Float/Double/FloatArray1D/DoubleArray1D parameters do not allow integer default values

[bkeryan]

Bug Report

In Python, int is generally duck-type compatible with float. In most cases, functions that expect a float will also accept an int.

Python type checking tools like mypy explicitly support "duck type compatibility" between int and float: https://mypy.readthedocs.io/en/stable/duck_type_compatibility.html

However, if a measurement plugin uses an int in the default value of a Float, Double, FloatArray1D, or DoubleArray1D configuration, it gets an error: TypeError: Unexpected type <class 'int'> in the default value for 'float_array_data'. Expected type: <class 'float'>.

Repro or Code Sample

@measurement_service.configuration("sample_rate", nims.DataType.Double, 1000)

Expected Behavior

It works.

Current Behavior

TypeError: Unexpected type <class 'int'> in the default value for 'sample_rate'. Expected type: <class 'float'>.

Possible Solution

Add this to _validate_default_value_type_for_basic_type:

if expected_type is float and isinstance(default_value, int):
    return

Update tests that check for an error in this case:

FAILED tests/unit/test_metadata.py::test___default_value_different_from_type___validate___raises_type_exception[DataType.Double-1-annotations6] - Failed: DID NOT RAISE <class 'TypeError'>
FAILED tests/unit/test_service.py::test___measurement_service___add_configuration_with_mismatch_default_value__raises_type_error[Float-DataType.Float-1] - Failed: DID NOT RAISE <class 'TypeError'>

Context

This came up in the code review for #807

Your Environment

• OS & Device: Windows
• ni-measurement-plugin-sdk-service version: 1.0.0
• MeasurementLink version: 24.8.x
• Python version: 3.9

AB#2819852