-
Notifications
You must be signed in to change notification settings - Fork 0
Home
OFITWOL is yet another morphological analyser for Finnish, especially the kind of Finnish that was written in the 19th and 20th centuries. The detailed documentation of OFITWOL is available in https://pytwolc.readthedocs.io/en/latest/index.html and further documentation will appear there.
OFITWOL used the Python 3 package called twol for establishing the morphophonemic representations for lexical and affix entries and for compiling and testing the morphophonemic rules. The documentation which was mentioned above covers also the twol package and the methods used in this project.
The creation of OFITWOL has benefited from several sources, among others Nykysuomen sanakirja (WSOY 1951-1961) and Suomen kielen käänteissanakirja (by Tuomo Tuomi, 1980, SKS) for the tables of inflectional classes and model words, Suomen kielioppi (by Aarni Penttilä 1963) and The Finnish N-grams 1820-2000 of the Newspaper and Periodical Corpus of the National Library of Finland [text]. Kielipankki. Retrieved from http://urn.fi/urn:nbn:fi:lb-2014073038. The OFITWOL files are results of heavy computer processing, lots of manual editing, selection and new material and they bear little resemblance with any existing word lists, lexicons or rule sets. OFITWOL is therefore free software according to the GPL 3 license, see the copyright data of this project.
OFITWOL uses heavily the Helsinki Finite-State Transducer https://github.com/hfst/hfst tools such as hfst-lexc, etc.
Ofitwol is related to OMORFI (by Tommi Pirinen et al.) an open Finnish morphological analyzer based on the HFST tools but is shares with it neither any code, lexicon nor rules.
In my opinion, OFITWOL appears to differ from OMORFI in some respects:
- Whereas OMORFI aims to follow the current norm of the Finnish language by excluding less frequent inflectional forms, OFITWOL aims to be tolerant and accept wider sets of forms that were used some 50 or 100 years ago and are still acceptable even if not common.
- OFITWOL aims to be more flexible so that a researcher may adjust it to accept even further dialectal variation. Special attention is paid to make it easier for scholars to modify OFITWOL so that it would better serve the needs of various applications, including historical and comparative linguistics.
- OMORFI aims to cover the vocabulary of present day Finnish as completely as possible whereas there are presently no plans to make OFITWOL become a practical spelling checker.
- In generation mode, OMORFI is prepared to produce the most preferred inflectional form (which is essential for using it as a part of machine translation into Finnish). OFITWOL has no such capabilities.