abydos
abydos copied to clipboard
chrislit
Abydos NLP/IR library for Python
In the name of reproducibility, perhaps authors' original formulae should be presented in the docs -- alongside info to translate their notation to standard set/abcd notation.
# other alignment styles: # TODO: verify that these are functioning correctly, since they don't # match thesis cmp2 = ALINE(mode='local') self.assertEqual( cmp2.alignment('aHpakosiHs', 'waHpikonoHha'), [(120.0, '‖ aH p a k...
See: https://www.aclweb.org/anthology/A00-2038 Double-articulations, like /w/, get special handling according to the paper. This wasn't implemented in Abydos 0.4.0.
- [x] PAQ (via https://github.com/observerss/paq) - [x] LZSS (https://github.com/rumbah/pylzss) - [ ] PPMd (cf. #32) - [ ] ZPAQ? - [ ] LZ4 (https://github.com/python-lz4/python-lz4) - [ ] LZO (https://github.com/jd-boyd/python-lzo)
I really want a PPMd compressor for compression distance, whether it's a pure python solution or via a library (possibly my own, based on the reference implementation C code).
- [ ] Biological sequence analysis: Probabilistic models of proteins and nucleic acids (book)?: - [ ] Repeated matches (24) - [ ] Overlap matches (26)
Reports to send upstream: - [ ] Apache Commons Code: Refined Soundex needs to delete vowels - [x] ~~BMPM: Unreachable rules (check coverage report)~~ (Perhaps I misremember. I don't see...