Add Check: Dataset Attributes should be fixed-length strings

[nemo794]

Currently, the QA XML checker reads in each Dataset's Attribute keys and values, only performing lightweight conversions if needed:

nisarqa/src/nisarqa/utils/file_verification/h5_parser.py

Lines 41 to 47 in 1fd188c

	for key in hdf5_object.attrs.keys():
	attribute = hdf5_object.attrs[key]
	# Attributes are held in the
	if type(attribute) in [bytearray, bytes, np.bytes_]:
	attributes[key] = attribute.decode("utf-8")
	else:
	attributes[key] = attribute

This allows Attributes with variable-length string dtypes to pass through, and go un-logged. By NISAR conventions, all strings should be saved as as fixed-length bytes strings (or lists of fixed-length byte strings).

So, we need to add a check that string Attributes are either correctly fixed-length byte strings (or lists of fixed-length byte strings), or that an error is logged.

[nemo794]

fwiw, the R4.0.4 sample products (GSLC, GCOV, all InSAR) contain some Attributes which would fail this check.