automated cleanup of python pep8 whitespace compliance
This fix makes the handling of type 'float' and 'double' the same. The implementation requires the C99 isfinite() macro, or otherwise some platform dependent equivalents, to be available.