Collator.lstripwords: Strip words off the start and append to the end. Collate.unicode, Collate.str: Convenience recoding routines.
strings: Microoptimizations, saves about 10% of runtime.
strings: Include deroman in import list.
Roman numeral parsing. More test cases. (Fixes issue #3)
strings.sortemes: Use a line break to separate letters and non-continuing-characters and a paragraph break to separate letters and numbers. (Fixes issue #2)
More release preparation. Docstrings and consistency work.
Fix typo, remove unneeded check.
Simplify algorithm. No changes to sort behavior or runtime.
Different algorithm, gives better results for numbers following grouping numbres but worse results for numbers following other punctuation.
sortemes: Simplify some logic.
More tweaks; notably try to insert paragraph breaks rather than a separate Python tuple when re-concatenating strings.