joom / Guguk

Turkish NLP library in Haskell.

Guguk

Turkish NLP library for Haskell. (pronounce: "goo gook".)

Note that this is a personal pet project, heavily influenced by the mighty zemberek-nlp.

Progress

• Syllabification (in Guguk.Syllabification)
  • Passes all the tests.
• Phonetics (in Guguk.Phonetics)
  • More usable set of functions for the existing data and types is needed.
• Turkish Alphabet (in Guguk.TurkishAlphabet)
  • ASCIIfying, deASCIIfying functions etc. needed.
• Phonology (in Guguk.Morphology.Phonology)
  • More usable set of functions for Turkish phonology and morphotactical rules.
• Tokenization (Guguk.Tokenization)
  • Basic functionality for Sentence boundary detector. (TODO: Handling ":" and "...", and changing from String to Text) This can be rewritten using Parsec.
  • Lexer needed.
• POS Tagger (Guguk.Syntax.PosTagger)

Contribution

I'm very open to any pull requests, issues or other kinds of suggestions. Feedback is especially important since I'm neither a Haskell nor Turkish NLP expert.

Projects that use Guguk

• Divan.hs: Ottoman Divan poetry vezin checker

License

MIT License