[REQ][python-nextgen] make float parsing more lenient

[tnielens]

If a response json field is an int (1), although the target field type is float (1.0) , the response parsing fails with an error like

ValidationError: 1 validation error for <ClassName>
<fieldName>
  value is not a valid float (type=type_error.float)

Describe the solution you'd like

Can the parsing be more lenient? Former python generators parsed this more leniently.

Describe alternatives you've considered

None

Additional context

I'm trying to use the latest python client code generator. Btw, great job on the source compatibility 👍 .

[spacether]

Fyi if you are open to using the python generator we have a CI verified test for this use case at:
https://github.com/OpenAPITools/openapi-generator/blob/master/samples/openapi3/client/3_0_3_unit_test/python/test/test_models/test_integer_type_matches_integers.py#L50

    def test_a_float_with_zero_fractional_part_is_an_integer_passes(self):
        # a float with zero fractional part is an integer
        IntegerTypeMatchesIntegers.from_openapi_data_oapg(
            1.0,
            _configuration=self._configuration
        )

[tnielens]

@spacether thanks. We have many projects based on clients generated by the python-prior generator. The python-nextgen sources look more compatible than the current python one. Also the models are based on pydantic which we have experience with and appreciate. python-nextgen looks like a better fit for us.

[nmoreaud]

It seems legitimate to parse {"field": 1} as a valid float value.

Note: generated code works fine for me if I post-process it like this:

find models -iname '*.py' -type f -print0 | xargs -0 sed -r -i 's/: StrictFloat/: float/g'
find models -iname '*.py' -type f -print0 | xargs -0 sed -r -i 's/: Optional\[StrictFloat\]/: Optional[float]/g'

[wing328]

I've filed #14618 to add an option to make float's strict type configurable.

[tnielens]

Tested successfully on 3746a20.