[libvoikko] Documentation for adding morphological analyzer to libvoikko

Harri Pitkänen hatapitk at iki.fi
Sat Nov 21 19:08:48 EET 2009


I have written a short document that describes the format in which the current
MalagaAnalyzer in libvoikko returns the results of morphological analysis:

http://voikko.svn.sourceforge.net/viewvc/voikko/trunk/libvoikko/doc/morphological-analysis.txt

It will not be necessary to implement a morphological analyzer at all to
provide spell checking capability through libvoikko. But right now it is still
needed for many things and even in the future it will be needed at least for
voikko_analyze_word_* public API functions. As can be seen from the
document, the language independent part is rather short (only two attributes)
and I expect that it is possible to provide those for most languages.
Current HfstAnalyzer does not correctly provide either of them.

The part of the document describing attributes for Finnish language support is
in Finnish. All of these attributes are used only in applications that are in
a very early stage of development, such as "WebVoikko 2.0", an experimental
web application for text analysis. They may change significantly and I would
not bother implementing them yet even in HfstAnalyzer for Finnish language.

Harri



More information about the Libvoikko mailing list