Merge 717ce93 into 4072b88

CHANGELOG.rst

@@ -4,6 +4,7 @@ Changelog
 `0.30.1`_ (未发布)
 ++++++++++++++++++++++++
 
+* **[Improved]** 对用户传入的已进行分词处理的数据进行二次分词以便提高准确性。
 * **[Improved]** 使用 `pinyin-data`_ v0.5.1 的拼音数据。
 
 

pypinyin/contrib/mmseg.py

@@ -77,6 +77,7 @@ def __contains__(self, key):
 #:    >>> text = '你好，我是中国人，我爱我的祖国'
 #:    >>> seg.cut(text)
 #:    <generator object Seg.cut at 0x10b2df2b0>
+#:
 #:    >>> list(seg.cut(text))
 #:    ['你好', '，', '我', '是', '中国人', '，', '我', '爱',
 #:     '我的', '祖', '国']

pypinyin/core.py

@@ -13,7 +13,7 @@
 )
 from pypinyin.contrib import mmseg
 from pypinyin.utils import simple_seg, _replace_tone2_style_dict_to_default
-from pypinyin.style import auto_discover, convert
+from pypinyin.style import auto_discover, convert as convert_style
 
 auto_discover()
 
@@ -77,15 +77,7 @@ def to_fixed(pinyin, style, strict=True):
     :return: 根据拼音风格格式化后的拼音字符串
     :rtype: unicode
     """
-    return convert(pinyin, style=style, strict=strict, default=pinyin)
-
-    # # 韵母
-    # elif style in [FINALS, FINALS_TONE, FINALS_TONE2, FINALS_TONE3]:
-    #     # 不处理鼻音: 'ḿ', 'ń', 'ň', 'ǹ'
-    #     if pinyin and pinyin[0] not in [
-    #         '\u1e3f', '\u0144', '\u0148', '\u01f9'
-    #     ]:
-    #         py = final(py, strict=strict)
+    return convert_style(pinyin, style=style, strict=strict, default=pinyin)
 
 
 def _handle_nopinyin_char(chars, errors='default'):
@@ -217,6 +209,8 @@ def foobar(char):
     :return: 拼音列表
     :rtype: list
 
+    :raise AssertionError: 当传入的字符串不是 unicode 字符时会抛出这个异常
+
     Usage::
 
       >>> from pypinyin import pinyin, Style
@@ -234,9 +228,11 @@ def foobar(char):
     """
     # 对字符串进行分词处理
     if isinstance(hans, text_type):
-        hans = seg(hans)
+        han_list = seg(hans)
+    else:
+        han_list = chain(*(seg(x) for x in hans))
     pys = []
-    for words in hans:
+    for words in han_list:
         pys.extend(_pinyin(words, style, heteronym, errors, strict=strict))
     return pys
 
@@ -256,6 +252,8 @@ def slug(hans, style=Style.NORMAL, heteronym=False, separator='-',
     :param strict: 是否严格遵照《汉语拼音方案》来处理声母和韵母，详见 :ref:`strict`
     :return: slug 字符串.
 
+    :raise AssertionError: 当传入的字符串不是 unicode 字符时会抛出这个异常
+
     ::
 
       >>> import pypinyin
@@ -290,6 +288,8 @@ def lazy_pinyin(hans, style=Style.NORMAL, errors='default', strict=True):
     :return: 拼音列表(e.g. ``['zhong', 'guo', 'ren']``)
     :rtype: list
 
+    :raise AssertionError: 当传入的字符串不是 unicode 字符时会抛出这个异常
+
     Usage::
 
       >>> from pypinyin import lazy_pinyin, Style

pypinyin/utils.py

@@ -44,7 +44,7 @@ def _seg(chars):
 
 
 def simple_seg(hans):
-    """将传入的字符串按是否有拼音来分割"""
+    """将传入的字符串按是否是汉字来分割"""
     assert not isinstance(hans, bytes_type), \
         'must be unicode string or [unicode, ...] list'
 

pytest.ini

@@ -2,5 +2,5 @@
 python_files = test_*.py
 python_classes = Test
 python_functions = test
-addopts = -slv --tb=short --pdb
+addopts = -slv --cov-report term-missing --tb=short --pdb
 norecursedirs = .git __pycache__

tests/test_pinyin.py

@@ -358,6 +358,11 @@ def test_simple_seg():
     ['侵略', {'style': BOPOMOFO}, ['ㄑㄧㄣ', 'ㄌㄩㄝˋ']],
     ['侵略', {'style': CYRILLIC}, ['цинь1', 'люэ4']],
     ['〇', {'style': TONE}, ['líng']],
+    # 二次分词
+    ['你要重新考虑', {'style': TONE}, [
+        'nǐ', 'yào', 'zhòng', 'xīn', 'kǎo', 'lǜ']],
+    [['你要', '重新考虑'], {'style': TONE}, [
+        'nǐ', 'yào', 'chóng', 'xīn', 'kǎo', 'lǜ']],
 ]
 
 

tox.ini

@@ -10,7 +10,7 @@ envlist = py26, py27, py33, py34, py35, py36, pypy, py3_env
 deps =
   pytest
   pytest-cov
-  pytest-random
+  pytest-random-order
 
 [testenv]
 deps = {[base]deps}
@@ -26,7 +26,7 @@ commands = py.test --cov pypinyin tests/_test_env.py
 deps =
   pytest<3.3.0
   pytest-cov
-  pytest-random
+  pytest-random-order
 
 [testenv:py26]
 deps = {[oldpy]deps}