docutils/utils/smartquotes.py

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