[Bug]: Boolean values are incorrectly accepted by integer and float validators #258

Contact details

No response

What happened?

The Response class uses _validate_non_negative_int() and _validate_non_negative_float() to validate integer and float parameters. However, these validators use isinstance(value, int) and isinstance(value, (float, int)).

Because bool is a subclass of int in Python, boolean values such as True and False pass these validations.

As a result, parameters that are expected to be integers or floats can incorrectly accept boolean values.

Steps to reproduce

1. Create a Response object with a boolean value for an integer parameter:

Response(top_k=True)

2. Create a Response object with a boolean value for a float parameter:

Response(temperature=True)

3. Similarly, boolean values can be passed to:

Response(score=True)
Response(top_p=False)
Response(inference_time=True)
Response(tokens=False)

Expected behavior

Boolean values should be rejected when an integer or float value is expected.

For example:

Response(top_k=True)

should raise MemorValidationError rather than accepting True as the integer 1.

Actual behavior

Boolean values are accepted as valid integer/float inputs because Python considers bool a subclass of int.

Operating system

Linux

Python version

Python 3.11

Memor version

Memor 1.3

Relevant log output