What is `safe_float` supposed to do?

[vict0rsch]

I have removed the trailing code from distribution_benchmark.py and stability_benchmark.py but this was the orginal. What was the context / intention behind this function?

def safe_float(value):
    """Safely convert value to float, handling None and NaN."""
    return value
    # if value is None:
    #     raise ValueError
    
    # float_val = float(value)
    # if math.isnan(float_val):
    #     raise ValueError
    # return float_val

[smglsn12]

Initially this was set up to prevent any of the metrics from returning NaN (since the defaults were zero). I went through and changed this (because metrics that default to zero are very dangerous when zero is a target value for many metrics!!) but I never cleaned this up throughout the workflow afterwards. I'll go through and fix now!