This fixes a regression introduced in PR #71:
If the clone ID that we look up in the exclusion list contains "-" or "0" characters, we need to check only the other characters. This is now handled by introducing a SimilaritySet class that uses exact lookups as long as there are no "-" or "0", but falls back to the old algorithm otherwise.
This adds a minute or so to processing the exclusion list, so it is slower, overall still much better than runtime on the order of days.
This fixes a regression introduced in PR #71: If the clone ID that we look up in the exclusion list contains "-" or "0" characters, we need to check only the other characters. This is now handled by introducing a SimilaritySet class that uses exact lookups as long as there are no "-" or "0", but falls back to the old algorithm otherwise.
This adds a minute or so to processing the exclusion list, so it is slower, overall still much better than runtime on the order of days.