Notes

pymorphy is a nice Python library from morphological analysis of Russian. Unfortunately, its documentation has verry little information in English. Here is a guide for non-Russian speakers.

1. Install pymorphy

pip install pymorphy

2. Download dictionaries

wget https://bitbucket.org/kmike/pymorphy/downloads/ru.sqlite-json.zip
unzip ru.sqlite-json.zip -d ru-dict

3. Analyze words

For each of the possible analyses of the word "прочим", we print its lemma, its part-of-speech class and its morphological information:

import pymorphy
morph = pymorphy.get_morph('ru-dict')
for analysis in morph.get_graminfo(u'ПРОЧИМ'):
    print analysis['norm'], analysis['class'], analysis['info']

ПРОЧЕЕ МС ср,ед,тв
ПРОЧИЙ МС мн,дт
ПРОЧИТЬ Г дст,нст,1л,мн
...

Queries have to be unicode uppercase strings.

If you only want to obtain the set of possible lemmas for a word, use:

morph.normalize(u'ПРОЧИМ') == set([u'ПРОЧЕЕ', u'ПРОЧИТЬ', u'ПРОЧИЙ'])

You can also run the reverse operation:

morph.inflect_ru(u'ПРОЧЕЕ', u'ср,ед,тв') == u'ПРОЧИМ'

or obtain the full set of possible realizations of a given lemma with:

for form in morph.decline(u'ГОВОРИТЬ'):
    print form['word'], form['class'], form['info']

ГОВОРИТЬ ИНФИНИТИВ дст
ГОВОРЮ Г дст,нст,1л,ед
ГОВОРИШЬ Г дст,нст,2л,ед
...

4. Using the analyses

The documentation in Russian contains a description of the part-of-speech tags (части речи) and morphological markers (граммемы) used.

Tagset conversion utilities have been written by the authors of pymorphy, and it is also possible to obtained simplified morphological information by using the standard=True argument when calling get_graminfo:

for analysis in morph.get_graminfo(u'ПРОЧИМ', standard=True):
    print analysis['norm'], analysis['class'], analysis['info']

ПРОЧЕЕ - n,ins,sg
ПРОЧИЙ - pl,dat
ПРОЧИТЬ V 1p,pres,pl
...