Finnish syllabifier and compound segmenter


Keywords
Finnish, syllabifier, compound, segmenter
License
BSD-3-Clause
Install
pip install FinnSyll==1.0.dev4

Documentation

https://travis-ci.org/tsnaomi/finnsyll.svg?branch=master

FinnSyll

FinnSyll is a Python library that syllabifies words according to Finnish syllabification principles. It is also equipped with a Finnish compound splitter. More details/docs to come.

Installation

$ pip install FinnSyll

Basic usage

First, instantiate a FinnSyll object.

>>> from finnsyll import FinnSyll
>>> f = FinnSyll()

To syllabify:

>>> f.syllabify('runoja')
['ru.no.ja']  # internal syllable boundaries are indicated with '.'

To segment compounds:

>>> f.split('sosiaalidemokraattien')
'sosiaali=demokraattien'  # internal word boundaries are indicated with '='

Optional arguments

The syllabifier can be customized along two different parameters: variation and compound splitting.

variation

Instantiating a FinnSyll object with variation=True (default) will allow the syllabifier to return multiple syllabifications if variation is predicted. When variation=True, the syllabifier will return a list. Setting variation to False will cause the syllabifier to return a string containing the first predicted syllabification.

Variation:

>>> f = FinnSyll(variation=True)
>>> f.syllabify('runoja')
['ru.no.ja']
>>> f.syllabify('vapaus')
['va.pa.us', 'va.paus']

No variation:

>>> f = FinnSyll(variation=False)
>>> f.syllabify('runoja')
'ru.no.ja'
>>> f.syllabify('vapaus')
'va.pa.us'

split_compounds

When instantiating a FinnSyll object with split_compounds=True (default), the syllabifier will first attempt to split the input into constituent words before syllabifying it. This forces the syllabifier to insert a syllable boundary in between identified constituent words. The syllabifier will skip this step if split_compounds is set to False.

Compound splitting:

>>> f = FinnSyll(split_compounds=True)
>>> f.syllabify('rahoituserien')  # rahoitus=erien
['ra.hoi.tus.e.ri.en']

No compound splitting:

>>> f = FinnSyll(split_compounds=False)
>>> f.syllabify('rahoituserien')
['ra.hoi.tu.se.ri.en']  # incorrect