Merge branch 'improve-FINALS' into develop

CHANGELOG.rst

@@ -1,6 +1,23 @@
 Changelog
 ---------
 
+0.19.0 (2017-05-05)
++++++++++++++++++++++
+
+* **[New]** 韵母风格下根据 `汉语拼音方案`_ 还原原始的 ``iou`` , ``uei`` , ``uen`` 韵母。
+
+    iou，uei，uen前面加声母的时候，写成iu，ui，un。
+    例如niu(牛)，gui(归)，lun(论)。即：
+
+    * niu 的韵母是 iou
+    * gui 的韵母是 uei
+    * lun 的韵母是 uen
+* **[Fixed]** 修复韵母相关风格下没有正确处理 ``wu`` 的韵母的问题
+  (比如: ``无`` 在 ``FINALS_TONE`` 风格下的结果是 ``uú`` 的问题) 。
+* **[Fixed]** 修复漏了 ǖ -> v1 的转换。
+
+
+
 0.18.2 (2017-04-25)
 +++++++++++++++++++++
 
@@ -390,7 +407,8 @@ __ https://github.com/mozillazg/python-pinyin/issues/8
 .. _#22: https://github.com/mozillazg/python-pinyin/pull/22
 .. _#26: https://github.com/mozillazg/python-pinyin/pull/26
 .. _@MingStar: https://github.com/MingStar
-.. _汉语拼音方案: http://www.edu.cn/20011114/3009777.shtml
+.. _汉语拼音方案: https://zh.wiktionary.org/wiki/%E9%99%84%E5%BD%95:%E6%B1%89%E8%AF%AD%E6%8B%BC%E9%9F%B3%E6%96%B9%E6%A1%88
+.. _汉语拼音方案.pdf: http://www.moe.edu.cn/s78/A19/yxs_left/moe_810/s230/195802/t19580201_186000.html
 .. _汉语拼音 - 维基百科: https://zh.wikipedia.org/wiki/%E6%B1%89%E8%AF%AD%E6%8B%BC%E9%9F%B3#cite_ref-10
 .. _@xulin97: https://github.com/xulin97
 .. _#31: https://github.com/mozillazg/python-pinyin/issues/31

pypinyin/__init__.py

@@ -5,17 +5,24 @@
 
 from __future__ import unicode_literals
 
-from .compat import PY2
-from .constants import (  # noqa
-    STYLE_NORMAL, NORMAL, STYLE_TONE, TONE,
-    STYLE_TONE2, TONE2, STYLE_TONE3, TONE3,
-    STYLE_INITIALS, INITIALS, STYLE_FIRST_LETTER, FIRST_LETTER,
-    STYLE_FINALS, FINALS, STYLE_FINALS_TONE, FINALS_TONE,
-    STYLE_FINALS_TONE2, FINALS_TONE2, STYLE_FINALS_TONE3, FINALS_TONE3,
-    STYLE_BOPOMOFO, BOPOMOFO, STYLE_BOPOMOFO_FIRST, BOPOMOFO_FIRST,
-    STYLE_CYRILLIC, CYRILLIC, STYLE_CYRILLIC_FIRST, CYRILLIC_FIRST
+from pypinyin.compat import PY2
+from pypinyin.constants import (  # noqa
+    STYLE_NORMAL, NORMAL,
+    STYLE_TONE, TONE,
+    STYLE_TONE2, TONE2,
+    STYLE_TONE3, TONE3,
+    STYLE_INITIALS, INITIALS,
+    STYLE_FIRST_LETTER, FIRST_LETTER,
+    STYLE_FINALS, FINALS,
+    STYLE_FINALS_TONE, FINALS_TONE,
+    STYLE_FINALS_TONE2, FINALS_TONE2,
+    STYLE_FINALS_TONE3, FINALS_TONE3,
+    STYLE_BOPOMOFO, BOPOMOFO,
+    STYLE_BOPOMOFO_FIRST, BOPOMOFO_FIRST,
+    STYLE_CYRILLIC, CYRILLIC,
+    STYLE_CYRILLIC_FIRST, CYRILLIC_FIRST
 )
-from .core import (     # noqa
+from pypinyin.core import (     # noqa
     pinyin, lazy_pinyin, slug, load_single_dict, load_phrases_dict
 )
 

pypinyin/__main__.py

@@ -1,8 +1,8 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
-
 import sys
-from .runner import main
+
+from pypinyin.runner import main
 
 if __name__ == '__main__':
     sys.exit(main())

pypinyin/constants.py

@@ -6,14 +6,14 @@
 import os
 import re
 
-from . import phonetic_symbol, pinyin_dict
-from .compat import SUPPORT_UCS4
+from pypinyin import phonetic_symbol, pinyin_dict
+from pypinyin.compat import SUPPORT_UCS4
 
 # 词语拼音库
 if os.environ.get('PYPINYIN_NO_PHRASES'):
     PHRASES_DICT = {}
 else:
-    from . import phrases_dict
+    from pypinyin import phrases_dict
     PHRASES_DICT = phrases_dict.phrases_dict.copy()
 
 # 单字拼音库
@@ -93,18 +93,10 @@
 BOPOMOFO = STYLE_BOPOMOFO = PINYIN_STYLE['BOPOMOFO']
 # 注音符号首字母
 BOPOMOFO_FIRST = STYLE_BOPOMOFO_FIRST = PINYIN_STYLE['BOPOMOFO_FIRST']
-
+# 俄语
 CYRILLIC = STYLE_CYRILLIC = PINYIN_STYLE['CYRILLIC']
-
 CYRILLIC_FIRST = STYLE_CYRILLIC_FIRST = PINYIN_STYLE['CYRILLIC_FIRST']
-
-U_FINALS_EXCEPTIONS_MAP = {
-    u'ū': u'ǖ',
-    u'ú': u'ǘ',
-    u'ǔ': u'ǚ',
-    u'ù': u'ǜ',
-}
-
+# 注音转换表
 BOPOMOFO_REPLACE = (
     (re.compile('^m(\d)$'), 'mu\\1'),  # 呣
     (re.compile('^n(\d)$'), 'N\\1'),  # 嗯
@@ -139,7 +131,7 @@
     'bpmfdtnlgkhjqxZCSrzcsiuvaoeEAIOUMNKGR2340',
     'ㄅㄆㄇㄈㄉㄊㄋㄌㄍㄎㄏㄐㄑㄒㄓㄔㄕㄖㄗㄘㄙㄧㄨㄩㄚㄛㄜㄝㄞㄟㄠㄡㄢㄣㄤㄥㄦˊˇˋ˙'
 ))
-
+# 俄语转换表
 CYRILLIC_REPLACE = (
     (re.compile('ong'), 'ung'),
     (re.compile('([zcs])i'), '\\1U'),

pypinyin/core.py

@@ -9,18 +9,19 @@
 import re
 import warnings
 
-from .compat import text_type, callable_check
-from .constants import (
+from pypinyin.compat import text_type, callable_check
+from pypinyin.constants import (
     PHRASES_DICT, PINYIN_DICT, _INITIALS, PHONETIC_SYMBOL, RE_PHONETIC_SYMBOL,
-    RE_TONE2, RE_TONE3, RE_HANS, U_FINALS_EXCEPTIONS_MAP,
+    RE_TONE2, RE_TONE3, RE_HANS,
     BOPOMOFO_REPLACE, BOPOMOFO_TABLE,
     CYRILLIC_REPLACE, CYRILLIC_TABLE,
     NORMAL, TONE, TONE2, TONE3, INITIALS, FIRST_LETTER,
     FINALS, FINALS_TONE, FINALS_TONE2, FINALS_TONE3,
     BOPOMOFO, BOPOMOFO_FIRST,
     CYRILLIC, CYRILLIC_FIRST
 )
-from .utils import simple_seg, _replace_tone2_style_dict_to_default
+from pypinyin.standard import convert_finals
+from pypinyin.utils import simple_seg, _replace_tone2_style_dict_to_default
 
 
 def seg(hans):
@@ -106,34 +107,14 @@ def final(pinyin):
     :return: 韵母
     :rtype: unicode
     """
-    initial_ = initial(pinyin) or None
+    initial_ = initial(pinyin) or ''
+    # 没有声母，整个都是韵母
     if not initial_:
-        return no_initial_final(pinyin)
-    # 特例 j/q/x
-    m = re.match(r'^(j|q|x)(ū|ú|ǔ|ù)$', pinyin)
-    if m:
-        return (U_FINALS_EXCEPTIONS_MAP[m.group(2)])
-    pinyin = re.sub(r'^(j|q|x)u(\d?)$', r'\1v\2', pinyin)
+        return pinyin
+    # 按声母分割，剩下的就是韵母
     return ''.join(pinyin.split(initial_, 1))
 
 
-def no_initial_final(pinyin):
-    # 特例 y/w
-    if pinyin.startswith('y'):
-        if pinyin.startswith('yu'):
-            pinyin = 'v' + pinyin[2:]
-        elif pinyin.startswith('yi'):
-            pinyin = pinyin[1:]
-        else:
-            pinyin = 'i' + pinyin[1:]
-    elif pinyin.startswith('w'):
-        if pinyin.startswith('wu'):
-            pinyin = pinyin[1:]
-        else:
-            pinyin = 'u' + pinyin[1:]
-    return pinyin
-
-
 def to_fixed(pinyin, style):
     """根据拼音风格格式化带声调的拼音.
 
@@ -145,6 +126,12 @@ def to_fixed(pinyin, style):
     # 声母
     if style == INITIALS:
         return initial(pinyin)
+    if style == TONE:
+        return pinyin
+
+    # 根据标准还原正确的韵母
+    if style in [FINALS, FINALS_TONE, FINALS_TONE2, FINALS_TONE3]:
+        pinyin = convert_finals(pinyin)
 
     def _replace(m):
         symbol = m.group(0)  # 带声调的字符

pypinyin/phonetic_symbol.py

@@ -27,6 +27,7 @@
     "ù": "u4",
     # üe
     "ü": "v",
+    "ǖ": "v1",
     "ǘ": "v2",
     "ǚ": "v3",
     "ǜ": "v4",

pypinyin/runner.py

@@ -5,13 +5,13 @@
 from argparse import ArgumentParser
 import sys
 
-from . import (                                    # noqa
+from pypinyin import (                                    # noqa
     __title__, __version__, pinyin, slug,
     NORMAL, TONE, TONE2, TONE3, INITIALS, FIRST_LETTER,
     FINALS, FINALS_TONE, FINALS_TONE2, FINALS_TONE3,
     BOPOMOFO, BOPOMOFO_FIRST, CYRILLIC, CYRILLIC_FIRST
 )
-from .compat import PY2
+from pypinyin.compat import PY2
 
 
 class NullWriter(object):

pypinyin/standard.py

@@ -0,0 +1,152 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""
+处理汉语拼音方案中的一些特殊情况
+
+
+汉语拼音方案:
+
+* https://zh.wiktionary.org/wiki/%E9%99%84%E5%BD%95:%E6%B1%89%E8%AF%AD%E6%8B%BC%E9%9F%B3%E6%96%B9%E6%A1%88
+* http://www.moe.edu.cn/s78/A19/yxs_left/moe_810/s230/195802/t19580201_186000.html
+"""  # noqa
+from __future__ import unicode_literals
+
+import re
+
+# u -> ü
+UV_MAP = {
+    'u': 'ü',
+    'ū': 'ǖ',
+    'ú': 'ǘ',
+    'ǔ': 'ǚ',
+    'ù': 'ǜ',
+}
+U_TONES = set(UV_MAP.keys())
+# ü行的韵跟声母j，q，x拼的时候，写成ju(居)，qu(区)，xu(虚)
+UV_RE = re.compile(r'^(j|q|x)({tones})$'.format(tones='|'.join(UV_MAP.keys())))
+I_TONES = set(['i', 'ī', 'í', 'ǐ', 'ì'])
+
+# iu -> iou
+IU_MAP = {
+    'iu': 'iou',
+    'iū': 'ioū',
+    'iú': 'ioú',
+    'iǔ': 'ioǔ',
+    'iù': 'ioù',
+}
+IU_TONES = set(IU_MAP.keys())
+IU_RE = re.compile(r'^([a-z]+)({tones})$'.format(tones='|'.join(IU_TONES)))
+
+# ui -> uei
+UI_MAP = {
+    'ui': 'uei',
+    'uī': 'ueī',
+    'uí': 'ueí',
+    'uǐ': 'ueǐ',
+    'uì': 'ueì',
+}
+UI_TONES = set(UI_MAP.keys())
+UI_RE = re.compile(r'([a-z]+)({tones})$'.format(tones='|'.join(UI_TONES)))
+
+# un -> uen
+UN_MAP = {
+    'un': 'uen',
+    'ūn': 'ūen',
+    'ún': 'úen',
+    'ǔn': 'ǔen',
+    'ùn': 'ùen',
+}
+UN_TONES = set(UN_MAP.keys())
+UN_RE = re.compile(r'([a-z]+)({tones})$'.format(tones='|'.join(UN_TONES)))
+
+
+def convert_zero_consonant(pinyin):
+    """零声母转换，还原原始的韵母
+
+    i行的韵母，前面没有声母的时候，写成yi(衣)，ya(呀)，ye(耶)，yao(腰)，
+    you(忧)，yan(烟)，yin(因)，yang(央)，ying(英)，yong(雍)。
+
+    u行的韵母，前面没有声母的时候，写成wu(乌)，wa(蛙)，wo(窝)，wai(歪)，
+    wei(威)，wan(弯)，wen(温)，wang(汪)，weng(翁)。
+
+    ü行的韵母，前面没有声母的时候，写成yu(迂)，yue(约)，yuan(冤)，
+    yun(晕)；ü上两点省略。
+    """
+    # y: yu -> v, yi -> i, y -> i
+    if pinyin.startswith('y'):
+        # 去除 y 后的拼音
+        no_y_py = pinyin[1:]
+        first_char = no_y_py[0] if len(no_y_py) > 0 else None
+
+        # yu -> ü: yue -> üe
+        if first_char in U_TONES:
+            pinyin = UV_MAP[first_char] + pinyin[2:]
+        # yi -> i: yi -> i
+        elif first_char in I_TONES:
+            pinyin = no_y_py
+        # y -> i: ya -> ia
+        else:
+            pinyin = 'i' + no_y_py
+        return pinyin
+
+    # w: wu -> u, w -> u
+    if pinyin.startswith('w'):
+        # 去除 w 后的拼音
+        no_w_py = pinyin[1:]
+        first_char = no_w_py[0] if len(no_w_py) > 0 else None
+
+        # wu -> u: wu -> u
+        if first_char in U_TONES:
+            pinyin = pinyin[1:]
+        # w -> u: wa -> ua
+        else:
+            pinyin = 'u' + pinyin[1:]
+        return pinyin
+
+    return pinyin
+
+
+def convert_uv(pinyin):
+    """ü 转换，还原原始的韵母
+
+    ü行的韵跟声母j，q，x拼的时候，写成ju(居)，qu(区)，xu(虚)，
+    ü上两点也省略；但是跟声母n，l拼的时候，仍然写成nü(女)，lü(吕)。
+    """
+    return UV_RE.sub(lambda m: m.group(1) + UV_MAP[m.group(2)], pinyin)
+
+
+def convert_iou(pinyin):
+    """iou 转换，还原原始的韵母
+
+    iou，uei，uen前面加声母的时候，写成iu，ui，un。
+    例如niu(牛)，gui(归)，lun(论)。
+    """
+    return IU_RE.sub(lambda m: m.group(1) + IU_MAP[m.group(2)], pinyin)
+
+
+def convert_uei(pinyin):
+    """uei 转换，还原原始的韵母
+
+    iou，uei，uen前面加声母的时候，写成iu，ui，un。
+    例如niu(牛)，gui(归)，lun(论)。
+    """
+    return UI_RE.sub(lambda m: m.group(1) + UI_MAP[m.group(2)], pinyin)
+
+
+def convert_uen(pinyin):
+    """uen 转换，还原原始的韵母
+
+    iou，uei，uen前面加声母的时候，写成iu，ui，un。
+    例如niu(牛)，gui(归)，lun(论)。
+    """
+    return UN_RE.sub(lambda m: m.group(1) + UN_MAP[m.group(2)], pinyin)
+
+
+def convert_finals(pinyin):
+    """还原原始的韵母"""
+    pinyin = convert_zero_consonant(pinyin)
+    pinyin = convert_uv(pinyin)
+    pinyin = convert_iou(pinyin)
+    pinyin = convert_uei(pinyin)
+    pinyin = convert_uen(pinyin)
+    return pinyin