Vabamorf - open source morphology tagger for Estonian

The software package consists of: 1. The morphoanalyzer shall determine the lemma, morphological structure, word type and morphological categories corresponding to the lemma, morphological structure, based on the form of the lemma. If the words are analyzed in many ways, several possible analyses are issued; The analyzer is also able to analyse the words corresponding to which are missing from the basic dictionary of morphological software. By calling out the analytical functions, it is possible to specify whether to analyse the forms of words missing from the basic dictionary (assume the output of the analysis); The analyzer can determine the pronunciation characteristics in the words analyzed: oppressive syllables, letters corresponding to palatalized pronunciations, and third-external syllables. In the case of foreign names, the determination of pronunciation characteristics is not required; 2. The morphosynthesizer synthesizes the morphological structure and pronunciation characteristics of the word change form (s), including the morphological structure and pronunciation characteristics of the change form, on the basis of a predetermined lemma or predetermined change forms. Morphosynthesizer can also synthesize transformation forms from words that are missing from the basic dictionary of morphology software. When summoning synthesis functions, it is possible to specify whether to synthesize the forms of words missing from the basic dictionary (assume the output of synthesis); Synthesis function can synthesize the whole paradigm of other forms at the same time. 3. The spelling module classifies words as orthodox and non-proper word forms. In addition, this module is able to find all the correct language word forms similar to a given word, using a predetermined similarity metric; 4. Estonian language speller suitable for Libre Office and OpenOffice.org 3.0.1 or later. Used on MS Windows, Linux, Macintosh platforms. 5. The Morphological Analysis Connector automatically finds the most likely analysis for multiple analysis words using sentence and/or document context. Preferably, the connector can assign context-dependent probabilities to all the analyses provided for the connection. The connector code shall be separated from the other morphological analysis code. 6. The dictionary. The format of the original text of the dictionary required for the work of the morphological software is human-readable and thoroughly documented. Complementing the basic dictionary of the morphological software by the user can ensure that the dictionary is in conformity with the format (e.g. a dictionary validator can be integrated, which checks whether the dictionary corresponds to the established format and indicates definite and possible errors). The dictionary uses the UTF-8 encoding.