Fixed bug with reading unicode characters from output files

[MJWeberg]

Output HDF5 files with Unicode characters in internal filepath strings would sometimes fail to load with eispac.read_fit(). Added fallback method of decoding strings that patches the issue.