Merge branch 'improve-tone-style' into develop

Old-Young233 · Jan 23, 2022 · 7e02fbe · 7e02fbe
2 parents a421a83 + 2b70c85
commit 7e02fbe
Show file tree

Hide file tree

Showing 23 changed files with 836 additions and 549 deletions.
diff --git a/pypinyin/constants.py b/pypinyin/constants.py
@@ -34,19 +34,28 @@
     RE_HANS = re.compile(
         r'^(?:['
         r'\u3007'                  # 〇
+        r'\ue815-\ue864'
+        r'\ufa18'
         r'\u3400-\u4dbf'           # CJK扩展A:[3400-4DBF]
         r'\u4e00-\u9fff'           # CJK基本:[4E00-9FFF]
         r'\uf900-\ufaff'           # CJK兼容:[F900-FAFF]
         r'\U00020000-\U0002A6DF'   # CJK扩展B:[20000-2A6DF]
         r'\U0002A703-\U0002B73F'   # CJK扩展C:[2A700-2B73F]
         r'\U0002B740-\U0002B81D'   # CJK扩展D:[2B740-2B81D]
+        r'\U0002B825-\U0002BF6E'
+        r'\U0002C029-\U0002CE93'
+        r'\U0002D016'
         r'\U0002F80A-\U0002FA1F'   # CJK兼容扩展:[2F800-2FA1F]
+        r'\U00030EDD'
+        r'\U00030EDE'
         r'])+$'
     )
 else:
-    RE_HANS = re.compile(  # pragma: no cover
+    RE_HANS = re.compile(
         r'^(?:['
         r'\u3007'                  # 〇
+        r'\ue815-\ue864'
+        r'\ufa18'
         r'\u3400-\u4dbf'           # CJK扩展A:[3400-4DBF]
         r'\u4e00-\u9fff'           # CJK基本:[4E00-9FFF]
         r'\uf900-\ufaff'           # CJK兼容:[F900-FAFF]

diff --git a/pypinyin/contrib/_tone_rule.py b/pypinyin/contrib/_tone_rule.py
@@ -1,45 +1,5 @@
 # -*- coding: utf-8 -*-
 from __future__ import unicode_literals
 
-"""
-标调位置
-
-    有 ɑ 不放过，
-
-　　没 ɑ 找 o、e；
-
-　　ɑ、o、e、i、u、ü
-
-　　标调就按这顺序；
-
-　　i、u 若是连在一起，
-
-　　谁在后面就标谁。
-
-http://www.hwjyw.com/resource/content/2010/06/04/8183.shtml
-https://www.zhihu.com/question/23655297
-https://github.com/mozillazg/python-pinyin/issues/160
-http://www.pinyin.info/rules/where.html
-"""
-
-
-def right_mark_index(pinyin_no_tone):
-    # 有 ɑ 不放过, 没 ɑ 找 o、e
-    for c in ['a', 'o', 'e']:
-        if c in pinyin_no_tone:
-            return pinyin_no_tone.index(c)
-
-    # i、u 若是连在一起，谁在后面就标谁
-    for c in ['iu', 'ui']:
-        if c in pinyin_no_tone:
-            return pinyin_no_tone.index(c) + 1
-
-    # ɑ、o、e、i、u、ü
-    for c in ['i', 'u', 'v', 'ü']:
-        if c in pinyin_no_tone:
-            return pinyin_no_tone.index(c)
-
-    # n, m, ê
-    for c in ['n', 'm', 'ê']:
-        if c in pinyin_no_tone:
-            return pinyin_no_tone.index(c)
+# 向后兼容
+from pypinyin.style._tone_rule import right_mark_index  # noqa