Mercurial > repos > shellac > guppy_basecaller

annotate env/lib/python3.7/site-packages/docutils/utils/smartquotes.py @ 0:26e78fe6e8c4 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
author shellac
date Sat, 02 May 2020 07:14:21 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1 #!/usr/bin/python
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
2 # -*- coding: utf-8 -*-
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
3
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
4 # :Id: $Id: smartquotes.py 8376 2019-08-27 19:49:29Z milde $
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
5 # :Copyright: © 2010 Günter Milde,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
6 # original `SmartyPants`_: © 2003 John Gruber
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
7 # smartypants.py: © 2004, 2007 Chad Miller
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
8 # :Maintainer: docutils-develop@lists.sourceforge.net
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
9 # :License: Released under the terms of the `2-Clause BSD license`_, in short:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
10 #
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
11 # Copying and distribution of this file, with or without modification,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
12 # are permitted in any medium without royalty provided the copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
13 # notices and this notice are preserved.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
14 # This file is offered as-is, without any warranty.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
15 #
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
16 # .. _2-Clause BSD license: http://www.spdx.org/licenses/BSD-2-Clause
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
17
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
18
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
19 r"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
20 =========================
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
21 Smart Quotes for Docutils
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
22 =========================
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
23
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
24 Synopsis
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
25 ========
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
26
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
27 "SmartyPants" is a free web publishing plug-in for Movable Type, Blosxom, and
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
28 BBEdit that easily translates plain ASCII punctuation characters into "smart"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
29 typographic punctuation characters.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
30
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
31 ``smartquotes.py`` is an adaption of "SmartyPants" to Docutils_.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
32
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
33 * Using Unicode instead of HTML entities for typographic punctuation
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
34 characters, it works for any output format that supports Unicode.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
35 * Supports `language specific quote characters`__.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
36
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
37 __ http://en.wikipedia.org/wiki/Non-English_usage_of_quotation_marks
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
38
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
39
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
40 Authors
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
41 =======
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
42
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
43 `John Gruber`_ did all of the hard work of writing this software in Perl for
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
44 `Movable Type`_ and almost all of this useful documentation. `Chad Miller`_
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
45 ported it to Python to use with Pyblosxom_.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
46 Adapted to Docutils_ by Günter Milde.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
47
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
48 Additional Credits
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
49 ==================
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
50
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
51 Portions of the SmartyPants original work are based on Brad Choate's nifty
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
52 MTRegex plug-in. `Brad Choate`_ also contributed a few bits of source code to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
53 this plug-in. Brad Choate is a fine hacker indeed.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
54
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
55 `Jeremy Hedley`_ and `Charles Wiltgen`_ deserve mention for exemplary beta
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
56 testing of the original SmartyPants.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
57
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
58 `Rael Dornfest`_ ported SmartyPants to Blosxom.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
59
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
60 .. _Brad Choate: http://bradchoate.com/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
61 .. _Jeremy Hedley: http://antipixel.com/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
62 .. _Charles Wiltgen: http://playbacktime.com/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
63 .. _Rael Dornfest: http://raelity.org/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
64
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
65
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
66 Copyright and License
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
67 =====================
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
68
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
69 SmartyPants_ license (3-Clause BSD license):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
70
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
71 Copyright (c) 2003 John Gruber (http://daringfireball.net/)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
72 All rights reserved.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
73
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
74 Redistribution and use in source and binary forms, with or without
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
75 modification, are permitted provided that the following conditions are
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
76 met:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
77
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
78 * Redistributions of source code must retain the above copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
79 notice, this list of conditions and the following disclaimer.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
80
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
81 * Redistributions in binary form must reproduce the above copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
82 notice, this list of conditions and the following disclaimer in
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
83 the documentation and/or other materials provided with the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
84 distribution.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
85
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
86 * Neither the name "SmartyPants" nor the names of its contributors
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
87 may be used to endorse or promote products derived from this
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
88 software without specific prior written permission.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
89
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
90 This software is provided by the copyright holders and contributors
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
91 "as is" and any express or implied warranties, including, but not
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
92 limited to, the implied warranties of merchantability and fitness for
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
93 a particular purpose are disclaimed. In no event shall the copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
94 owner or contributors be liable for any direct, indirect, incidental,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
95 special, exemplary, or consequential damages (including, but not
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
96 limited to, procurement of substitute goods or services; loss of use,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
97 data, or profits; or business interruption) however caused and on any
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
98 theory of liability, whether in contract, strict liability, or tort
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
99 (including negligence or otherwise) arising in any way out of the use
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
100 of this software, even if advised of the possibility of such damage.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
101
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
102 smartypants.py license (2-Clause BSD license):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
103
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
104 smartypants.py is a derivative work of SmartyPants.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
105
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
106 Redistribution and use in source and binary forms, with or without
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
107 modification, are permitted provided that the following conditions are
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
108 met:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
109
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
110 * Redistributions of source code must retain the above copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
111 notice, this list of conditions and the following disclaimer.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
112
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
113 * Redistributions in binary form must reproduce the above copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
114 notice, this list of conditions and the following disclaimer in
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
115 the documentation and/or other materials provided with the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
116 distribution.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
117
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
118 This software is provided by the copyright holders and contributors
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
119 "as is" and any express or implied warranties, including, but not
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
120 limited to, the implied warranties of merchantability and fitness for
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
121 a particular purpose are disclaimed. In no event shall the copyright
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
122 owner or contributors be liable for any direct, indirect, incidental,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
123 special, exemplary, or consequential damages (including, but not
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
124 limited to, procurement of substitute goods or services; loss of use,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
125 data, or profits; or business interruption) however caused and on any
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
126 theory of liability, whether in contract, strict liability, or tort
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
127 (including negligence or otherwise) arising in any way out of the use
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
128 of this software, even if advised of the possibility of such damage.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
129
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
130 .. _John Gruber: http://daringfireball.net/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
131 .. _Chad Miller: http://web.chad.org/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
132
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
133 .. _Pyblosxom: http://pyblosxom.bluesock.org/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
134 .. _SmartyPants: http://daringfireball.net/projects/smartypants/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
135 .. _Movable Type: http://www.movabletype.org/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
136 .. _2-Clause BSD license: http://www.spdx.org/licenses/BSD-2-Clause
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
137 .. _Docutils: http://docutils.sf.net/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
138
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
139 Description
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
140 ===========
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
141
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
142 SmartyPants can perform the following transformations:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
143
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
144 - Straight quotes ( " and ' ) into "curly" quote characters
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
145 - Backticks-style quotes (\`\`like this'') into "curly" quote characters
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
146 - Dashes (``--`` and ``---``) into en- and em-dash entities
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
147 - Three consecutive dots (``...`` or ``. . .``) into an ellipsis entity
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
148
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
149 This means you can write, edit, and save your posts using plain old
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
150 ASCII straight quotes, plain dashes, and plain dots, but your published
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
151 posts (and final HTML output) will appear with smart quotes, em-dashes,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
152 and proper ellipses.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
153
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
154 SmartyPants does not modify characters within ``<pre>``, ``<code>``, ``<kbd>``,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
155 ``<math>`` or ``<script>`` tag blocks. Typically, these tags are used to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
156 display text where smart quotes and other "smart punctuation" would not be
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
157 appropriate, such as source code or example markup.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
158
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
159
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
160 Backslash Escapes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
161 =================
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
162
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
163 If you need to use literal straight quotes (or plain hyphens and periods),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
164 `smartquotes` accepts the following backslash escape sequences to force
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
165 ASCII-punctuation. Mind, that you need two backslashes as Docutils expands it,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
166 too.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
167
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
168 ======== =========
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
169 Escape Character
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
170 ======== =========
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
171 ``\\`` \\
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
172 ``\\"`` \\"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
173 ``\\'`` \\'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
174 ``\\.`` \\.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
175 ``\\-`` \\-
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
176 ``\\``` \\`
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
177 ======== =========
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
178
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
179 This is useful, for example, when you want to use straight quotes as
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
180 foot and inch marks: 6\\'2\\" tall; a 17\\" iMac.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
181
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
182
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
183 Caveats
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
184 =======
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
185
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
186 Why You Might Not Want to Use Smart Quotes in Your Weblog
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
187 ---------------------------------------------------------
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
188
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
189 For one thing, you might not care.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
190
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
191 Most normal, mentally stable individuals do not take notice of proper
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
192 typographic punctuation. Many design and typography nerds, however, break
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
193 out in a nasty rash when they encounter, say, a restaurant sign that uses
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
194 a straight apostrophe to spell "Joe's".
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
195
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
196 If you're the sort of person who just doesn't care, you might well want to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
197 continue not caring. Using straight quotes -- and sticking to the 7-bit
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
198 ASCII character set in general -- is certainly a simpler way to live.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
199
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
200 Even if you *do* care about accurate typography, you still might want to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
201 think twice before educating the quote characters in your weblog. One side
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
202 effect of publishing curly quote characters is that it makes your
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
203 weblog a bit harder for others to quote from using copy-and-paste. What
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
204 happens is that when someone copies text from your blog, the copied text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
205 contains the 8-bit curly quote characters (as well as the 8-bit characters
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
206 for em-dashes and ellipses, if you use these options). These characters
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
207 are not standard across different text encoding methods, which is why they
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
208 need to be encoded as characters.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
209
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
210 People copying text from your weblog, however, may not notice that you're
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
211 using curly quotes, and they'll go ahead and paste the unencoded 8-bit
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
212 characters copied from their browser into an email message or their own
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
213 weblog. When pasted as raw "smart quotes", these characters are likely to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
214 get mangled beyond recognition.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
215
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
216 That said, my own opinion is that any decent text editor or email client
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
217 makes it easy to stupefy smart quote characters into their 7-bit
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
218 equivalents, and I don't consider it my problem if you're using an
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
219 indecent text editor or email client.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
220
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
221
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
222 Algorithmic Shortcomings
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
223 ------------------------
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
224
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
225 One situation in which quotes will get curled the wrong way is when
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
226 apostrophes are used at the start of leading contractions. For example::
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
227
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
228 'Twas the night before Christmas.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
229
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
230 In the case above, SmartyPants will turn the apostrophe into an opening
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
231 single-quote, when in fact it should be the `right single quotation mark`
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
232 character which is also "the preferred character to use for apostrophe"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
233 (Unicode). I don't think this problem can be solved in the general case --
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
234 every word processor I've tried gets this wrong as well. In such cases, it's
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
235 best to use the proper character for closing single-quotes (’) by hand.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
236
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
237 In English, the same character is used for apostrophe and closing single
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
238 quote (both plain and "smart" ones). For other locales (French, Italean,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
239 Swiss, ...) "smart" single closing quotes differ from the curly apostrophe.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
240
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
241 .. class:: language-fr
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
242
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
243 Il dit : "C'est 'super' !"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
244
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
245 If the apostrophe is used at the end of a word, it cannot be distinguished
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
246 from a single quote by the algorithm. Therefore, a text like::
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
247
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
248 .. class:: language-de-CH
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
249
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
250 "Er sagt: 'Ich fass' es nicht.'"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
251
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
252 will get a single closing guillemet instead of an apostrophe.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
253
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
254 This can be prevented by use use of the curly apostrophe character (’) in
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
255 the source::
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
256
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
257 - "Er sagt: 'Ich fass' es nicht.'"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
258 + "Er sagt: 'Ich fass’ es nicht.'"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
259
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
260
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
261 Version History
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
262 ===============
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
263
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
264 1.8.1 2017-10-25
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
265 - Use open quote after Unicode whitespace, ZWSP, and ZWNJ.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
266 - Code cleanup.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
267
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
268 1.8: 2017-04-24
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
269 - Command line front-end.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
270
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
271 1.7.1: 2017-03-19
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
272 - Update and extend language-dependent quotes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
273 - Differentiate apostrophe from single quote.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
274
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
275 1.7: 2012-11-19
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
276 - Internationalization: language-dependent quotes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
277
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
278 1.6.1: 2012-11-06
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
279 - Refactor code, code cleanup,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
280 - `educate_tokens()` generator as interface for Docutils.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
281
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
282 1.6: 2010-08-26
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
283 - Adaption to Docutils:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
284 - Use Unicode instead of HTML entities,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
285 - Remove code special to pyblosxom.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
286
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
287 1.5_1.6: Fri, 27 Jul 2007 07:06:40 -0400
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
288 - Fixed bug where blocks of precious unalterable text was instead
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
289 interpreted. Thanks to Le Roux and Dirk van Oosterbosch.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
290
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
291 1.5_1.5: Sat, 13 Aug 2005 15:50:24 -0400
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
292 - Fix bogus magical quotation when there is no hint that the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
293 user wants it, e.g., in "21st century". Thanks to Nathan Hamblen.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
294 - Be smarter about quotes before terminating numbers in an en-dash'ed
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
295 range.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
296
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
297 1.5_1.4: Thu, 10 Feb 2005 20:24:36 -0500
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
298 - Fix a date-processing bug, as reported by jacob childress.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
299 - Begin a test-suite for ensuring correct output.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
300 - Removed import of "string", since I didn't really need it.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
301 (This was my first every Python program. Sue me!)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
302
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
303 1.5_1.3: Wed, 15 Sep 2004 18:25:58 -0400
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
304 - Abort processing if the flavour is in forbidden-list. Default of
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
305 [ "rss" ] (Idea of Wolfgang SCHNERRING.)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
306 - Remove stray virgules from en-dashes. Patch by Wolfgang SCHNERRING.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
307
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
308 1.5_1.2: Mon, 24 May 2004 08:14:54 -0400
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
309 - Some single quotes weren't replaced properly. Diff-tesuji played
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
310 by Benjamin GEIGER.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
311
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
312 1.5_1.1: Sun, 14 Mar 2004 14:38:28 -0500
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
313 - Support upcoming pyblosxom 0.9 plugin verification feature.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
314
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
315 1.5_1.0: Tue, 09 Mar 2004 08:08:35 -0500
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
316 - Initial release
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
317 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
318 from __future__ import print_function
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
319
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
320 options = r"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
321 Options
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
322 =======
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
323
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
324 Numeric values are the easiest way to configure SmartyPants' behavior:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
325
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
326 :0: Suppress all transformations. (Do nothing.)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
327
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
328 :1: Performs default SmartyPants transformations: quotes (including
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
329 \`\`backticks'' -style), em-dashes, and ellipses. "``--``" (dash dash)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
330 is used to signify an em-dash; there is no support for en-dashes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
331
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
332 :2: Same as smarty_pants="1", except that it uses the old-school typewriter
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
333 shorthand for dashes: "``--``" (dash dash) for en-dashes, "``---``"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
334 (dash dash dash)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
335 for em-dashes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
336
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
337 :3: Same as smarty_pants="2", but inverts the shorthand for dashes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
338 "``--``" (dash dash) for em-dashes, and "``---``" (dash dash dash) for
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
339 en-dashes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
340
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
341 :-1: Stupefy mode. Reverses the SmartyPants transformation process, turning
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
342 the characters produced by SmartyPants into their ASCII equivalents.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
343 E.g. the LEFT DOUBLE QUOTATION MARK (“) is turned into a simple
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
344 double-quote (\"), "—" is turned into two dashes, etc.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
345
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
346
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
347 The following single-character attribute values can be combined to toggle
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
348 individual transformations from within the smarty_pants attribute. For
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
349 example, ``"1"`` is equivalent to ``"qBde"``.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
350
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
351 :q: Educates normal quote characters: (") and (').
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
352
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
353 :b: Educates \`\`backticks'' -style double quotes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
354
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
355 :B: Educates \`\`backticks'' -style double quotes and \`single' quotes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
356
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
357 :d: Educates em-dashes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
358
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
359 :D: Educates em-dashes and en-dashes, using old-school typewriter shorthand:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
360 (dash dash) for en-dashes, (dash dash dash) for em-dashes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
361
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
362 :i: Educates em-dashes and en-dashes, using inverted old-school typewriter
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
363 shorthand: (dash dash) for em-dashes, (dash dash dash) for en-dashes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
364
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
365 :e: Educates ellipses.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
366
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
367 :w: Translates any instance of ``&quot;`` into a normal double-quote character.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
368 This should be of no interest to most people, but of particular interest
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
369 to anyone who writes their posts using Dreamweaver, as Dreamweaver
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
370 inexplicably uses this entity to represent a literal double-quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
371 character. SmartyPants only educates normal quotes, not entities (because
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
372 ordinarily, entities are used for the explicit purpose of representing the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
373 specific character they represent). The "w" option must be used in
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
374 conjunction with one (or both) of the other quote options ("q" or "b").
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
375 Thus, if you wish to apply all SmartyPants transformations (quotes, en-
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
376 and em-dashes, and ellipses) and also translate ``&quot;`` entities into
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
377 regular quotes so SmartyPants can educate them, you should pass the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
378 following to the smarty_pants attribute:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
379 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
380
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
381
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
382 default_smartypants_attr = "1"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
383
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
384
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
385 import re, sys
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
386
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
387 class smartchars(object):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
388 """Smart quotes and dashes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
389 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
390
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
391 endash = u'–' # "&#8211;" EN DASH
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
392 emdash = u'—' # "&#8212;" EM DASH
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
393 ellipsis = u'…' # "&#8230;" HORIZONTAL ELLIPSIS
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
394 apostrophe = u'’' # "&#8217;" RIGHT SINGLE QUOTATION MARK
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
395
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
396 # quote characters (language-specific, set in __init__())
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
397 # [1] http://en.wikipedia.org/wiki/Non-English_usage_of_quotation_marks
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
398 # [2] http://de.wikipedia.org/wiki/Anf%C3%BChrungszeichen#Andere_Sprachen
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
399 # [3] https://fr.wikipedia.org/wiki/Guillemet
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
400 # [4] http://typographisme.net/post/Les-espaces-typographiques-et-le-web
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
401 # [5] http://www.btb.termiumplus.gc.ca/tpv2guides/guides/redac/index-fra.html
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
402 # [6] https://en.wikipedia.org/wiki/Hebrew_punctuation#Quotation_marks
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
403 # [7] http://www.tustep.uni-tuebingen.de/bi/bi00/bi001t1-anfuehrung.pdf
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
404 # [8] http://www.korrekturavdelingen.no/anforselstegn.htm
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
405 # [9] Typografisk håndbok. Oslo: Spartacus. 2000. s. 67. ISBN 8243001530.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
406 # [10] http://www.typografi.org/sitat/sitatart.html
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
407 #
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
408 # See also configuration option "smartquote-locales".
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
409 quotes = {'af': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
410 'af-x-altquot': u'„”‚’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
411 'bg': u'„“‚‘', # Bulgarian, https://bg.wikipedia.org/wiki/Кавички
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
412 'ca': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
413 'ca-x-altquot': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
414 'cs': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
415 'cs-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
416 'da': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
417 'da-x-altquot': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
418 # 'da-x-altquot2': u'””’’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
419 'de': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
420 'de-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
421 'de-ch': u'«»‹›',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
422 'el': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
423 'en': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
424 'en-uk-x-altquot': u'‘’“”', # Attention: " → ‘ and ' → “ !
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
425 'eo': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
426 'es': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
427 'es-x-altquot': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
428 'et': u'„“‚‘', # no secondary quote listed in
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
429 'et-x-altquot': u'«»‹›', # the sources above (wikipedia.org)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
430 'eu': u'«»‹›',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
431 'fi': u'””’’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
432 'fi-x-altquot': u'»»››',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
433 'fr': (u'« ', u' »', u'“', u'”'), # full no-break space
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
434 'fr-x-altquot': (u'« ', u' »', u'“', u'”'), # narrow no-break space
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
435 'fr-ch': u'«»‹›',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
436 'fr-ch-x-altquot': (u'« ', u' »', u'‹ ', u' ›'), # narrow no-break space, http://typoguide.ch/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
437 'gl': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
438 'he': u'”“»«', # Hebrew is RTL, test position:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
439 'he-x-altquot': u'„”‚’', # low quotation marks are opening.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
440 # 'he-x-altquot': u'“„‘‚', # RTL: low quotation marks opening
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
441 'hr': u'„”‘’', # http://hrvatska-tipografija.com/polunavodnici/
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
442 'hr-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
443 'hsb': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
444 'hsb-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
445 'hu': u'„”«»',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
446 'is': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
447 'it': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
448 'it-ch': u'«»‹›',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
449 'it-x-altquot': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
450 # 'it-x-altquot2': u'“„‘‚', # [7] in headlines
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
451 'ja': u'「」『』',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
452 'ko': u'《》〈〉',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
453 'lt': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
454 'lv': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
455 'mk': u'„“‚‘', # Macedonian, https://mk.wikipedia.org/wiki/Правопис_и_правоговор_на_македонскиот_јазик
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
456 'nl': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
457 'nl-x-altquot': u'„”‚’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
458 # 'nl-x-altquot2': u'””’’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
459 'nb': u'«»’’', # Norsk bokmål (canonical form 'no')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
460 'nn': u'«»’’', # Nynorsk [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
461 'nn-x-altquot': u'«»‘’', # [8], [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
462 # 'nn-x-altquot2': u'«»«»', # [9], [10
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
463 # 'nn-x-altquot3': u'„“‚‘', # [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
464 'no': u'«»’’', # Norsk bokmål [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
465 'no-x-altquot': u'«»‘’', # [8], [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
466 # 'no-x-altquot2': u'«»«»', # [9], [10
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
467 # 'no-x-altquot3': u'„“‚‘', # [10]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
468 'pl': u'„”«»',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
469 'pl-x-altquot': u'«»‚’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
470 # 'pl-x-altquot2': u'„”‚’', # https://pl.wikipedia.org/wiki/Cudzys%C5%82%C3%B3w
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
471 'pt': u'«»“”',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
472 'pt-br': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
473 'ro': u'„”«»',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
474 'ru': u'«»„“',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
475 'sh': u'„”‚’', # Serbo-Croatian
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
476 'sh-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
477 'sk': u'„“‚‘', # Slovak
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
478 'sk-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
479 'sl': u'„“‚‘', # Slovenian
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
480 'sl-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
481 'sq': u'«»‹›', # Albanian
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
482 'sq-x-altquot': u'“„‘‚',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
483 'sr': u'„”’’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
484 'sr-x-altquot': u'»«›‹',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
485 'sv': u'””’’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
486 'sv-x-altquot': u'»»››',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
487 'tr': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
488 'tr-x-altquot': u'«»‹›',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
489 # 'tr-x-altquot2': u'“„‘‚', # [7] antiquated?
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
490 'uk': u'«»„“',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
491 'uk-x-altquot': u'„“‚‘',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
492 'zh-cn': u'“”‘’',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
493 'zh-tw': u'「」『』',
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
494 }
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
495
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
496 def __init__(self, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
497 self.language = language
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
498 try:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
499 (self.opquote, self.cpquote,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
500 self.osquote, self.csquote) = self.quotes[language.lower()]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
501 except KeyError:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
502 self.opquote, self.cpquote, self.osquote, self.csquote = u'""\'\''
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
503
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
504
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
505 def smartyPants(text, attr=default_smartypants_attr, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
506 """Main function for "traditional" use."""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
507
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
508 return "".join([t for t in educate_tokens(tokenize(text),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
509 attr, language)])
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
510
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
511
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
512 def educate_tokens(text_tokens, attr=default_smartypants_attr, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
513 """Return iterator that "educates" the items of `text_tokens`.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
514 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
515
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
516 # Parse attributes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
517 # 0 : do nothing
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
518 # 1 : set all
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
519 # 2 : set all, using old school en- and em- dash shortcuts
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
520 # 3 : set all, using inverted old school en and em- dash shortcuts
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
521 #
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
522 # q : quotes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
523 # b : backtick quotes (``double'' only)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
524 # B : backtick quotes (``double'' and `single')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
525 # d : dashes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
526 # D : old school dashes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
527 # i : inverted old school dashes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
528 # e : ellipses
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
529 # w : convert &quot; entities to " for Dreamweaver users
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
530
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
531 convert_quot = False # translate &quot; entities into normal quotes?
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
532 do_dashes = False
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
533 do_backticks = False
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
534 do_quotes = False
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
535 do_ellipses = False
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
536 do_stupefy = False
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
537
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
538 # if attr == "0": # pass tokens unchanged (see below).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
539 if attr == "1": # Do everything, turn all options on.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
540 do_quotes = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
541 do_backticks = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
542 do_dashes = 1
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
543 do_ellipses = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
544 elif attr == "2":
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
545 # Do everything, turn all options on, use old school dash shorthand.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
546 do_quotes = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
547 do_backticks = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
548 do_dashes = 2
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
549 do_ellipses = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
550 elif attr == "3":
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
551 # Do everything, use inverted old school dash shorthand.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
552 do_quotes = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
553 do_backticks = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
554 do_dashes = 3
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
555 do_ellipses = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
556 elif attr == "-1": # Special "stupefy" mode.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
557 do_stupefy = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
558 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
559 if "q" in attr: do_quotes = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
560 if "b" in attr: do_backticks = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
561 if "B" in attr: do_backticks = 2
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
562 if "d" in attr: do_dashes = 1
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
563 if "D" in attr: do_dashes = 2
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
564 if "i" in attr: do_dashes = 3
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
565 if "e" in attr: do_ellipses = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
566 if "w" in attr: convert_quot = True
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
567
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
568 prev_token_last_char = " "
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
569 # Last character of the previous text token. Used as
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
570 # context to curl leading quote characters correctly.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
571
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
572 for (ttype, text) in text_tokens:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
573
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
574 # skip HTML and/or XML tags as well as emtpy text tokens
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
575 # without updating the last character
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
576 if ttype == 'tag' or not text:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
577 yield text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
578 continue
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
579
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
580 # skip literal text (math, literal, raw, ...)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
581 if ttype == 'literal':
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
582 prev_token_last_char = text[-1:]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
583 yield text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
584 continue
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
585
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
586 last_char = text[-1:] # Remember last char before processing.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
587
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
588 text = processEscapes(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
589
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
590 if convert_quot:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
591 text = re.sub('&quot;', '"', text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
592
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
593 if do_dashes == 1:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
594 text = educateDashes(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
595 elif do_dashes == 2:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
596 text = educateDashesOldSchool(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
597 elif do_dashes == 3:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
598 text = educateDashesOldSchoolInverted(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
599
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
600 if do_ellipses:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
601 text = educateEllipses(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
602
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
603 # Note: backticks need to be processed before quotes.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
604 if do_backticks:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
605 text = educateBackticks(text, language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
606
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
607 if do_backticks == 2:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
608 text = educateSingleBackticks(text, language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
609
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
610 if do_quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
611 # Replace plain quotes in context to prevent converstion to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
612 # 2-character sequence in French.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
613 context = prev_token_last_char.replace('"', ';').replace("'", ';')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
614 text = educateQuotes(context+text, language)[1:]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
615
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
616 if do_stupefy:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
617 text = stupefyEntities(text, language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
618
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
619 # Remember last char as context for the next token
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
620 prev_token_last_char = last_char
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
621
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
622 text = processEscapes(text, restore=True)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
623
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
624 yield text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
625
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
626
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
627
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
628 def educateQuotes(text, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
629 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
630 Parameter: - text string (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
631 - language (`BCP 47` language tag.)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
632 Returns: The `text`, with "educated" curly quote characters.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
633
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
634 Example input: "Isn't this fun?"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
635 Example output: “Isn’t this fun?“;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
636 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
637
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
638 smart = smartchars(language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
639
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
640 punct_class = r"""[!"#\$\%'()*+,-.\/:;<=>?\@\[\\\]\^_`{|}~]"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
641 close_class = r"""[^\ \t\r\n\[\{\(\-]"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
642 open_class = u'[\u200B\u200C]' # ZWSP, ZWNJ
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
643 dec_dashes = r"""&#8211;|&#8212;"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
644
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
645 # Special case if the very first character is a quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
646 # followed by punctuation at a non-word-break.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
647 # Close the quotes by brute force:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
648 text = re.sub(r"^'(?=%s\\B)" % (punct_class,), smart.csquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
649 text = re.sub(r'^"(?=%s\\B)' % (punct_class,), smart.cpquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
650
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
651 # Special case for double sets of quotes, e.g.:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
652 # <p>He said, "'Quoted' words in a larger quote."</p>
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
653 text = re.sub(r""""'(?=\w)""", smart.opquote+smart.osquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
654 text = re.sub(r"""'"(?=\w)""", smart.osquote+smart.opquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
655
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
656 # Special case for decade abbreviations (the '80s):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
657 if language.startswith('en'): # TODO similar cases in other languages?
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
658 text = re.sub(r"'(?=\d{2}s)", smart.apostrophe, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
659
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
660 # Get most opening single quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
661 opening_single_quotes_regex = re.compile(u"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
662 (# ?<= # look behind fails: requires fixed-width pattern
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
663 \\s | # a whitespace char, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
664 %s | # another separating char, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
665 &nbsp; | # a non-breaking space entity, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
666 [\u2013 \u2014 ] | # literal dashes, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
667 -- | # dumb dashes, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
668 &[mn]dash; | # dash entities (named or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
669 %s | # decimal or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
670 &\\#x201[34]; # hex)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
671 )
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
672 ' # the quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
673 (?=\\w) # followed by a word character
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
674 """ % (open_class, dec_dashes), re.VERBOSE | re.UNICODE)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
675
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
676 text = opening_single_quotes_regex.sub(r'\1'+smart.osquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
677
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
678 # In many locales, single closing quotes are different from apostrophe:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
679 if smart.csquote != smart.apostrophe:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
680 apostrophe_regex = re.compile(r"(?<=(\w|\d))'(?=\w)", re.UNICODE)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
681 text = apostrophe_regex.sub(smart.apostrophe, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
682 # TODO: keep track of quoting level to recognize apostrophe in, e.g.,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
683 # "Ich fass' es nicht."
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
684
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
685 closing_single_quotes_regex = re.compile(r"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
686 (?<=%s)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
687 '
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
688 """ % close_class, re.VERBOSE)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
689 text = closing_single_quotes_regex.sub(smart.csquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
690
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
691 # Any remaining single quotes should be opening ones:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
692 text = re.sub(r"""'""", smart.osquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
693
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
694 # Get most opening double quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
695 opening_double_quotes_regex = re.compile(u"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
696 (
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
697 \\s | # a whitespace char, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
698 %s | # another separating char, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
699 &nbsp; | # a non-breaking space entity, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
700 [\u2013 \u2014 ] | # literal dashes, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
701 -- | # dumb dashes, or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
702 &[mn]dash; | # dash entities (named or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
703 %s | # decimal or
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
704 &\\#x201[34]; # hex)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
705 )
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
706 " # the quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
707 (?=\\w) # followed by a word character
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
708 """ % (open_class, dec_dashes), re.VERBOSE | re.UNICODE)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
709
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
710 text = opening_double_quotes_regex.sub(r'\1'+smart.opquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
711
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
712 # Double closing quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
713 closing_double_quotes_regex = re.compile(r"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
714 (
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
715 (?<=%s)" | # char indicating the quote should be closing
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
716 "(?=\s) # whitespace behind
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
717 )
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
718 """ % (close_class,), re.VERBOSE | re.UNICODE)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
719 text = closing_double_quotes_regex.sub(smart.cpquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
720
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
721 # Any remaining quotes should be opening ones.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
722 text = re.sub(r'"', smart.opquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
723
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
724 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
725
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
726
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
727 def educateBackticks(text, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
728 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
729 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
730 Returns: The `text`, with ``backticks'' -style double quotes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
731 translated into HTML curly quote entities.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
732 Example input: ``Isn't this fun?''
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
733 Example output: “Isn't this fun?“;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
734 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
735 smart = smartchars(language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
736
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
737 text = re.sub(r"""``""", smart.opquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
738 text = re.sub(r"""''""", smart.cpquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
739 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
740
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
741
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
742 def educateSingleBackticks(text, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
743 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
744 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
745 Returns: The `text`, with `backticks' -style single quotes
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
746 translated into HTML curly quote entities.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
747
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
748 Example input: `Isn't this fun?'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
749 Example output: ‘Isn’t this fun?’
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
750 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
751 smart = smartchars(language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
752
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
753 text = re.sub(r"""`""", smart.osquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
754 text = re.sub(r"""'""", smart.csquote, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
755 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
756
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
757
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
758 def educateDashes(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
759 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
760 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
761 Returns: The `text`, with each instance of "--" translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
762 an em-dash character.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
763 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
764
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
765 text = re.sub(r"""---""", smartchars.endash, text) # en (yes, backwards)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
766 text = re.sub(r"""--""", smartchars.emdash, text) # em (yes, backwards)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
767 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
768
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
769
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
770 def educateDashesOldSchool(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
771 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
772 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
773 Returns: The `text`, with each instance of "--" translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
774 an en-dash character, and each "---" translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
775 an em-dash character.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
776 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
777
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
778 text = re.sub(r"""---""", smartchars.emdash, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
779 text = re.sub(r"""--""", smartchars.endash, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
780 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
781
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
782
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
783 def educateDashesOldSchoolInverted(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
784 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
785 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
786 Returns: The `text`, with each instance of "--" translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
787 an em-dash character, and each "---" translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
788 an en-dash character. Two reasons why: First, unlike the
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
789 en- and em-dash syntax supported by
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
790 EducateDashesOldSchool(), it's compatible with existing
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
791 entries written before SmartyPants 1.1, back when "--" was
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
792 only used for em-dashes. Second, em-dashes are more
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
793 common than en-dashes, and so it sort of makes sense that
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
794 the shortcut should be shorter to type. (Thanks to Aaron
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
795 Swartz for the idea.)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
796 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
797 text = re.sub(r"""---""", smartchars.endash, text) # em
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
798 text = re.sub(r"""--""", smartchars.emdash, text) # en
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
799 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
800
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
801
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
802
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
803 def educateEllipses(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
804 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
805 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
806 Returns: The `text`, with each instance of "..." translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
807 an ellipsis character.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
808
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
809 Example input: Huh...?
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
810 Example output: Huh&#8230;?
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
811 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
812
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
813 text = re.sub(r"""\.\.\.""", smartchars.ellipsis, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
814 text = re.sub(r"""\. \. \.""", smartchars.ellipsis, text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
815 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
816
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
817
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
818 def stupefyEntities(text, language='en'):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
819 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
820 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
821 Returns: The `text`, with each SmartyPants character translated to
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
822 its ASCII counterpart.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
823
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
824 Example input: “Hello — world.”
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
825 Example output: "Hello -- world."
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
826 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
827 smart = smartchars(language)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
828
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
829 text = re.sub(smart.endash, "-", text) # en-dash
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
830 text = re.sub(smart.emdash, "--", text) # em-dash
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
831
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
832 text = re.sub(smart.osquote, "'", text) # open single quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
833 text = re.sub(smart.csquote, "'", text) # close single quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
834
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
835 text = re.sub(smart.opquote, '"', text) # open double quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
836 text = re.sub(smart.cpquote, '"', text) # close double quote
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
837
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
838 text = re.sub(smart.ellipsis, '...', text)# ellipsis
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
839
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
840 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
841
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
842
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
843 def processEscapes(text, restore=False):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
844 r"""
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
845 Parameter: String (unicode or bytes).
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
846 Returns: The `text`, with after processing the following backslash
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
847 escape sequences. This is useful if you want to force a "dumb"
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
848 quote or other character to appear.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
849
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
850 Escape Value
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
851 ------ -----
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
852 \\ &#92;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
853 \" &#34;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
854 \' &#39;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
855 \. &#46;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
856 \- &#45;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
857 \` &#96;
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
858 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
859 replacements = ((r'\\', r'&#92;'),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
860 (r'\"', r'&#34;'),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
861 (r"\'", r'&#39;'),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
862 (r'\.', r'&#46;'),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
863 (r'\-', r'&#45;'),
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
864 (r'\`', r'&#96;'))
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
865 if restore:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
866 for (ch, rep) in replacements:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
867 text = text.replace(rep, ch[1])
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
868 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
869 for (ch, rep) in replacements:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
870 text = text.replace(ch, rep)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
871
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
872 return text
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
873
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
874
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
875 def tokenize(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
876 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
877 Parameter: String containing HTML markup.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
878 Returns: An iterator that yields the tokens comprising the input
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
879 string. Each token is either a tag (possibly with nested,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
880 tags contained therein, such as <a href="<MTFoo>">, or a
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
881 run of text between tags. Each yielded element is a
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
882 two-element tuple; the first is either 'tag' or 'text';
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
883 the second is the actual value.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
884
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
885 Based on the _tokenize() subroutine from Brad Choate's MTRegex plugin.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
886 <http://www.bradchoate.com/past/mtregex.php>
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
887 """
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
888
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
889 pos = 0
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
890 length = len(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
891 # tokens = []
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
892
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
893 depth = 6
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
894 nested_tags = "|".join(['(?:<(?:[^<>]',] * depth) + (')*>)' * depth)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
895 #match = r"""(?: <! ( -- .*? -- \s* )+ > ) | # comments
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
896 # (?: <\? .*? \?> ) | # directives
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
897 # %s # nested tags """ % (nested_tags,)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
898 tag_soup = re.compile(r"""([^<]*)(<[^>]*>)""")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
899
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
900 token_match = tag_soup.search(text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
901
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
902 previous_end = 0
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
903 while token_match is not None:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
904 if token_match.group(1):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
905 yield ('text', token_match.group(1))
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
906
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
907 yield ('tag', token_match.group(2))
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
908
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
909 previous_end = token_match.end()
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
910 token_match = tag_soup.search(text, token_match.end())
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
911
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
912 if previous_end < len(text):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
913 yield ('text', text[previous_end:])
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
914
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
915
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
916
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
917 if __name__ == "__main__":
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
918
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
919 import itertools
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
920 try:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
921 import locale # module missing in Jython
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
922 locale.setlocale(locale.LC_ALL, '') # set to user defaults
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
923 defaultlanguage = locale.getdefaultlocale()[0]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
924 except:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
925 defaultlanguage = 'en'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
926
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
927 # Normalize and drop unsupported subtags:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
928 defaultlanguage = defaultlanguage.lower().replace('-', '_')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
929 # split (except singletons, which mark the following tag as non-standard):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
930 defaultlanguage = re.sub(r'_([a-zA-Z0-9])_', r'_\1-', defaultlanguage)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
931 _subtags = [subtag for subtag in defaultlanguage.split('_')]
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
932 _basetag = _subtags.pop(0)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
933 # find all combinations of subtags
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
934 for n in range(len(_subtags), 0, -1):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
935 for tags in itertools.combinations(_subtags, n):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
936 _tag = '-'.join((_basetag,)+tags)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
937 if _tag in smartchars.quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
938 defaultlanguage = _tag
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
939 break
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
940 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
941 if _basetag in smartchars.quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
942 defaultlanguage = _basetag
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
943 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
944 defaultlanguage = 'en'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
945
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
946
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
947 import argparse
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
948 parser = argparse.ArgumentParser(
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
949 description='Filter stdin making ASCII punctuation "smart".')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
950 # parser.add_argument("text", help="text to be acted on")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
951 parser.add_argument("-a", "--action", default="1",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
952 help="what to do with the input (see --actionhelp)")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
953 parser.add_argument("-e", "--encoding", default="utf8",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
954 help="text encoding")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
955 parser.add_argument("-l", "--language", default=defaultlanguage,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
956 help="text language (BCP47 tag), "
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
957 "Default: %s"% defaultlanguage)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
958 parser.add_argument("-q", "--alternative-quotes", action="store_true",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
959 help="use alternative quote style")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
960 parser.add_argument("--doc", action="store_true",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
961 help="print documentation")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
962 parser.add_argument("--actionhelp", action="store_true",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
963 help="list available actions")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
964 parser.add_argument("--stylehelp", action="store_true",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
965 help="list available quote styles")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
966 parser.add_argument("--test", action="store_true",
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
967 help="perform short self-test")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
968 args = parser.parse_args()
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
969
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
970 if args.doc:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
971 print(__doc__)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
972 elif args.actionhelp:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
973 print(options)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
974 elif args.stylehelp:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
975 print()
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
976 print("Available styles (primary open/close, secondary open/close)")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
977 print("language tag quotes")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
978 print("============ ======")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
979 for key in sorted(smartchars.quotes.keys()):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
980 print("%-14s %s" % (key, smartchars.quotes[key]))
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
981 elif args.test:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
982 # Unit test output goes to stderr.
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
983 import unittest
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
984
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
985 class TestSmartypantsAllAttributes(unittest.TestCase):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
986 # the default attribute is "1", which means "all".
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
987 def test_dates(self):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
988 self.assertEqual(smartyPants("1440-80's"), u"1440-80’s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
989 self.assertEqual(smartyPants("1440-'80s"), u"1440-’80s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
990 self.assertEqual(smartyPants("1440---'80s"), u"1440–’80s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
991 self.assertEqual(smartyPants("1960's"), u"1960’s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
992 self.assertEqual(smartyPants("one two '60s"), u"one two ’60s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
993 self.assertEqual(smartyPants("'60s"), u"’60s")
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
994
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
995 def test_educated_quotes(self):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
996 self.assertEqual(smartyPants('"Isn\'t this fun?"'), u'“Isn’t this fun?”')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
997
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
998 def test_html_tags(self):
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
999 text = '<a src="foo">more</a>'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1000 self.assertEqual(smartyPants(text), text)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1001
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1002 suite = unittest.TestLoader().loadTestsFromTestCase(
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1003 TestSmartypantsAllAttributes)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1004 unittest.TextTestRunner().run(suite)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1005
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1006 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1007 if args.alternative_quotes:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1008 if '-x-altquot' in args.language:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1009 args.language = args.language.replace('-x-altquot', '')
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1010 else:
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1011 args.language += '-x-altquot'
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1012 text = sys.stdin.read().decode(args.encoding)
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1013 print(smartyPants(text, attr=args.action,
26e78fe6e8c4 "planemo upload commit c699937486c35866861690329de38ec1a5d9f783"
shellac
parents:
diff changeset
1014 language=args.language).encode(args.encoding))