Repository files navigation

(C) 2018 byDamir Cavar

Version: 0.1

This is part of my Julia code and libs for Natural Language Processing (NLP). I am using Julia 1.0 or newer.

This is the beginning of an interface to theFoma library in Julia.

The code here is accompanied by an example morphology in form of aFoma Finite State Transducer.

english.fst

There are a few morphemes and words in this morphology, just for testing purposes.

Make sure thatFoma is installed on your machine, and in particular the dynamic libraries. These libraries needto be in your system's library path such that Julia can find them.

This is an extremely fast morphological analyzer. This combination of Julia and Foma FST-based morphological analysiscan process more than 300,000 tokens per second on modern Intel i7 CPUs on a single thread.

I will extend the library and functions soon.

For any token that you process using the FST, the output will be of the form:

call+N+Plcall+V+3P+Sg

This is the output for the input token "calls". The two lines mean that:

• there are two analyses forcalls
• the lemma forcalls is in both casescall
• the main part of speech isN (noun) andV (verb)
• the nouncalls has a morphosyntactic feature+Pl (+plural)
• the verbcalls has two morphosyntactic features, that is+3P (third person) andSg (singular)

Test some more examples using the compiled mini-morphology for English inenglish.fst.