[BUG] normalize_severity crashes when given a Severity Enum Object

[initions-david]

📦 Environment

Python: 3.12, SparkDQ: 0.11.0

📝 Bug Description

The severity config option gets normalized by converting a string to the Enum and calling upper on it. This doesn't work, if the normalization function is passed an Enum. This happens in the below cases.

🔁 Steps to Reproduce

1. If the severity is configures via the Enum```python

from sparkdq.plugin.check_factory import CheckFactory
from sparkdq.core.severity import Severity

config_data = {
    "check-id": "test",
    "check": "null-check",
    "columns": ["foo"],
     "severity": Severity.WARNING
}

check = CheckFactory._from_dict(config_data)

2. If the config_data is used in more than one CheckFactory, because the config is mutated in place:

from sparkdq.plugin.check_factory import CheckFactory

config_data = {
    "check-id": "test",
    "check": "null-check",
    "columns": ["foo"],
    "severity": "warning",
}

check1 = CheckFactory._from_dict(config_data)
check2 = CheckFactory._from_dict(config_data)

✅ Expected Behavior

Return the Enum in the normalization step, if its already an Enum.

📋 Relevant Logs / Error Messages

💬 Additional Context

No response

[initions-david]

I lack the permissions to push the fix, but this is the required change in sparkdq/core/severity.py:

from typing import Union

def normalize_severity(value: Union[str, Severity]) -> Severity:
    """
    Convert string representations to validated Severity enum instances.

    Provides robust parsing of severity values from external sources such as
    configuration files, user input, or API requests. The function performs
    case-insensitive matching and provides clear error handling for invalid
    severity specifications.

    This normalization ensures consistent severity handling across all framework
    components while providing clear feedback for configuration errors.

    Args:
        value (Union[str, Severity]): String representation of the severity level,
            processed case-insensitively for flexible input handling, or an
            existing Severity enum instance (which is returned as-is).

    Returns:
        Severity: Validated severity enum instance corresponding to the input string.

    Raises:
        InvalidSeverityLevelError: When the input string does not match any
            recognized severity level, providing clear feedback for correction.
    """
    if isinstance(value, Severity):
        return value

    try:
        return Severity[value.upper()]
    except KeyError:
        raise InvalidSeverityLevelError(value)

[initions-david]

@flitzpiepe93 , @tongqqiu , @quatpv

We rely heavily on SparkDQ in our implementations, and I’d prefer to avoid forking the project so we can continue contributing to and benefiting from a shared codebase. However, not receiving any response for two weeks makes me a bit uneasy, especially when it comes to blocking or more critical issues.