Index: incubator/apertium-bel-rus/dev/process_nouns.py
===================================================================
--- incubator/apertium-bel-rus/dev/process_nouns.py	(nonexistent)
+++ incubator/apertium-bel-rus/dev/process_nouns.py	(revision 66399)
@@ -0,0 +1,102 @@
+# usage: python process_nouns.py N.xml meanings.txt 1> success.txt 2> errors.txt
+# where:
+#   N.xml – file with nouns (bnkorpus grammar DB)
+#   meanings.txt – file for dump words with multiple paradigms
+# it prints speling-file to stdout and errors to stderr
+
+import sys, xml
+import xml.etree.ElementTree as ET
+from BelarusianTagsAnalyzer.BelarusianTagsAnalyzer import Analyzer
+
+class Form(object):
+    lemma = None
+    form = None
+    pr1 = None
+    pr2 = None
+    def __init__(self, lemma, form, parameter1, parameter2):
+        self.lemma = lemma
+        self.form = form
+        self.pr1 = parameter1
+        self.pr2 = parameter2
+    def __str__(self):
+        return '; '.join([self.lemma, self.form, self.pr1, self.pr2])
+
+
+class Paradigm:
+    forms = []
+    def __init__(self, forms):
+        self.forms = forms
+    def __str__(self):
+        return "\n".join([str(form) for form in self.forms])
+    def add_suffix(self, suffix):
+        self.forms = [Form(f.lemma + suffix, f.form, f.pr1, f.pr2) for f in self.forms]
+
+
+gender = {'мужчынскі': 'm', 'жаночы': 'f', 'ніякі': 'nt'}
+animacy = {'адушаўлёны': 'aa', 'неадушаўлёны': 'nn'}
+case = {'назоўны': 'nom', 'родны': 'gen', 'давальны': 'dat', 'вінавальны': 'acc', 'творны': 'ins', 'месны': 'loc'}
+num = {'адзіночны': 'sg', 'множны': 'pl'}
+nums = ['¹', '²', '³', '⁴', '⁵']
+        
+f = ET.parse(sys.argv[1])
+file_meanings = open(sys.argv[2], 'w')
+
+paradigms = []
+lem_cnt = {}
+path = './/Paradigm'
+analyzer = Analyzer()
+for paradigm in f.findall(path):
+    lem = paradigm.attrib['Lemma'].replace('´', '')
+    tag = paradigm.attrib['Tag']
+    meaning = paradigm.get('meaning', '')
+
+    try:
+        p = analyzer.describe(tag)
+    except:
+        continue
+
+    lem_cnt[lem] = [(p[2] + "." + p[5], meaning)] if lem not in lem_cnt else lem_cnt[lem] + [(p[2] + "." + p[5], meaning)]
+
+    forms = []
+    for form in paradigm.findall('.//Form'): 
+        full_tag = tag + form.attrib['Tag']
+        forma = ''
+        for i in form.itertext(): 
+            forma = forma + i
+        forma = forma.replace('´', '')
+        try:
+            p = analyzer.describe(full_tag)
+        except:
+            print("{} | {} | {}".format(lem, forma, full_tag), file=sys.stderr)
+        else:
+            canim = animacy[p[2]] if p[2] in animacy else None
+            cgen = gender[p[5]] if p[5] in gender else None
+            ccase = case[p[7]] if p[7] in case else None
+            cnum = num[p[8]] if p[8] in num else None
+            if canim is None or cgen is None or ccase is None or cnum is None:
+                print("{} | {} | {}".format(lem, forma, full_tag), file=sys.stderr)
+            else:
+                forms += [Form(lem, forma, '.'.join([cnum, ccase]), '.'.join(['n', cgen, canim]))]
+    if len(forms) > 0:
+        paradigms += [Paradigm(forms)]
+
+lem_us = dict(lem_cnt)
+for x in lem_us:
+    lem_us[x] = 0
+
+for p in paradigms:
+    lem = p.forms[0].lemma
+    if len(lem_cnt[lem]) > 1:
+        mng = set([x[0] for x in lem_cnt[lem]])
+        if len(mng) != 1:
+            continue
+        p.add_suffix(nums[lem_us[lem]])
+        file_meanings.write("{}; {}\n".format(p.forms[0].lemma, lem_cnt[lem][lem_us[lem]]))
+        lem_us[lem] += 1
+
+for p in paradigms:
+    print(p)
+
+file_meanings.close()
+
+