Simplify float validation logic

[molisani]

Resolves #58

In is_float_none, the except block is not necessary. It exists to capture a ValueError and return None if value is None or "". However, value can never be None here because float(None) throws a TypeError (see issue #58) which is not caught. As well, if value is "" then we return None and re-raise the error for any other value.

Instead, this conditional can be moved to the beginning of the function. Then, the float(value) will continue to raise exceptions for all non-"" values, and None will correctly return None.

[jarondl]

Thank you!