add some comment for easily unstanding

pinyin/hmm/train.py

@@ -33,25 +33,26 @@ def init_start():
 def init_emission():
     """
     初始化发射概率
+    character_pinyin_map: {u'长': {u'chang': 1000, u'zhang': 500}}
     """
     character_pinyin_map = {}
     for phrase, frequency in iter_dict():
-        pinyins = pinyin(phrase, style=NORMAL)
+        pinyins = pinyin(phrase, style=NORMAL) # e.g. [[u'zhang'], [u'da']]
         for character, py in zip(phrase, pinyins):
             character_pinyin_count = len(py)
             if character not in character_pinyin_map:
                 character_pinyin_map[character] = \
-                    {x: frequency/character_pinyin_count for x in py}
+                    {x: frequency / character_pinyin_count for x in py}
             else:
                 pinyin_freq_map = character_pinyin_map[character]
                 for x in py:
                     pinyin_freq_map[x] = pinyin_freq_map.get(x, 0) + \
-                                         frequency/character_pinyin_count
+                                         frequency / character_pinyin_count
 
     for character, pinyin_map in character_pinyin_map.iteritems():
         sum_frequency = sum(pinyin_map.values())
         for py, frequency in pinyin_map.iteritems():
-            Emission.add(character, py, log(frequency/sum_frequency))
+            Emission.add(character, py, log(frequency / sum_frequency))
 
 
 def init_transition():

pinyin/phrase/train.py

@@ -15,6 +15,7 @@
 def int_phrase_pinyin_map():
     """
     初始化字典中短语的拼音
+    result: {u'zhi dao': {u'指导': 1, u'知道':2, u'直到': 3}}
     """
     result = {}
     total = 0

pinyin/utils.py

@@ -8,7 +8,7 @@ def iter_dict():
     """
     遍历dict.txt文件
     """
-    with open(dict_path, 'r', ) as f:
+    with open(dict_path, 'r') as f:
         for line in f:
             phrase, frequency, tag = line.split()
             yield phrase.decode('utf8'), int(frequency)