to our clever naming, they have different name...
There was a problem hiding this comment.
I feel I'm missing something obvious. np.minimum is an extension of the functionality of min, so I don't understand your first sentence. Unless my assumption is wrong our naming is indeed clever as we chose the numpy name for all builtins and not sometimes the python ones in case they exist.
There was a problem hiding this comment.
Ah now, I get it, you want to do the scalar functions using numpy?
Enables field view in ffront_tests
New exclusion markers for some cases
Adds the following features to embedded:
__ne__and__eq__to FieldTODOs: