The license_serializer program normalizes and serializes the known
licenseclassifier licenses into a compressed archive. The hash values for
the licenses are calculated and added to the archive. These can then be used
to determine where in unknown text is a good offset to run through the
Levenshtein Distance algorithm.