Merge pull request #796 from nschneid/patch-11

WordNet: lemma name lookup bugfix (#454) was clobbered

nltk/corpus/reader/wordnet.py

@@ -1177,7 +1177,9 @@ class WordNetCorpusReader(CorpusReader):
 
     def lemma(self, name, lang='en'):
         '''Return lemma object that matches the name'''
-        synset_name, lemma_name = name.rsplit('.', 1)
+        # cannot simply split on first '.', e.g.: '.45_caliber.a.01..45_caliber'
+        separator = SENSENUM_RE.search(name).start()
+        synset_name, lemma_name = name[:separator+3], name[separator+4:]
         synset = self.synset(synset_name)
         for lemma in synset.lemmas(lang):
             if lemma._name == lemma_name: