blob: c6252540707094c3736b954b0c8c30528d0e0561 [file] [log] [blame]
Georg Brandl116aa622007-08-15 14:28:22 +00001:mod:`textwrap` --- Text wrapping and filling
2=============================================
3
4.. module:: textwrap
5 :synopsis: Text wrapping and filling
6.. moduleauthor:: Greg Ward <gward@python.net>
7.. sectionauthor:: Greg Ward <gward@python.net>
8
Raymond Hettinger10480942011-01-10 03:26:08 +00009**Source code:** :source:`Lib/textwrap.py`
Georg Brandl116aa622007-08-15 14:28:22 +000010
Raymond Hettinger4f707fd2011-01-10 19:54:11 +000011--------------
12
Georg Brandl116aa622007-08-15 14:28:22 +000013The :mod:`textwrap` module provides two convenience functions, :func:`wrap` and
14:func:`fill`, as well as :class:`TextWrapper`, the class that does all the work,
Nick Coghlan4fae8cd2012-06-11 23:07:51 +100015and two utility functions, :func:`dedent` and :func:`indent`. If you're just wrapping or filling one
Georg Brandl116aa622007-08-15 14:28:22 +000016or two text strings, the convenience functions should be good enough;
17otherwise, you should use an instance of :class:`TextWrapper` for efficiency.
18
Georg Brandl7f01a132009-09-16 15:58:14 +000019.. function:: wrap(text, width=70, **kwargs)
Georg Brandl116aa622007-08-15 14:28:22 +000020
Georg Brandl7f01a132009-09-16 15:58:14 +000021 Wraps the single paragraph in *text* (a string) so every line is at most
22 *width* characters long. Returns a list of output lines, without final
23 newlines.
Georg Brandl116aa622007-08-15 14:28:22 +000024
25 Optional keyword arguments correspond to the instance attributes of
26 :class:`TextWrapper`, documented below. *width* defaults to ``70``.
27
R David Murray1585b702012-09-08 13:13:25 -040028 See the :meth:`TextWrapper.wrap` method for additional details on how
29 :func:`wrap` behaves.
30
Georg Brandl116aa622007-08-15 14:28:22 +000031
Georg Brandl7f01a132009-09-16 15:58:14 +000032.. function:: fill(text, width=70, **kwargs)
Georg Brandl116aa622007-08-15 14:28:22 +000033
34 Wraps the single paragraph in *text*, and returns a single string containing the
35 wrapped paragraph. :func:`fill` is shorthand for ::
36
37 "\n".join(wrap(text, ...))
38
39 In particular, :func:`fill` accepts exactly the same keyword arguments as
40 :func:`wrap`.
41
42Both :func:`wrap` and :func:`fill` work by creating a :class:`TextWrapper`
43instance and calling a single method on it. That instance is not reused, so for
44applications that wrap/fill many text strings, it will be more efficient for you
45to create your own :class:`TextWrapper` object.
46
Alexandre Vassalotti5f8ced22008-05-16 00:03:33 +000047Text is preferably wrapped on whitespaces and right after the hyphens in
48hyphenated words; only then will long words be broken if necessary, unless
49:attr:`TextWrapper.break_long_words` is set to false.
50
Nick Coghlan4fae8cd2012-06-11 23:07:51 +100051Two additional utility function, :func:`dedent` and :func:`indent`, are
52provided to remove indentation from strings that have unwanted whitespace
53to the left of the text and to add an arbitrary prefix to selected lines
54in a block of text.
Georg Brandl116aa622007-08-15 14:28:22 +000055
56.. function:: dedent(text)
57
58 Remove any common leading whitespace from every line in *text*.
59
60 This can be used to make triple-quoted strings line up with the left edge of the
61 display, while still presenting them in the source code in indented form.
62
63 Note that tabs and spaces are both treated as whitespace, but they are not
64 equal: the lines ``" hello"`` and ``"\thello"`` are considered to have no
Georg Brandle6bcc912008-05-12 18:05:20 +000065 common leading whitespace.
Georg Brandl116aa622007-08-15 14:28:22 +000066
67 For example::
68
69 def test():
70 # end first line with \ to avoid the empty line!
71 s = '''\
72 hello
73 world
74 '''
Collin Winterc79461b2007-09-01 23:34:30 +000075 print(repr(s)) # prints ' hello\n world\n '
76 print(repr(dedent(s))) # prints 'hello\n world\n'
Georg Brandl116aa622007-08-15 14:28:22 +000077
78
Nick Coghlan4fae8cd2012-06-11 23:07:51 +100079.. function:: indent(text, prefix, predicate=None)
80
81 Add *prefix* to the beginning of selected lines in *text*.
82
83 Lines are separated by calling ``text.splitlines(True)``.
84
85 By default, *prefix* is added to all lines that do not consist
86 solely of whitespace (including any line endings).
87
88 For example::
89
90 >>> s = 'hello\n\n \nworld'
91 >>> indent(s, ' ')
92 ' hello\n\n \n world'
93
94 The optional *predicate* argument can be used to control which lines
95 are indented. For example, it is easy to add *prefix* to even empty
96 and whitespace-only lines::
97
98 >>> print(indent(s, '+ ', lambda line: True))
99 + hello
100 +
101 +
102 + world
103
104
Georg Brandl7f01a132009-09-16 15:58:14 +0000105.. class:: TextWrapper(**kwargs)
Georg Brandl116aa622007-08-15 14:28:22 +0000106
107 The :class:`TextWrapper` constructor accepts a number of optional keyword
Georg Brandl7f01a132009-09-16 15:58:14 +0000108 arguments. Each keyword argument corresponds to an instance attribute, so
109 for example ::
Georg Brandl116aa622007-08-15 14:28:22 +0000110
111 wrapper = TextWrapper(initial_indent="* ")
112
113 is the same as ::
114
115 wrapper = TextWrapper()
116 wrapper.initial_indent = "* "
117
118 You can re-use the same :class:`TextWrapper` object many times, and you can
119 change any of its options through direct assignment to instance attributes
120 between uses.
121
Benjamin Petersone41251e2008-04-25 01:59:09 +0000122 The :class:`TextWrapper` instance attributes (and keyword arguments to the
123 constructor) are as follows:
Georg Brandl116aa622007-08-15 14:28:22 +0000124
125
Benjamin Petersone41251e2008-04-25 01:59:09 +0000126 .. attribute:: width
Georg Brandl116aa622007-08-15 14:28:22 +0000127
Benjamin Petersone41251e2008-04-25 01:59:09 +0000128 (default: ``70``) The maximum length of wrapped lines. As long as there
129 are no individual words in the input text longer than :attr:`width`,
130 :class:`TextWrapper` guarantees that no output line will be longer than
131 :attr:`width` characters.
Georg Brandl116aa622007-08-15 14:28:22 +0000132
133
Benjamin Petersone41251e2008-04-25 01:59:09 +0000134 .. attribute:: expand_tabs
Georg Brandl116aa622007-08-15 14:28:22 +0000135
Benjamin Petersone41251e2008-04-25 01:59:09 +0000136 (default: ``True``) If true, then all tab characters in *text* will be
137 expanded to spaces using the :meth:`expandtabs` method of *text*.
Georg Brandl116aa622007-08-15 14:28:22 +0000138
139
Hynek Schlawackd5272592012-05-19 13:33:11 +0200140 .. attribute:: tabsize
141
142 (default: ``8``) If :attr:`expand_tabs` is true, then all tab characters
143 in *text* will be expanded to zero or more spaces, depending on the
144 current column and the given tab size.
145
146 .. versionadded:: 3.3
147
148
Benjamin Petersone41251e2008-04-25 01:59:09 +0000149 .. attribute:: replace_whitespace
Georg Brandl116aa622007-08-15 14:28:22 +0000150
Andrew Svetlov59db4012012-08-13 23:22:23 +0300151 (default: ``True``) If true, after tab expansion but before wrapping,
152 the :meth:`wrap` method will replace each whitespace character
153 with a single space. The whitespace characters replaced are
154 as follows: tab, newline, vertical tab, formfeed, and carriage
155 return (``'\t\n\v\f\r'``).
Georg Brandl116aa622007-08-15 14:28:22 +0000156
Benjamin Petersone41251e2008-04-25 01:59:09 +0000157 .. note::
Georg Brandl116aa622007-08-15 14:28:22 +0000158
Benjamin Petersone41251e2008-04-25 01:59:09 +0000159 If :attr:`expand_tabs` is false and :attr:`replace_whitespace` is true,
160 each tab character will be replaced by a single space, which is *not*
161 the same as tab expansion.
Georg Brandl116aa622007-08-15 14:28:22 +0000162
Terry Reedy6d2ab712010-11-23 20:17:24 +0000163 .. note::
164
165 If :attr:`replace_whitespace` is false, newlines may appear in the
166 middle of a line and cause strange output. For this reason, text should
167 be split into paragraphs (using :meth:`str.splitlines` or similar)
168 which are wrapped separately.
169
Georg Brandl116aa622007-08-15 14:28:22 +0000170
Benjamin Petersone41251e2008-04-25 01:59:09 +0000171 .. attribute:: drop_whitespace
Georg Brandl116aa622007-08-15 14:28:22 +0000172
R David Murray1585b702012-09-08 13:13:25 -0400173 (default: ``True``) If true, whitespace at the beginning and ending of
174 every line (after wrapping but before indenting) is dropped.
175 Whitespace at the beginning of the paragraph, however, is not dropped
176 if non-whitespace follows it. If whitespace being dropped takes up an
177 entire line, the whole line is dropped.
Georg Brandl116aa622007-08-15 14:28:22 +0000178
Georg Brandl116aa622007-08-15 14:28:22 +0000179
Benjamin Petersone41251e2008-04-25 01:59:09 +0000180 .. attribute:: initial_indent
Georg Brandl116aa622007-08-15 14:28:22 +0000181
Benjamin Petersone41251e2008-04-25 01:59:09 +0000182 (default: ``''``) String that will be prepended to the first line of
R David Murray1585b702012-09-08 13:13:25 -0400183 wrapped output. Counts towards the length of the first line. The empty
184 string is not indented.
Georg Brandl116aa622007-08-15 14:28:22 +0000185
186
Benjamin Petersone41251e2008-04-25 01:59:09 +0000187 .. attribute:: subsequent_indent
Georg Brandl116aa622007-08-15 14:28:22 +0000188
Benjamin Petersone41251e2008-04-25 01:59:09 +0000189 (default: ``''``) String that will be prepended to all lines of wrapped
190 output except the first. Counts towards the length of each line except
191 the first.
Georg Brandl116aa622007-08-15 14:28:22 +0000192
193
Benjamin Petersone41251e2008-04-25 01:59:09 +0000194 .. attribute:: fix_sentence_endings
Georg Brandl116aa622007-08-15 14:28:22 +0000195
Benjamin Petersone41251e2008-04-25 01:59:09 +0000196 (default: ``False``) If true, :class:`TextWrapper` attempts to detect
197 sentence endings and ensure that sentences are always separated by exactly
198 two spaces. This is generally desired for text in a monospaced font.
199 However, the sentence detection algorithm is imperfect: it assumes that a
200 sentence ending consists of a lowercase letter followed by one of ``'.'``,
201 ``'!'``, or ``'?'``, possibly followed by one of ``'"'`` or ``"'"``,
202 followed by a space. One problem with this is algorithm is that it is
203 unable to detect the difference between "Dr." in ::
Georg Brandl116aa622007-08-15 14:28:22 +0000204
Benjamin Petersone41251e2008-04-25 01:59:09 +0000205 [...] Dr. Frankenstein's monster [...]
Georg Brandl116aa622007-08-15 14:28:22 +0000206
Benjamin Petersone41251e2008-04-25 01:59:09 +0000207 and "Spot." in ::
Georg Brandl116aa622007-08-15 14:28:22 +0000208
Benjamin Petersone41251e2008-04-25 01:59:09 +0000209 [...] See Spot. See Spot run [...]
Georg Brandl116aa622007-08-15 14:28:22 +0000210
Benjamin Petersone41251e2008-04-25 01:59:09 +0000211 :attr:`fix_sentence_endings` is false by default.
Georg Brandl116aa622007-08-15 14:28:22 +0000212
Benjamin Petersone41251e2008-04-25 01:59:09 +0000213 Since the sentence detection algorithm relies on ``string.lowercase`` for
214 the definition of "lowercase letter," and a convention of using two spaces
215 after a period to separate sentences on the same line, it is specific to
216 English-language texts.
Georg Brandl116aa622007-08-15 14:28:22 +0000217
218
Benjamin Petersone41251e2008-04-25 01:59:09 +0000219 .. attribute:: break_long_words
Georg Brandl116aa622007-08-15 14:28:22 +0000220
Benjamin Petersone41251e2008-04-25 01:59:09 +0000221 (default: ``True``) If true, then words longer than :attr:`width` will be
222 broken in order to ensure that no lines are longer than :attr:`width`. If
223 it is false, long words will not be broken, and some lines may be longer
224 than :attr:`width`. (Long words will be put on a line by themselves, in
225 order to minimize the amount by which :attr:`width` is exceeded.)
Georg Brandl116aa622007-08-15 14:28:22 +0000226
Alexandre Vassalotti5f8ced22008-05-16 00:03:33 +0000227
228 .. attribute:: break_on_hyphens
229
230 (default: ``True``) If true, wrapping will occur preferably on whitespaces
231 and right after hyphens in compound words, as it is customary in English.
232 If false, only whitespaces will be considered as potentially good places
233 for line breaks, but you need to set :attr:`break_long_words` to false if
234 you want truly insecable words. Default behaviour in previous versions
235 was to always allow breaking hyphenated words.
236
Alexandre Vassalotti5f8ced22008-05-16 00:03:33 +0000237
Benjamin Petersone41251e2008-04-25 01:59:09 +0000238 :class:`TextWrapper` also provides two public methods, analogous to the
239 module-level convenience functions:
Georg Brandl116aa622007-08-15 14:28:22 +0000240
Benjamin Petersone41251e2008-04-25 01:59:09 +0000241 .. method:: wrap(text)
Georg Brandl116aa622007-08-15 14:28:22 +0000242
Benjamin Petersone41251e2008-04-25 01:59:09 +0000243 Wraps the single paragraph in *text* (a string) so every line is at most
244 :attr:`width` characters long. All wrapping options are taken from
R David Murray1585b702012-09-08 13:13:25 -0400245 instance attributes of the :class:`TextWrapper` instance. Returns a list
246 of output lines, without final newlines. If the wrapped output has no
247 content, the returned list is empty.
Georg Brandl116aa622007-08-15 14:28:22 +0000248
249
Benjamin Petersone41251e2008-04-25 01:59:09 +0000250 .. method:: fill(text)
Georg Brandl116aa622007-08-15 14:28:22 +0000251
Benjamin Petersone41251e2008-04-25 01:59:09 +0000252 Wraps the single paragraph in *text*, and returns a single string
253 containing the wrapped paragraph.
Georg Brandl116aa622007-08-15 14:28:22 +0000254