cleaning up crap left off

git-svn-id: https://scm.gforge.inria.fr/authscm/cfourrie/svn/lingwb/metagger/trunk@2690 dc05b511-7f1d-0410-9f1c-d6f32a2df9e4

src/metagger/Makefile.am

@@ -18,7 +18,15 @@ CLEANFILES = 		__init__.py		\
 			megam_classifier.py	\
 			result_sink.py 		\
 			utils.py		\
-			corpus_reader.py
+			corpus_reader.py	\
+			__init__.pyc		\
+			pos_tagger.pyc		\
+			instance.pyc		\
+			mytoken.pyc		\
+			megam_classifier.pyc	\
+			result_sink.pyc 	\
+			utils.pyc		\
+			corpus_reader.pyc
 
 BUILT_SOURCES = 	__init__.py		\
 			pos_tagger.py		\

src/metagger/corpus_reader.py

-""" Corpus reader """
-
-import sys
-import codecs
-import re
-
-WD_TAG_RE = re.compile(r'^(.+?)/(.+)$')
-
-
-class CorpusReader:
-    pass
-
-
-class BrownReader(CorpusReader):
-
-    """
-    Data reader for corpus in the Brown format:
-
-    Le/DET prix/NC de/P vente/NC du/P+D journal/NC n'/ADV a/V pas/ADV <E9>t<E9>/VPP divulgu<E9>/VPP ./PONCT 
-    Le/DET cabinet/NC Arthur/NPP Andersen/NPP ,/PONCT administrateur/NC judiciaire/ADJ du/P+D titre/NC depuis/P l'/DET effondrement/NC de/P l'/DET empire/NC Maxwell/NPP en/P d<E9>cembre/NC 1991/NC ,/PO
-    NCT a/V indiqu<E9>/VPP que/CS les/DET nouveaux/ADJ propri<E9>taires/NC allaient/V poursuivre/VINF la/DET publication/NC ./PONCT 
-
-    """
-
-    def __init__(self,infile, encoding='latin-1'):
-        self.stream = codecs.open(infile, 'r', encoding)
-        return
-
-    def __iter__(self):
-        return self
-
-    def next(self):
-        line = self.stream.readline().strip()
-        if (line == ''):
-            self.stream.seek(0)
-            raise StopIteration
-        token_list = []
-        for item in line.split():
-            wd,tag = WD_TAG_RE.match(item).groups()
-            token_list.append( (wd,tag) )
-        return token_list
-
-
-
-
-