Small PEP 8 changes

morfessor.py

@@ -5,6 +5,11 @@
 Smallest pieces are "atoms" (e.g. characters or words)
 Lexicon contains "items" (e.g. morphs or phrases)
 """
+
+__all__ = ['InputFormatError','batch_train','online_train',
+           'corpus_segmentation_dict', 'Lexicon','BaselineModel', 'Corpus',
+           'Annotations']
+
 __version__ = '2.0.0pre1'
 __author__ = 'Sami Virpioja, Peter Smit'
 __author_email__ = "[email protected]"
@@ -191,7 +196,7 @@ def __init__(self, forcesplit_list = [], corpusweight = 1.0,
             self.counter = {}       # Counter for random skipping
         self.corpuscostweight = corpusweight
         self.forcesplit_list = forcesplit_list
-        if annotations != None:
+        if annotations is not None:
             self.set_annotations(annotations, supervisedcorpusweight)
         else:
             self.supervised = False
@@ -208,7 +213,7 @@ def set_annotations(self, annotations, supervisedcorpusweight):
         self.supervisedtokens = 0
         self.supervisedlogtokensum = 0.0
         self.supervisedcorpuscost = 0.0
-        if supervisedcorpusweight == None:
+        if supervisedcorpusweight is None:
             self.supervisedcorpusweight = 1.0
             self.sweightbalance = True
         else:
@@ -225,7 +230,7 @@ def load_segmentations(self, segfile):
         <count> <item1> + <item2> + ... + <itemN>
 
         """
-        if segfile[-3:] == '.gz':
+        if segfile.endswith('.gz'):
             fobj = gzip.open(segfile, 'r')
         else:
             fobj = open(segfile, 'r')
@@ -253,7 +258,7 @@ def save_segmentations(self, segfile):
         <count> <item1> + <item2> + ... + <itemN>
 
         """
-        if segfile[-3:] == '.gz':
+        if segfile.endswith('.gz'):
             fobj = gzip.open(segfile, 'w')
         else:
             fobj = open(segfile, 'w')
@@ -458,7 +463,7 @@ def best_analysis(self, choices):
                         math.log(self.analyses[m][1])
                 else:
                     cost -= self.penaltylogprob # penaltylogprob is negative
-            if bestcost == None or cost < bestcost:
+            if bestcost is None or cost < bestcost:
                 bestcost = cost
                 bestanalysis = analysis
         return bestanalysis, bestcost
@@ -614,7 +619,7 @@ def get_viterbi_segments(self, compound, allow_new_items = True):
             bestpath = None
             bestcost = None
             for pt in range(0, t):
-                if grid[pt][0] == None:
+                if grid[pt][0] is None:
                     continue
                 cost = grid[pt][0]
                 item = compound[pt:t]
@@ -631,14 +636,14 @@ def get_viterbi_segments(self, compound, allow_new_items = True):
                     cost += badlikelihood
                 else:
                     continue
-                if bestcost == None or cost < bestcost:
+                if bestcost is None or cost < bestcost:
                     bestcost = cost
                     bestpath = pt
             grid.append((bestcost, bestpath))
         items = []
         path = grid[-1][1]
         lt = clen + 1
-        while path != None:
+        while path is not None:
             t = path
             items.append(compound[t:lt])
             path = grid[t][1]
@@ -682,7 +687,7 @@ def get_compound_str(self, i):
 
     def get_compound_atoms(self, i):
         """Return the atom representation of the compound at index i."""
-        if self.atom_sep == None:
+        if self.atom_sep is None:
             return self.compounds[i] # string
         else:
             return tuple(re.split(self.atom_sep, self.compounds[i])) # tuple
@@ -715,7 +720,7 @@ def get_max_compound_len(self):
 
     def get_compound_len(self, c):
         """Return the number of atoms in the compound."""
-        if self.atom_sep == None:
+        if self.atom_sep is None:
             return len(c)
         else:
             return len(re.split(self.atom_sep, c))
@@ -740,7 +745,7 @@ def load(self, datafile, compound_sep = ' *', comment_re = "^#"):
         for line in fobj:
             if re.search(comment_re, line):
                 continue
-            if compound_sep == None or compound_sep == '':
+            if compound_sep is None or compound_sep == '':
                 # Line is one compound
                 compounds = [line.rstrip()]
             else:
@@ -789,7 +794,7 @@ def generator(self, datafiles, compound_sep = ' *', comment_re = "^#"):
             for line in fobj:
                 if re.search(comment_re, line):
                     continue
-                if compound_sep == None or compound_sep == '':
+                if compound_sep is None or compound_sep == '':
                     # Line is one compound
                     compounds = [line.rstrip()]
                 else:
@@ -978,14 +983,14 @@ def online_train(model, corpusiter, epochinterval = 10000, dampfunc = None):
 
     """
     model.epoch_update(0)
-    if dampfunc != None:
+    if dampfunc is not None:
         counts = {}
     _vprint("Starting online training\n", 1)
     i = 0
     epochs = 0
     dotfreq = int(math.ceil(epochinterval / 70.0))
     for w in corpusiter:
-        if dampfunc != None:
+        if dampfunc is not None:
             if not counts.has_key(w):
                 c = 0
                 counts[w] = 1
@@ -1175,30 +1180,30 @@ def main(argv):
     global _verboselevel
     _verboselevel = args.verbose
 
-    if args.loadfile == None and args.loadsegfile == None and \
+    if args.loadfile is None and args.loadsegfile is None and \
             len(args.trainfiles) == 0:
         parser.error("either model file or training data should be defined")
 
-    if args.randseed != None:
+    if args.randseed is not None:
         random.seed(args.randseed)
 
     # Load annotated data if specified
-    if args.annofile != None:
+    if args.annofile is not None:
         annotations = Annotations()
         annotations.load(args.annofile)
     else:
         annotations = None
 
     # Load exisiting model or create a new one
-    if args.loadfile != None:
+    if args.loadfile is not None:
         _vprint("Loading model from '%s'..." % args.loadfile, 1)
         with open(args.loadfile, 'rb') as fobj:
             model = pickle.load(fobj)
         _vprint(" Done.\n", 1)
-        if annotations != None:
+        if annotations is not None:
             # Add annotated data to model
             model.set_annotations(annotations, args.scorpusweight)
-    elif args.loadsegfile != None:
+    elif args.loadsegfile is not None:
         _vprint("Loading model from '%s'..." % args.loadsegfile, 1)
         model = BaselineModel(forcesplit_list = args.forcesplit,
                               corpusweight = args.corpusweight,
@@ -1240,7 +1245,7 @@ def main(argv):
                     data.load(f, args.cseparator)
                 _vprint(" Done.\n", 1)
             model.batch_init(data, args.freqthreshold, dampfunc)
-            if args.splitprob != None:
+            if args.splitprob is not None:
                 model.random_split_init(data, args.splitprob)
             e, c = batch_train(model, data, freqthreshold = args.freqthreshold)
         elif args.trainmode == 'online':
@@ -1259,20 +1264,20 @@ def main(argv):
                      (e, c, te-ts), 1)
 
     # Save model
-    if args.savefile != None:
+    if args.savefile is not None:
         _vprint("Saving model to '%s'..." % args.savefile, 1)
         with open(args.savefile, 'wb') as fobj:
             pickle.dump(model, fobj, pickle.HIGHEST_PROTOCOL)
         _vprint(" Done.\n", 1)
 
-    if args.savesegfile != None:
+    if args.savesegfile is not None:
         _vprint("Saving model segmentations to '%s'..." %
                      args.savesegfile, 1)
         model.save_segmentations(args.savesegfile)
         _vprint(" Done.\n", 1)
 
     # Output lexicon
-    if args.lexfile != None:
+    if args.lexfile is not None:
         if args.lexfile == '-':
             fobj = sys.stdout
         elif args.lexfile[-3:] == '.gz':