docutils/writers/html4css1/__init__.py

   1 # $Id$
   2 # Author: David Goodger
   3 # Maintainer: docutils-develop@lists.sourceforge.net
   4 # Copyright: This module has been placed in the public domain.
   5
   6 """
   7 Simple HyperText Markup Language document tree Writer.
   8
   9 The output conforms to the XHTML version 1.0 Transitional DTD
  10 (*almost* strict).  The output contains a minimum of formatting
  11 information.  The cascading style sheet "html4css1.css" is required
  12 for proper viewing with a modern graphical browser.
  13 """
  14
  15 __docformat__ = 'reStructuredText'
  16
  17
  18 import sys
  19 import os
  20 import os.path
  21 import time
  22 import re
  23 import urllib
  24 try: # check for the Python Imaging Library
  25     import PIL.Image
  26 except ImportError:
  27     try:  # sometimes PIL modules are put in PYTHONPATH's root
  28         import Image
  29         class PIL(object): pass  # dummy wrapper
  30         PIL.Image = Image
  31     except ImportError:
  32         PIL = None
  33 import docutils
  34 from docutils import frontend, nodes, utils, writers, languages, io
  35 from docutils.utils.error_reporting import SafeString
  36 from docutils.transforms import writer_aux
  37 from docutils.utils.math import unichar2tex, pick_math_environment, math2html
  38 from docutils.utils.math.latex2mathml import parse_latex_math
  39
  40 class Writer(writers.Writer):
  41
  42     supported = ('html', 'html4css1', 'xhtml')
  43     """Formats this writer supports."""
  44
  45     default_stylesheet = 'html4css1.css'
  46     default_stylesheet_dirs = ['.', utils.relative_path(
  47         os.path.join(os.getcwd(), 'dummy'), os.path.dirname(__file__))]
  48
  49     default_template = 'template.txt'
  50
  51     default_template_path = utils.relative_path(
  52         os.path.join(os.getcwd(), 'dummy'),
  53         os.path.join(os.path.dirname(__file__), default_template))
  54
  55     settings_spec = (
  56         'HTML-Specific Options',
  57         None,
  58         (('Specify the template file (UTF-8 encoded).  Default is "%s".'
  59           % default_template_path,
  60           ['--template'],
  61           {'default': default_template_path, 'metavar': '<file>'}),
  62          ('Comma separated list of stylesheet URLs. '
  63           'Overrides previous --stylesheet and --stylesheet-path settings.',
  64           ['--stylesheet'],
  65           {'metavar': '<URL[,URL,...]>', 'overrides': 'stylesheet_path',
  66            'validator': frontend.validate_comma_separated_list}),
  67          ('Comma separated list of stylesheet paths. '
  68           'Relative paths are expanded if a matching file is found in '
  69           'the --stylesheet-dirs. With --link-stylesheet, '
  70           'the path is rewritten relative to the output HTML file. '
  71           'Default: "%s"' % default_stylesheet,
  72           ['--stylesheet-path'],
  73           {'metavar': '<file[,file,...]>', 'overrides': 'stylesheet',
  74            'validator': frontend.validate_comma_separated_list,
  75            'default': [default_stylesheet]}),
  76          ('Embed the stylesheet(s) in the output HTML file.  The stylesheet '
  77           'files must be accessible during processing. This is the default.',
  78           ['--embed-stylesheet'],
  79           {'default': 1, 'action': 'store_true',
  80            'validator': frontend.validate_boolean}),
  81          ('Link to the stylesheet(s) in the output HTML file. '
  82           'Default: embed stylesheets.',
  83           ['--link-stylesheet'],
  84           {'dest': 'embed_stylesheet', 'action': 'store_false'}),
  85          ('Comma-separated list of directories where stylesheets are found. '
  86           'Used by --stylesheet-path when expanding relative path arguments. '
  87           'Default: "%s"' % default_stylesheet_dirs,
  88           ['--stylesheet-dirs'],
  89           {'metavar': '<dir[,dir,...]>',
  90            'validator': frontend.validate_comma_separated_list,
  91            'default': default_stylesheet_dirs}),
  92          ('Specify the initial header level.  Default is 1 for "<h1>".  '
  93           'Does not affect document title & subtitle (see --no-doc-title).',
  94           ['--initial-header-level'],
  95           {'choices': '1 2 3 4 5 6'.split(), 'default': '1',
  96            'metavar': '<level>'}),
  97          ('Specify the maximum width (in characters) for one-column field '
  98           'names.  Longer field names will span an entire row of the table '
  99           'used to render the field list.  Default is 14 characters.  '
 100           'Use 0 for "no limit".',
 101           ['--field-name-limit'],
 102           {'default': 14, 'metavar': '<level>',
 103            'validator': frontend.validate_nonnegative_int}),
 104          ('Specify the maximum width (in characters) for options in option '
 105           'lists.  Longer options will span an entire row of the table used '
 106           'to render the option list.  Default is 14 characters.  '
 107           'Use 0 for "no limit".',
 108           ['--option-limit'],
 109           {'default': 14, 'metavar': '<level>',
 110            'validator': frontend.validate_nonnegative_int}),
 111          ('Format for footnote references: one of "superscript" or '
 112           '"brackets".  Default is "brackets".',
 113           ['--footnote-references'],
 114           {'choices': ['superscript', 'brackets'], 'default': 'brackets',
 115            'metavar': '<format>',
 116            'overrides': 'trim_footnote_reference_space'}),
 117          ('Format for block quote attributions: one of "dash" (em-dash '
 118           'prefix), "parentheses"/"parens", or "none".  Default is "dash".',
 119           ['--attribution'],
 120           {'choices': ['dash', 'parentheses', 'parens', 'none'],
 121            'default': 'dash', 'metavar': '<format>'}),
 122          ('Remove extra vertical whitespace between items of "simple" bullet '
 123           'lists and enumerated lists.  Default: enabled.',
 124           ['--compact-lists'],
 125           {'default': 1, 'action': 'store_true',
 126            'validator': frontend.validate_boolean}),
 127          ('Disable compact simple bullet and enumerated lists.',
 128           ['--no-compact-lists'],
 129           {'dest': 'compact_lists', 'action': 'store_false'}),
 130          ('Remove extra vertical whitespace between items of simple field '
 131           'lists.  Default: enabled.',
 132           ['--compact-field-lists'],
 133           {'default': 1, 'action': 'store_true',
 134            'validator': frontend.validate_boolean}),
 135          ('Disable compact simple field lists.',
 136           ['--no-compact-field-lists'],
 137           {'dest': 'compact_field_lists', 'action': 'store_false'}),
 138          ('Added to standard table classes. '
 139           'Defined styles: "borderless". Default: ""',
 140           ['--table-style'],
 141           {'default': ''}),
 142          ('Math output format, one of "MathML", "HTML", "MathJax" '
 143           'or "LaTeX". Default: "HTML math.css"',
 144           ['--math-output'],
 145           {'default': 'HTML math.css'}),
 146          ('Omit the XML declaration.  Use with caution.',
 147           ['--no-xml-declaration'],
 148           {'dest': 'xml_declaration', 'default': 1, 'action': 'store_false',
 149            'validator': frontend.validate_boolean}),
 150          ('Obfuscate email addresses to confuse harvesters while still '
 151           'keeping email links usable with standards-compliant browsers.',
 152           ['--cloak-email-addresses'],
 153           {'action': 'store_true', 'validator': frontend.validate_boolean}),))
 154
 155     settings_defaults = {'output_encoding_error_handler': 'xmlcharrefreplace'}
 156
 157     config_section = 'html4css1 writer'
 158     config_section_dependencies = ('writers',)
 159
 160     visitor_attributes = (
 161         'head_prefix', 'head', 'stylesheet', 'body_prefix',
 162         'body_pre_docinfo', 'docinfo', 'body', 'body_suffix',
 163         'title', 'subtitle', 'header', 'footer', 'meta', 'fragment',
 164         'html_prolog', 'html_head', 'html_title', 'html_subtitle',
 165         'html_body')
 166
 167     def get_transforms(self):
 168         return writers.Writer.get_transforms(self) + [writer_aux.Admonitions]
 169
 170     def __init__(self):
 171         writers.Writer.__init__(self)
 172         self.translator_class = HTMLTranslator
 173
 174     def translate(self):
 175         self.visitor = visitor = self.translator_class(self.document)
 176         self.document.walkabout(visitor)
 177         for attr in self.visitor_attributes:
 178             setattr(self, attr, getattr(visitor, attr))
 179         self.output = self.apply_template()
 180
 181     def apply_template(self):
 182         template_file = open(self.document.settings.template, 'rb')
 183         template = unicode(template_file.read(), 'utf-8')
 184         template_file.close()
 185         subs = self.interpolation_dict()
 186         return template % subs
 187
 188     def interpolation_dict(self):
 189         subs = {}
 190         settings = self.document.settings
 191         for attr in self.visitor_attributes:
 192             subs[attr] = ''.join(getattr(self, attr)).rstrip('\n')
 193         subs['encoding'] = settings.output_encoding
 194         subs['version'] = docutils.__version__
 195         return subs
 196
 197     def assemble_parts(self):
 198         writers.Writer.assemble_parts(self)
 199         for part in self.visitor_attributes:
 200             self.parts[part] = ''.join(getattr(self, part))
 201
 202
 203 class HTMLTranslator(nodes.NodeVisitor):
 204
 205     """
 206     This HTML writer has been optimized to produce visually compact
 207     lists (less vertical whitespace).  HTML's mixed content models
 208     allow list items to contain "<li><p>body elements</p></li>" or
 209     "<li>just text</li>" or even "<li>text<p>and body
 210     elements</p>combined</li>", each with different effects.  It would
 211     be best to stick with strict body elements in list items, but they
 212     affect vertical spacing in browsers (although they really
 213     shouldn't).
 214
 215     Here is an outline of the optimization:
 216
 217     - Check for and omit <p> tags in "simple" lists: list items
 218       contain either a single paragraph, a nested simple list, or a
 219       paragraph followed by a nested simple list.  This means that
 220       this list can be compact:
 221
 222           - Item 1.
 223           - Item 2.
 224
 225       But this list cannot be compact:
 226
 227           - Item 1.
 228
 229             This second paragraph forces space between list items.
 230
 231           - Item 2.
 232
 233     - In non-list contexts, omit <p> tags on a paragraph if that
 234       paragraph is the only child of its parent (footnotes & citations
 235       are allowed a label first).
 236
 237     - Regardless of the above, in definitions, table cells, field bodies,
 238       option descriptions, and list items, mark the first child with
 239       'class="first"' and the last child with 'class="last"'.  The stylesheet
 240       sets the margins (top & bottom respectively) to 0 for these elements.
 241
 242     The ``no_compact_lists`` setting (``--no-compact-lists`` command-line
 243     option) disables list whitespace optimization.
 244     """
 245
 246     xml_declaration = '<?xml version="1.0" encoding="%s" ?>\n'
 247     doctype = (
 248         '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'
 249         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">\n')
 250     doctype_mathml = doctype
 251
 252     head_prefix_template = ('<html xmlns="http://www.w3.org/1999/xhtml"'
 253                             ' xml:lang="%(lang)s" lang="%(lang)s">\n<head>\n')
 254     content_type = ('<meta http-equiv="Content-Type"'
 255                     ' content="text/html; charset=%s" />\n')
 256     content_type_mathml = ('<meta http-equiv="Content-Type"'
 257                            ' content="application/xhtml+xml; charset=%s" />\n')
 258
 259     generator = ('<meta name="generator" content="Docutils %s: '
 260                  'http://docutils.sourceforge.net/" />\n')
 261
 262     # Template for the MathJax script in the header:
 263     mathjax_script = '<script type="text/javascript" src="%s"></script>\n'
 264     # The latest version of MathJax from the distributed server:
 265     # avaliable to the public under the `MathJax CDN Terms of Service`__
 266     # __http://www.mathjax.org/download/mathjax-cdn-terms-of-service/
 267     mathjax_url = ('http://cdn.mathjax.org/mathjax/latest/MathJax.js?'
 268                    'config=TeX-AMS-MML_HTMLorMML')
 269     # may be overwritten by custom URL appended to "mathjax"
 270
 271     stylesheet_link = '<link rel="stylesheet" href="%s" type="text/css" />\n'
 272     embedded_stylesheet = '<style type="text/css">\n\n%s\n</style>\n'
 273     words_and_spaces = re.compile(r'\S+| +|\n')
 274     sollbruchstelle = re.compile(r'.+\W\W.+|[-?].+', re.U) # wrap point inside word
 275     lang_attribute = 'lang' # name changes to 'xml:lang' in XHTML 1.1
 276
 277     def __init__(self, document):
 278         nodes.NodeVisitor.__init__(self, document)
 279         self.settings = settings = document.settings
 280         lcode = settings.language_code
 281         self.language = languages.get_language(lcode, document.reporter)
 282         self.meta = [self.generator % docutils.__version__]
 283         self.head_prefix = []
 284         self.html_prolog = []
 285         if settings.xml_declaration:
 286             self.head_prefix.append(self.xml_declaration
 287                                     % settings.output_encoding)
 288             # encoding not interpolated:
 289             self.html_prolog.append(self.xml_declaration)
 290         self.head = self.meta[:]
 291         self.stylesheet = [self.stylesheet_call(path)
 292                            for path in utils.get_stylesheet_list(settings)]
 293         self.body_prefix = ['</head>\n<body>\n']
 294         # document title, subtitle display
 295         self.body_pre_docinfo = []
 296         # author, date, etc.
 297         self.docinfo = []
 298         self.body = []
 299         self.fragment = []
 300         self.body_suffix = ['</body>\n</html>\n']
 301         self.section_level = 0
 302         self.initial_header_level = int(settings.initial_header_level)
 303
 304         self.math_output = settings.math_output.split()
 305         self.math_output_options = self.math_output[1:]
 306         self.math_output = self.math_output[0].lower()
 307
 308         # A heterogenous stack used in conjunction with the tree traversal.
 309         # Make sure that the pops correspond to the pushes:
 310         self.context = []
 311         self.topic_classes = []
 312         self.colspecs = []
 313         self.compact_p = True
 314         self.compact_simple = False
 315         self.compact_field_list = False
 316         self.in_docinfo = False
 317         self.in_sidebar = False
 318         self.title = []
 319         self.subtitle = []
 320         self.header = []
 321         self.footer = []
 322         self.html_head = [self.content_type] # charset not interpolated
 323         self.html_title = []
 324         self.html_subtitle = []
 325         self.html_body = []
 326         self.in_document_title = 0   # len(self.body) or 0
 327         self.in_mailto = False
 328         self.author_in_authors = False
 329         self.math_header = []
 330
 331     def astext(self):
 332         return ''.join(self.head_prefix + self.head
 333                        + self.stylesheet + self.body_prefix
 334                        + self.body_pre_docinfo + self.docinfo
 335                        + self.body + self.body_suffix)
 336
 337     def encode(self, text):
 338         """Encode special characters in `text` & return."""
 339         # @@@ A codec to do these and all other HTML entities would be nice.
 340         text = unicode(text)
 341         return text.translate({
 342             ord('&'): u'&amp;',
 343             ord('<'): u'&lt;',
 344             ord('"'): u'&quot;',
 345             ord('>'): u'&gt;',
 346             ord('@'): u'&#64;', # may thwart some address harvesters
 347             # TODO: convert non-breaking space only if needed?
 348             0xa0: u'&nbsp;'}) # non-breaking space
 349
 350     def cloak_mailto(self, uri):
 351         """Try to hide a mailto: URL from harvesters."""
 352         # Encode "@" using a URL octet reference (see RFC 1738).
 353         # Further cloaking with HTML entities will be done in the
 354         # `attval` function.
 355         return uri.replace('@', '%40')
 356
 357     def cloak_email(self, addr):
 358         """Try to hide the link text of a email link from harversters."""
 359         # Surround at-signs and periods with <span> tags.  ("@" has
 360         # already been encoded to "&#64;" by the `encode` method.)
 361         addr = addr.replace('&#64;', '<span>&#64;</span>')
 362         addr = addr.replace('.', '<span>&#46;</span>')
 363         return addr
 364
 365     def attval(self, text,
 366                whitespace=re.compile('[\n\r\t\v\f]')):
 367         """Cleanse, HTML encode, and return attribute value text."""
 368         encoded = self.encode(whitespace.sub(' ', text))
 369         if self.in_mailto and self.settings.cloak_email_addresses:
 370             # Cloak at-signs ("%40") and periods with HTML entities.
 371             encoded = encoded.replace('%40', '&#37;&#52;&#48;')
 372             encoded = encoded.replace('.', '&#46;')
 373         return encoded
 374
 375     def stylesheet_call(self, path):
 376         """Return code to reference or embed stylesheet file `path`"""
 377         if self.settings.embed_stylesheet:
 378             try:
 379                 content = io.FileInput(source_path=path,
 380                                        encoding='utf-8').read()
 381                 self.settings.record_dependencies.add(path)
 382             except IOError, err:
 383                 msg = u"Cannot embed stylesheet '%s': %s." % (
 384                                 path, SafeString(err.strerror))
 385                 self.document.reporter.error(msg)
 386                 return '<--- %s --->\n' % msg
 387             return self.embedded_stylesheet % content
 388         # else link to style file:
 389         if self.settings.stylesheet_path:
 390             # adapt path relative to output (cf. config.html#stylesheet-path)
 391             path = utils.relative_path(self.settings._destination, path)
 392         return self.stylesheet_link % self.encode(path)
 393
 394     def starttag(self, node, tagname, suffix='\n', empty=False, **attributes):
 395         """
 396         Construct and return a start tag given a node (id & class attributes
 397         are extracted), tag name, and optional attributes.
 398         """
 399         tagname = tagname.lower()
 400         prefix = []
 401         atts = {}
 402         ids = []
 403         for (name, value) in attributes.items():
 404             atts[name.lower()] = value
 405         classes = node.get('classes', [])
 406         if 'class' in atts:
 407             classes.append(atts.pop('class'))
 408         # move language specification to 'lang' attribute
 409         languages = [cls for cls in classes
 410                      if cls.startswith('language-')]
 411         if languages:
 412             # attribute name is 'lang' in XHTML 1.0 but 'xml:lang' in 1.1
 413             atts[self.lang_attribute] = languages[0][9:]
 414             classes.pop(classes.index(languages[0]))
 415         classes = ' '.join(classes).strip()
 416         if classes:
 417             atts['class'] = classes
 418         assert 'id' not in atts
 419         ids.extend(node.get('ids', []))
 420         if 'ids' in atts:
 421             ids.extend(atts['ids'])
 422             del atts['ids']
 423         if ids:
 424             atts['id'] = ids[0]
 425             for id in ids[1:]:
 426                 # Add empty "span" elements for additional IDs.  Note
 427                 # that we cannot use empty "a" elements because there
 428                 # may be targets inside of references, but nested "a"
 429                 # elements aren't allowed in XHTML (even if they do
 430                 # not all have a "href" attribute).
 431                 if empty:
 432                     # Empty tag.  Insert target right in front of element.
 433                     prefix.append('<span id="%s"></span>' % id)
 434                 else:
 435                     # Non-empty tag.  Place the auxiliary <span> tag
 436                     # *inside* the element, as the first child.
 437                     suffix += '<span id="%s"></span>' % id
 438         attlist = atts.items()
 439         attlist.sort()
 440         parts = [tagname]
 441         for name, value in attlist:
 442             # value=None was used for boolean attributes without
 443             # value, but this isn't supported by XHTML.
 444             assert value is not None
 445             if isinstance(value, list):
 446                 values = [unicode(v) for v in value]
 447                 parts.append('%s="%s"' % (name.lower(),
 448                                           self.attval(' '.join(values))))
 449             else:
 450                 parts.append('%s="%s"' % (name.lower(),
 451                                           self.attval(unicode(value))))
 452         if empty:
 453             infix = ' /'
 454         else:
 455             infix = ''
 456         return ''.join(prefix) + '<%s%s>' % (' '.join(parts), infix) + suffix
 457
 458     def emptytag(self, node, tagname, suffix='\n', **attributes):
 459         """Construct and return an XML-compatible empty tag."""
 460         return self.starttag(node, tagname, suffix, empty=True, **attributes)
 461
 462     def set_class_on_child(self, node, class_, index=0):
 463         """
 464         Set class `class_` on the visible child no. index of `node`.
 465         Do nothing if node has fewer children than `index`.
 466         """
 467         children = [n for n in node if not isinstance(n, nodes.Invisible)]
 468         try:
 469             child = children[index]
 470         except IndexError:
 471             return
 472         child['classes'].append(class_)
 473
 474     def set_first_last(self, node):
 475         self.set_class_on_child(node, 'first', 0)
 476         self.set_class_on_child(node, 'last', -1)
 477
 478     def visit_Text(self, node):
 479         text = node.astext()
 480         encoded = self.encode(text)
 481         if self.in_mailto and self.settings.cloak_email_addresses:
 482             encoded = self.cloak_email(encoded)
 483         self.body.append(encoded)
 484
 485     def depart_Text(self, node):
 486         pass
 487
 488     def visit_abbreviation(self, node):
 489         # @@@ implementation incomplete ("title" attribute)
 490         self.body.append(self.starttag(node, 'abbr', ''))
 491
 492     def depart_abbreviation(self, node):
 493         self.body.append('</abbr>')
 494
 495     def visit_acronym(self, node):
 496         # @@@ implementation incomplete ("title" attribute)
 497         self.body.append(self.starttag(node, 'acronym', ''))
 498
 499     def depart_acronym(self, node):
 500         self.body.append('</acronym>')
 501
 502     def visit_address(self, node):
 503         self.visit_docinfo_item(node, 'address', meta=False)
 504         self.body.append(self.starttag(node, 'pre', CLASS='address'))
 505
 506     def depart_address(self, node):
 507         self.body.append('\n</pre>\n')
 508         self.depart_docinfo_item()
 509
 510     def visit_admonition(self, node):
 511         self.body.append(self.starttag(node, 'div'))
 512         self.set_first_last(node)
 513
 514     def depart_admonition(self, node=None):
 515         self.body.append('</div>\n')
 516
 517     attribution_formats = {'dash': ('&mdash;', ''),
 518                            'parentheses': ('(', ')'),
 519                            'parens': ('(', ')'),
 520                            'none': ('', '')}
 521
 522     def visit_attribution(self, node):
 523         prefix, suffix = self.attribution_formats[self.settings.attribution]
 524         self.context.append(suffix)
 525         self.body.append(
 526             self.starttag(node, 'p', prefix, CLASS='attribution'))
 527
 528     def depart_attribution(self, node):
 529         self.body.append(self.context.pop() + '</p>\n')
 530
 531     def visit_author(self, node):
 532         if isinstance(node.parent, nodes.authors):
 533             if self.author_in_authors:
 534                 self.body.append('\n<br />')
 535         else:
 536             self.visit_docinfo_item(node, 'author')
 537
 538     def depart_author(self, node):
 539         if isinstance(node.parent, nodes.authors):
 540             self.author_in_authors = True
 541         else:
 542             self.depart_docinfo_item()
 543
 544     def visit_authors(self, node):
 545         self.visit_docinfo_item(node, 'authors')
 546         self.author_in_authors = False  # initialize
 547
 548     def depart_authors(self, node):
 549         self.depart_docinfo_item()
 550
 551     def visit_block_quote(self, node):
 552         self.body.append(self.starttag(node, 'blockquote'))
 553
 554     def depart_block_quote(self, node):
 555         self.body.append('</blockquote>\n')
 556
 557     def check_simple_list(self, node):
 558         """Check for a simple list that can be rendered compactly."""
 559         visitor = SimpleListChecker(self.document)
 560         try:
 561             node.walk(visitor)
 562         except nodes.NodeFound:
 563             return None
 564         else:
 565             return 1
 566
 567     def is_compactable(self, node):
 568         return ('compact' in node['classes']
 569                 or (self.settings.compact_lists
 570                     and 'open' not in node['classes']
 571                     and (self.compact_simple
 572                          or self.topic_classes == ['contents']
 573                          or self.check_simple_list(node))))
 574
 575     def visit_bullet_list(self, node):
 576         atts = {}
 577         old_compact_simple = self.compact_simple
 578         self.context.append((self.compact_simple, self.compact_p))
 579         self.compact_p = None
 580         self.compact_simple = self.is_compactable(node)
 581         if self.compact_simple and not old_compact_simple:
 582             atts['class'] = 'simple'
 583         self.body.append(self.starttag(node, 'ul', **atts))
 584
 585     def depart_bullet_list(self, node):
 586         self.compact_simple, self.compact_p = self.context.pop()
 587         self.body.append('</ul>\n')
 588
 589     def visit_caption(self, node):
 590         self.body.append(self.starttag(node, 'p', '', CLASS='caption'))
 591
 592     def depart_caption(self, node):
 593         self.body.append('</p>\n')
 594
 595     def visit_citation(self, node):
 596         self.body.append(self.starttag(node, 'table',
 597                                        CLASS='docutils citation',
 598                                        frame="void", rules="none"))
 599         self.body.append('<colgroup><col class="label" /><col /></colgroup>\n'
 600                          '<tbody valign="top">\n'
 601                          '<tr>')
 602         self.footnote_backrefs(node)
 603
 604     def depart_citation(self, node):
 605         self.body.append('</td></tr>\n'
 606                          '</tbody>\n</table>\n')
 607
 608     def visit_citation_reference(self, node):
 609         href = '#'
 610         if 'refid' in node:
 611             href += node['refid']
 612         elif 'refname' in node:
 613             href += self.document.nameids[node['refname']]
 614         # else: # TODO system message (or already in the transform)?
 615         # 'Citation reference missing.'
 616         self.body.append(self.starttag(
 617             node, 'a', '[', CLASS='citation-reference', href=href))
 618
 619     def depart_citation_reference(self, node):
 620         self.body.append(']</a>')
 621
 622     def visit_classifier(self, node):
 623         self.body.append(' <span class="classifier-delimiter">:</span> ')
 624         self.body.append(self.starttag(node, 'span', '', CLASS='classifier'))
 625
 626     def depart_classifier(self, node):
 627         self.body.append('</span>')
 628
 629     def visit_colspec(self, node):
 630         self.colspecs.append(node)
 631         # "stubs" list is an attribute of the tgroup element:
 632         node.parent.stubs.append(node.attributes.get('stub'))
 633
 634     def depart_colspec(self, node):
 635         pass
 636
 637     def write_colspecs(self):
 638         width = 0
 639         for node in self.colspecs:
 640             width += node['colwidth']
 641         for node in self.colspecs:
 642             colwidth = int(node['colwidth'] * 100.0 / width + 0.5)
 643             self.body.append(self.emptytag(node, 'col',
 644                                            width='%i%%' % colwidth))
 645         self.colspecs = []
 646
 647     def visit_comment(self, node,
 648                       sub=re.compile('-(?=-)').sub):
 649         """Escape double-dashes in comment text."""
 650         self.body.append('<!-- %s -->\n' % sub('- ', node.astext()))
 651         # Content already processed:
 652         raise nodes.SkipNode
 653
 654     def visit_compound(self, node):
 655         self.body.append(self.starttag(node, 'div', CLASS='compound'))
 656         if len(node) > 1:
 657             node[0]['classes'].append('compound-first')
 658             node[-1]['classes'].append('compound-last')
 659             for child in node[1:-1]:
 660                 child['classes'].append('compound-middle')
 661
 662     def depart_compound(self, node):
 663         self.body.append('</div>\n')
 664
 665     def visit_container(self, node):
 666         self.body.append(self.starttag(node, 'div', CLASS='container'))
 667
 668     def depart_container(self, node):
 669         self.body.append('</div>\n')
 670
 671     def visit_contact(self, node):
 672         self.visit_docinfo_item(node, 'contact', meta=False)
 673
 674     def depart_contact(self, node):
 675         self.depart_docinfo_item()
 676
 677     def visit_copyright(self, node):
 678         self.visit_docinfo_item(node, 'copyright')
 679
 680     def depart_copyright(self, node):
 681         self.depart_docinfo_item()
 682
 683     def visit_date(self, node):
 684         self.visit_docinfo_item(node, 'date')
 685
 686     def depart_date(self, node):
 687         self.depart_docinfo_item()
 688
 689     def visit_decoration(self, node):
 690         pass
 691
 692     def depart_decoration(self, node):
 693         pass
 694
 695     def visit_definition(self, node):
 696         self.body.append('</dt>\n')
 697         self.body.append(self.starttag(node, 'dd', ''))
 698         self.set_first_last(node)
 699
 700     def depart_definition(self, node):
 701         self.body.append('</dd>\n')
 702
 703     def visit_definition_list(self, node):
 704         self.body.append(self.starttag(node, 'dl', CLASS='docutils'))
 705
 706     def depart_definition_list(self, node):
 707         self.body.append('</dl>\n')
 708
 709     def visit_definition_list_item(self, node):
 710         pass
 711
 712     def depart_definition_list_item(self, node):
 713         pass
 714
 715     def visit_description(self, node):
 716         self.body.append(self.starttag(node, 'td', ''))
 717         self.set_first_last(node)
 718
 719     def depart_description(self, node):
 720         self.body.append('</td>')
 721
 722     def visit_docinfo(self, node):
 723         self.context.append(len(self.body))
 724         self.body.append(self.starttag(node, 'table',
 725                                        CLASS='docinfo',
 726                                        frame="void", rules="none"))
 727         self.body.append('<col class="docinfo-name" />\n'
 728                          '<col class="docinfo-content" />\n'
 729                          '<tbody valign="top">\n')
 730         self.in_docinfo = True
 731
 732     def depart_docinfo(self, node):
 733         self.body.append('</tbody>\n</table>\n')
 734         self.in_docinfo = False
 735         start = self.context.pop()
 736         self.docinfo = self.body[start:]
 737         self.body = []
 738
 739     def visit_docinfo_item(self, node, name, meta=True):
 740         if meta:
 741             meta_tag = '<meta name="%s" content="%s" />\n' \
 742                        % (name, self.attval(node.astext()))
 743             self.add_meta(meta_tag)
 744         self.body.append(self.starttag(node, 'tr', ''))
 745         self.body.append('<th class="docinfo-name">%s:</th>\n<td>'
 746                          % self.language.labels[name])
 747         if len(node):
 748             if isinstance(node[0], nodes.Element):
 749                 node[0]['classes'].append('first')
 750             if isinstance(node[-1], nodes.Element):
 751                 node[-1]['classes'].append('last')
 752
 753     def depart_docinfo_item(self):
 754         self.body.append('</td></tr>\n')
 755
 756     def visit_doctest_block(self, node):
 757         self.body.append(self.starttag(node, 'pre', CLASS='doctest-block'))
 758
 759     def depart_doctest_block(self, node):
 760         self.body.append('\n</pre>\n')
 761
 762     def visit_document(self, node):
 763         self.head.append('<title>%s</title>\n'
 764                          % self.encode(node.get('title', '')))
 765
 766     def depart_document(self, node):
 767         self.head_prefix.extend([self.doctype,
 768                                  self.head_prefix_template %
 769                                  {'lang': self.settings.language_code}])
 770         self.html_prolog.append(self.doctype)
 771         self.meta.insert(0, self.content_type % self.settings.output_encoding)
 772         self.head.insert(0, self.content_type % self.settings.output_encoding)
 773         if self.math_header:
 774             if self.math_output == 'mathjax':
 775                 self.head.extend(self.math_header)
 776             else:
 777                 self.stylesheet.extend(self.math_header)
 778         # skip content-type meta tag with interpolated charset value:
 779         self.html_head.extend(self.head[1:])
 780         self.body_prefix.append(self.starttag(node, 'div', CLASS='document'))
 781         self.body_suffix.insert(0, '</div>\n')
 782         self.fragment.extend(self.body) # self.fragment is the "naked" body
 783         self.html_body.extend(self.body_prefix[1:] + self.body_pre_docinfo
 784                               + self.docinfo + self.body
 785                               + self.body_suffix[:-1])
 786         assert not self.context, 'len(context) = %s' % len(self.context)
 787
 788     def visit_emphasis(self, node):
 789         self.body.append(self.starttag(node, 'em', ''))
 790
 791     def depart_emphasis(self, node):
 792         self.body.append('</em>')
 793
 794     def visit_entry(self, node):
 795         atts = {'class': []}
 796         if isinstance(node.parent.parent, nodes.thead):
 797             atts['class'].append('head')
 798         if node.parent.parent.parent.stubs[node.parent.column]:
 799             # "stubs" list is an attribute of the tgroup element
 800             atts['class'].append('stub')
 801         if atts['class']:
 802             tagname = 'th'
 803             atts['class'] = ' '.join(atts['class'])
 804         else:
 805             tagname = 'td'
 806             del atts['class']
 807         node.parent.column += 1
 808         if 'morerows' in node:
 809             atts['rowspan'] = node['morerows'] + 1
 810         if 'morecols' in node:
 811             atts['colspan'] = node['morecols'] + 1
 812             node.parent.column += node['morecols']
 813         self.body.append(self.starttag(node, tagname, '', **atts))
 814         self.context.append('</%s>\n' % tagname.lower())
 815         if len(node) == 0:              # empty cell
 816             self.body.append('&nbsp;')
 817         self.set_first_last(node)
 818
 819     def depart_entry(self, node):
 820         self.body.append(self.context.pop())
 821
 822     def visit_enumerated_list(self, node):
 823         """
 824         The 'start' attribute does not conform to HTML 4.01's strict.dtd, but
 825         CSS1 doesn't help. CSS2 isn't widely enough supported yet to be
 826         usable.
 827         """
 828         atts = {}
 829         if 'start' in node:
 830             atts['start'] = node['start']
 831         if 'enumtype' in node:
 832             atts['class'] = node['enumtype']
 833         # @@@ To do: prefix, suffix. How? Change prefix/suffix to a
 834         # single "format" attribute? Use CSS2?
 835         old_compact_simple = self.compact_simple
 836         self.context.append((self.compact_simple, self.compact_p))
 837         self.compact_p = None
 838         self.compact_simple = self.is_compactable(node)
 839         if self.compact_simple and not old_compact_simple:
 840             atts['class'] = (atts.get('class', '') + ' simple').strip()
 841         self.body.append(self.starttag(node, 'ol', **atts))
 842
 843     def depart_enumerated_list(self, node):
 844         self.compact_simple, self.compact_p = self.context.pop()
 845         self.body.append('</ol>\n')
 846
 847     def visit_field(self, node):
 848         self.body.append(self.starttag(node, 'tr', '', CLASS='field'))
 849
 850     def depart_field(self, node):
 851         self.body.append('</tr>\n')
 852
 853     def visit_field_body(self, node):
 854         self.body.append(self.starttag(node, 'td', '', CLASS='field-body'))
 855         self.set_class_on_child(node, 'first', 0)
 856         field = node.parent
 857         if (self.compact_field_list or
 858             isinstance(field.parent, nodes.docinfo) or
 859             field.parent.index(field) == len(field.parent) - 1):
 860             # If we are in a compact list, the docinfo, or if this is
 861             # the last field of the field list, do not add vertical
 862             # space after last element.
 863             self.set_class_on_child(node, 'last', -1)
 864
 865     def depart_field_body(self, node):
 866         self.body.append('</td>\n')
 867
 868     def visit_field_list(self, node):
 869         self.context.append((self.compact_field_list, self.compact_p))
 870         self.compact_p = None
 871         if 'compact' in node['classes']:
 872             self.compact_field_list = True
 873         elif (self.settings.compact_field_lists
 874               and 'open' not in node['classes']):
 875             self.compact_field_list = True
 876         if self.compact_field_list:
 877             for field in node:
 878                 field_body = field[-1]
 879                 assert isinstance(field_body, nodes.field_body)
 880                 children = [n for n in field_body
 881                             if not isinstance(n, nodes.Invisible)]
 882                 if not (len(children) == 0 or
 883                         len(children) == 1 and
 884                         isinstance(children[0],
 885                                    (nodes.paragraph, nodes.line_block))):
 886                     self.compact_field_list = False
 887                     break
 888         self.body.append(self.starttag(node, 'table', frame='void',
 889                                        rules='none',
 890                                        CLASS='docutils field-list'))
 891         self.body.append('<col class="field-name" />\n'
 892                          '<col class="field-body" />\n'
 893                          '<tbody valign="top">\n')
 894
 895     def depart_field_list(self, node):
 896         self.body.append('</tbody>\n</table>\n')
 897         self.compact_field_list, self.compact_p = self.context.pop()
 898
 899     def visit_field_name(self, node):
 900         atts = {}
 901         if self.in_docinfo:
 902             atts['class'] = 'docinfo-name'
 903         else:
 904             atts['class'] = 'field-name'
 905         if ( self.settings.field_name_limit
 906              and len(node.astext()) > self.settings.field_name_limit):
 907             atts['colspan'] = 2
 908             self.context.append('</tr>\n'
 909                                 + self.starttag(node.parent, 'tr', '')
 910                                 + '<td>&nbsp;</td>')
 911         else:
 912             self.context.append('')
 913         self.body.append(self.starttag(node, 'th', '', **atts))
 914
 915     def depart_field_name(self, node):
 916         self.body.append(':</th>')
 917         self.body.append(self.context.pop())
 918
 919     def visit_figure(self, node):
 920         atts = {'class': 'figure'}
 921         if node.get('width'):
 922             atts['style'] = 'width: %s' % node['width']
 923         if node.get('align'):
 924             atts['class'] += " align-" + node['align']
 925         self.body.append(self.starttag(node, 'div', **atts))
 926
 927     def depart_figure(self, node):
 928         self.body.append('</div>\n')
 929
 930     def visit_footer(self, node):
 931         self.context.append(len(self.body))
 932
 933     def depart_footer(self, node):
 934         start = self.context.pop()
 935         footer = [self.starttag(node, 'div', CLASS='footer'),
 936                   '<hr class="footer" />\n']
 937         footer.extend(self.body[start:])
 938         footer.append('\n</div>\n')
 939         self.footer.extend(footer)
 940         self.body_suffix[:0] = footer
 941         del self.body[start:]
 942
 943     def visit_footnote(self, node):
 944         self.body.append(self.starttag(node, 'table',
 945                                        CLASS='docutils footnote',
 946                                        frame="void", rules="none"))
 947         self.body.append('<colgroup><col class="label" /><col /></colgroup>\n'
 948                          '<tbody valign="top">\n'
 949                          '<tr>')
 950         self.footnote_backrefs(node)
 951
 952     def footnote_backrefs(self, node):
 953         backlinks = []
 954         backrefs = node['backrefs']
 955         if self.settings.footnote_backlinks and backrefs:
 956             if len(backrefs) == 1:
 957                 self.context.append('')
 958                 self.context.append('</a>')
 959                 self.context.append('<a class="fn-backref" href="#%s">'
 960                                     % backrefs[0])
 961             else:
 962                 i = 1
 963                 for backref in backrefs:
 964                     backlinks.append('<a class="fn-backref" href="#%s">%s</a>'
 965                                      % (backref, i))
 966                     i += 1
 967                 self.context.append('<em>(%s)</em> ' % ', '.join(backlinks))
 968                 self.context += ['', '']
 969         else:
 970             self.context.append('')
 971             self.context += ['', '']
 972         # If the node does not only consist of a label.
 973         if len(node) > 1:
 974             # If there are preceding backlinks, we do not set class
 975             # 'first', because we need to retain the top-margin.
 976             if not backlinks:
 977                 node[1]['classes'].append('first')
 978             node[-1]['classes'].append('last')
 979
 980     def depart_footnote(self, node):
 981         self.body.append('</td></tr>\n'
 982                          '</tbody>\n</table>\n')
 983
 984     def visit_footnote_reference(self, node):
 985         href = '#' + node['refid']
 986         format = self.settings.footnote_references
 987         if format == 'brackets':
 988             suffix = '['
 989             self.context.append(']')
 990         else:
 991             assert format == 'superscript'
 992             suffix = '<sup>'
 993             self.context.append('</sup>')
 994         self.body.append(self.starttag(node, 'a', suffix,
 995                                        CLASS='footnote-reference', href=href))
 996
 997     def depart_footnote_reference(self, node):
 998         self.body.append(self.context.pop() + '</a>')
 999
1000     def visit_generated(self, node):
1001         pass
1002
1003     def depart_generated(self, node):
1004         pass
1005
1006     def visit_header(self, node):
1007         self.context.append(len(self.body))
1008
1009     def depart_header(self, node):
1010         start = self.context.pop()
1011         header = [self.starttag(node, 'div', CLASS='header')]
1012         header.extend(self.body[start:])
1013         header.append('\n<hr class="header"/>\n</div>\n')
1014         self.body_prefix.extend(header)
1015         self.header.extend(header)
1016         del self.body[start:]
1017
1018     def visit_image(self, node):
1019         atts = {}
1020         uri = node['uri']
1021         # place SVG and SWF images in an <object> element
1022         types = {'.svg': 'image/svg+xml',
1023                  '.swf': 'application/x-shockwave-flash'}
1024         ext = os.path.splitext(uri)[1].lower()
1025         if ext in ('.svg', '.swf'):
1026             atts['data'] = uri
1027             atts['type'] = types[ext]
1028         else:
1029             atts['src'] = uri
1030             atts['alt'] = node.get('alt', uri)
1031         # image size
1032         if 'width' in node:
1033             atts['width'] = node['width']
1034         if 'height' in node:
1035             atts['height'] = node['height']
1036         if 'scale' in node:
1037             if (PIL and not ('width' in node and 'height' in node)
1038                 and self.settings.file_insertion_enabled):
1039                 imagepath = urllib.url2pathname(uri)
1040                 try:
1041                     img = PIL.Image.open(
1042                             imagepath.encode(sys.getfilesystemencoding()))
1043                 except (IOError, UnicodeEncodeError):
1044                     pass # TODO: warn?
1045                 else:
1046                     self.settings.record_dependencies.add(
1047                         imagepath.replace('\\', '/'))
1048                     if 'width' not in atts:
1049                         atts['width'] = str(img.size[0])
1050                     if 'height' not in atts:
1051                         atts['height'] = str(img.size[1])
1052                     del img
1053             for att_name in 'width', 'height':
1054                 if att_name in atts:
1055                     match = re.match(r'([0-9.]+)(\S*)$', atts[att_name])
1056                     assert match
1057                     atts[att_name] = '%s%s' % (
1058                         float(match.group(1)) * (float(node['scale']) / 100),
1059                         match.group(2))
1060         style = []
1061         for att_name in 'width', 'height':
1062             if att_name in atts:
1063                 if re.match(r'^[0-9.]+$', atts[att_name]):
1064                     # Interpret unitless values as pixels.
1065                     atts[att_name] += 'px'
1066                 style.append('%s: %s;' % (att_name, atts[att_name]))
1067                 del atts[att_name]
1068         if style:
1069             atts['style'] = ' '.join(style)
1070         if (isinstance(node.parent, nodes.TextElement) or
1071             (isinstance(node.parent, nodes.reference) and
1072              not isinstance(node.parent.parent, nodes.TextElement))):
1073             # Inline context or surrounded by <a>...</a>.
1074             suffix = ''
1075         else:
1076             suffix = '\n'
1077         if 'align' in node:
1078             atts['class'] = 'align-%s' % node['align']
1079         self.context.append('')
1080         if ext in ('.svg', '.swf'): # place in an object element,
1081             # do NOT use an empty tag: incorrect rendering in browsers
1082             self.body.append(self.starttag(node, 'object', suffix, **atts) +
1083                              node.get('alt', uri) + '</object>' + suffix)
1084         else:
1085             self.body.append(self.emptytag(node, 'img', suffix, **atts))
1086
1087     def depart_image(self, node):
1088         self.body.append(self.context.pop())
1089
1090     def visit_inline(self, node):
1091         self.body.append(self.starttag(node, 'span', ''))
1092
1093     def depart_inline(self, node):
1094         self.body.append('</span>')
1095
1096     def visit_label(self, node):
1097         # Context added in footnote_backrefs.
1098         self.body.append(self.starttag(node, 'td', '%s[' % self.context.pop(),
1099                                        CLASS='label'))
1100
1101     def depart_label(self, node):
1102         # Context added in footnote_backrefs.
1103         self.body.append(']%s</td><td>%s' % (self.context.pop(), self.context.pop()))
1104
1105     def visit_legend(self, node):
1106         self.body.append(self.starttag(node, 'div', CLASS='legend'))
1107
1108     def depart_legend(self, node):
1109         self.body.append('</div>\n')
1110
1111     def visit_line(self, node):
1112         self.body.append(self.starttag(node, 'div', suffix='', CLASS='line'))
1113         if not len(node):
1114             self.body.append('<br />')
1115
1116     def depart_line(self, node):
1117         self.body.append('</div>\n')
1118
1119     def visit_line_block(self, node):
1120         self.body.append(self.starttag(node, 'div', CLASS='line-block'))
1121
1122     def depart_line_block(self, node):
1123         self.body.append('</div>\n')
1124
1125     def visit_list_item(self, node):
1126         self.body.append(self.starttag(node, 'li', ''))
1127         if len(node):
1128             node[0]['classes'].append('first')
1129
1130     def depart_list_item(self, node):
1131         self.body.append('</li>\n')
1132
1133     def visit_literal(self, node):
1134         # special case: "code" role
1135         classes = node.get('classes', [])
1136         if 'code' in classes:
1137             # filter 'code' from class arguments
1138             node['classes'] = [cls for cls in classes if cls != 'code']
1139             self.body.append(self.starttag(node, 'code', ''))
1140             return
1141         self.body.append(
1142             self.starttag(node, 'tt', '', CLASS='docutils literal'))
1143         text = node.astext()
1144         for token in self.words_and_spaces.findall(text):
1145             if token.strip():
1146                 # Protect text like "--an-option" and the regular expression
1147                 # ``[+]?(\d+(\.\d*)?|\.\d+)`` from bad line wrapping
1148                 if self.sollbruchstelle.search(token):
1149                     self.body.append('<span class="pre">%s</span>'
1150                                      % self.encode(token))
1151                 else:
1152                     self.body.append(self.encode(token))
1153             elif token in ('\n', ' '):
1154                 # Allow breaks at whitespace:
1155                 self.body.append(token)
1156             else:
1157                 # Protect runs of multiple spaces; the last space can wrap:
1158                 self.body.append('&nbsp;' * (len(token) - 1) + ' ')
1159         self.body.append('</tt>')
1160         # Content already processed:
1161         raise nodes.SkipNode
1162
1163     def depart_literal(self, node):
1164         # skipped unless literal element is from "code" role:
1165         self.body.append('</code>')
1166
1167     def visit_literal_block(self, node):
1168         self.body.append(self.starttag(node, 'pre', CLASS='literal-block'))
1169
1170     def depart_literal_block(self, node):
1171         self.body.append('\n</pre>\n')
1172
1173     def visit_math(self, node, math_env=''):
1174         # If the method is called from visit_math_block(), math_env != ''.
1175
1176         # As there is no native HTML math support, we provide alternatives:
1177         # LaTeX and MathJax math_output modes simply wrap the content,
1178         # HTML and MathML math_output modes also convert the math_code.
1179         if self.math_output not in ('mathml', 'html', 'mathjax', 'latex'):
1180             self.document.reporter.error(
1181                 'math-output format "%s" not supported '
1182                 'falling back to "latex"'% self.math_output)
1183             self.math_output = 'latex'
1184         #
1185         # HTML container
1186         tags = {# math_output: (block, inline, class-arguments)
1187                 'mathml':      ('div', '', ''),
1188                 'html':        ('div', 'span', 'formula'),
1189                 'mathjax':     ('div', 'span', 'math'),
1190                 'latex':       ('pre', 'tt',   'math'),
1191                }
1192         tag = tags[self.math_output][math_env == '']
1193         clsarg = tags[self.math_output][2]
1194         # LaTeX container
1195         wrappers = {# math_mode: (inline, block)
1196                     'mathml':  (None,     None),
1197                     'html':    ('$%s$',   u'\\begin{%s}\n%s\n\\end{%s}'),
1198                     'mathjax': ('\(%s\)', u'\\begin{%s}\n%s\n\\end{%s}'),
1199                     'latex':   (None,     None),
1200                    }
1201         wrapper = wrappers[self.math_output][math_env != '']
1202         # get and wrap content
1203         math_code = node.astext().translate(unichar2tex.uni2tex_table)
1204         if wrapper and math_env:
1205             math_code = wrapper % (math_env, math_code, math_env)
1206         elif wrapper:
1207             math_code = wrapper % math_code
1208         # settings and conversion
1209         if self.math_output in ('latex', 'mathjax'):
1210             math_code = self.encode(math_code)
1211         if self.math_output == 'mathjax' and not self.math_header:
1212             if self.math_output_options:
1213                 self.mathjax_url = self.math_output_options[0]
1214             self.math_header = [self.mathjax_script % self.mathjax_url]
1215         elif self.math_output == 'html':
1216             if self.math_output_options and not self.math_header:
1217                 self.math_header = [self.stylesheet_call(
1218                     utils.find_file_in_dirs(s, self.settings.stylesheet_dirs))
1219                     for s in self.math_output_options[0].split(',')]
1220             # TODO: fix display mode in matrices and fractions
1221             math2html.DocumentParameters.displaymode = (math_env != '')
1222             math_code = math2html.math2html(math_code)
1223         elif self.math_output == 'mathml':
1224             self.doctype = self.doctype_mathml
1225             self.content_type = self.content_type_mathml
1226             try:
1227                 mathml_tree = parse_latex_math(math_code, inline=not(math_env))
1228                 math_code = ''.join(mathml_tree.xml())
1229             except SyntaxError, err:
1230                 err_node = self.document.reporter.error(err, base_node=node)
1231                 self.visit_system_message(err_node)
1232                 self.body.append(self.starttag(node, 'p'))
1233                 self.body.append(u','.join(err.args))
1234                 self.body.append('</p>\n')
1235                 self.body.append(self.starttag(node, 'pre',
1236                                                CLASS='literal-block'))
1237                 self.body.append(self.encode(math_code))
1238                 self.body.append('\n</pre>\n')
1239                 self.depart_system_message(err_node)
1240                 raise nodes.SkipNode
1241         # append to document body
1242         if tag:
1243             self.body.append(self.starttag(node, tag,
1244                                            suffix='\n'*bool(math_env),
1245                                            CLASS=clsarg))
1246         self.body.append(math_code)
1247         if math_env:
1248             self.body.append('\n')
1249         if tag:
1250             self.body.append('</%s>\n' % tag)
1251         # Content already processed:
1252         raise nodes.SkipNode
1253
1254     def depart_math(self, node):
1255         pass # never reached
1256
1257     def visit_math_block(self, node):
1258         # print node.astext().encode('utf8')
1259         math_env = pick_math_environment(node.astext())
1260         self.visit_math(node, math_env=math_env)
1261
1262     def depart_math_block(self, node):
1263         pass # never reached
1264
1265     def visit_meta(self, node):
1266         meta = self.emptytag(node, 'meta', **node.non_default_attributes())
1267         self.add_meta(meta)
1268
1269     def depart_meta(self, node):
1270         pass
1271
1272     def add_meta(self, tag):
1273         self.meta.append(tag)
1274         self.head.append(tag)
1275
1276     def visit_option(self, node):
1277         if self.context[-1]:
1278             self.body.append(', ')
1279         self.body.append(self.starttag(node, 'span', '', CLASS='option'))
1280
1281     def depart_option(self, node):
1282         self.body.append('</span>')
1283         self.context[-1] += 1
1284
1285     def visit_option_argument(self, node):
1286         self.body.append(node.get('delimiter', ' '))
1287         self.body.append(self.starttag(node, 'var', ''))
1288
1289     def depart_option_argument(self, node):
1290         self.body.append('</var>')
1291
1292     def visit_option_group(self, node):
1293         atts = {}
1294         if ( self.settings.option_limit
1295              and len(node.astext()) > self.settings.option_limit):
1296             atts['colspan'] = 2
1297             self.context.append('</tr>\n<tr><td>&nbsp;</td>')
1298         else:
1299             self.context.append('')
1300         self.body.append(
1301             self.starttag(node, 'td', CLASS='option-group', **atts))
1302         self.body.append('<kbd>')
1303         self.context.append(0)          # count number of options
1304
1305     def depart_option_group(self, node):
1306         self.context.pop()
1307         self.body.append('</kbd></td>\n')
1308         self.body.append(self.context.pop())
1309
1310     def visit_option_list(self, node):
1311         self.body.append(
1312               self.starttag(node, 'table', CLASS='docutils option-list',
1313                             frame="void", rules="none"))
1314         self.body.append('<col class="option" />\n'
1315                          '<col class="description" />\n'
1316                          '<tbody valign="top">\n')
1317
1318     def depart_option_list(self, node):
1319         self.body.append('</tbody>\n</table>\n')
1320
1321     def visit_option_list_item(self, node):
1322         self.body.append(self.starttag(node, 'tr', ''))
1323
1324     def depart_option_list_item(self, node):
1325         self.body.append('</tr>\n')
1326
1327     def visit_option_string(self, node):
1328         pass
1329
1330     def depart_option_string(self, node):
1331         pass
1332
1333     def visit_organization(self, node):
1334         self.visit_docinfo_item(node, 'organization')
1335
1336     def depart_organization(self, node):
1337         self.depart_docinfo_item()
1338
1339     def should_be_compact_paragraph(self, node):
1340         """
1341         Determine if the <p> tags around paragraph ``node`` can be omitted.
1342         """
1343         if (isinstance(node.parent, nodes.document) or
1344             isinstance(node.parent, nodes.compound)):
1345             # Never compact paragraphs in document or compound.
1346             return False
1347         for key, value in node.attlist():
1348             if (node.is_not_default(key) and
1349                 not (key == 'classes' and value in
1350                      ([], ['first'], ['last'], ['first', 'last']))):
1351                 # Attribute which needs to survive.
1352                 return False
1353         first = isinstance(node.parent[0], nodes.label) # skip label
1354         for child in node.parent.children[first:]:
1355             # only first paragraph can be compact
1356             if isinstance(child, nodes.Invisible):
1357                 continue
1358             if child is node:
1359                 break
1360             return False
1361         parent_length = len([n for n in node.parent if not isinstance(
1362             n, (nodes.Invisible, nodes.label))])
1363         if ( self.compact_simple
1364              or self.compact_field_list
1365              or self.compact_p and parent_length == 1):
1366             return True
1367         return False
1368
1369     def visit_paragraph(self, node):
1370         if self.should_be_compact_paragraph(node):
1371             self.context.append('')
1372         else:
1373             self.body.append(self.starttag(node, 'p', ''))
1374             self.context.append('</p>\n')
1375
1376     def depart_paragraph(self, node):
1377         self.body.append(self.context.pop())
1378
1379     def visit_problematic(self, node):
1380         if node.hasattr('refid'):
1381             self.body.append('<a href="#%s">' % node['refid'])
1382             self.context.append('</a>')
1383         else:
1384             self.context.append('')
1385         self.body.append(self.starttag(node, 'span', '', CLASS='problematic'))
1386
1387     def depart_problematic(self, node):
1388         self.body.append('</span>')
1389         self.body.append(self.context.pop())
1390
1391     def visit_raw(self, node):
1392         if 'html' in node.get('format', '').split():
1393             t = isinstance(node.parent, nodes.TextElement) and 'span' or 'div'
1394             if node['classes']:
1395                 self.body.append(self.starttag(node, t, suffix=''))
1396             self.body.append(node.astext())
1397             if node['classes']:
1398                 self.body.append('</%s>' % t)
1399         # Keep non-HTML raw text out of output:
1400         raise nodes.SkipNode
1401
1402     def visit_reference(self, node):
1403         atts = {'class': 'reference'}
1404         if 'refuri' in node:
1405             atts['href'] = node['refuri']
1406             if ( self.settings.cloak_email_addresses
1407                  and atts['href'].startswith('mailto:')):
1408                 atts['href'] = self.cloak_mailto(atts['href'])
1409                 self.in_mailto = True
1410             atts['class'] += ' external'
1411         else:
1412             assert 'refid' in node, \
1413                    'References must have "refuri" or "refid" attribute.'
1414             atts['href'] = '#' + node['refid']
1415             atts['class'] += ' internal'
1416         if not isinstance(node.parent, nodes.TextElement):
1417             assert len(node) == 1 and isinstance(node[0], nodes.image)
1418             atts['class'] += ' image-reference'
1419         self.body.append(self.starttag(node, 'a', '', **atts))
1420
1421     def depart_reference(self, node):
1422         self.body.append('</a>')
1423         if not isinstance(node.parent, nodes.TextElement):
1424             self.body.append('\n')
1425         self.in_mailto = False
1426
1427     def visit_revision(self, node):
1428         self.visit_docinfo_item(node, 'revision', meta=False)
1429
1430     def depart_revision(self, node):
1431         self.depart_docinfo_item()
1432
1433     def visit_row(self, node):
1434         self.body.append(self.starttag(node, 'tr', ''))
1435         node.column = 0
1436
1437     def depart_row(self, node):
1438         self.body.append('</tr>\n')
1439
1440     def visit_rubric(self, node):
1441         self.body.append(self.starttag(node, 'p', '', CLASS='rubric'))
1442
1443     def depart_rubric(self, node):
1444         self.body.append('</p>\n')
1445
1446     def visit_section(self, node):
1447         self.section_level += 1
1448         self.body.append(
1449             self.starttag(node, 'div', CLASS='section'))
1450
1451     def depart_section(self, node):
1452         self.section_level -= 1
1453         self.body.append('</div>\n')
1454
1455     def visit_sidebar(self, node):
1456         self.body.append(
1457             self.starttag(node, 'div', CLASS='sidebar'))
1458         self.set_first_last(node)
1459         self.in_sidebar = True
1460
1461     def depart_sidebar(self, node):
1462         self.body.append('</div>\n')
1463         self.in_sidebar = False
1464
1465     def visit_status(self, node):
1466         self.visit_docinfo_item(node, 'status', meta=False)
1467
1468     def depart_status(self, node):
1469         self.depart_docinfo_item()
1470
1471     def visit_strong(self, node):
1472         self.body.append(self.starttag(node, 'strong', ''))
1473
1474     def depart_strong(self, node):
1475         self.body.append('</strong>')
1476
1477     def visit_subscript(self, node):
1478         self.body.append(self.starttag(node, 'sub', ''))
1479
1480     def depart_subscript(self, node):
1481         self.body.append('</sub>')
1482
1483     def visit_substitution_definition(self, node):
1484         """Internal only."""
1485         raise nodes.SkipNode
1486
1487     def visit_substitution_reference(self, node):
1488         self.unimplemented_visit(node)
1489
1490     def visit_subtitle(self, node):
1491         if isinstance(node.parent, nodes.sidebar):
1492             self.body.append(self.starttag(node, 'p', '',
1493                                            CLASS='sidebar-subtitle'))
1494             self.context.append('</p>\n')
1495         elif isinstance(node.parent, nodes.document):
1496             self.body.append(self.starttag(node, 'h2', '', CLASS='subtitle'))
1497             self.context.append('</h2>\n')
1498             self.in_document_title = len(self.body)
1499         elif isinstance(node.parent, nodes.section):
1500             tag = 'h%s' % (self.section_level + self.initial_header_level - 1)
1501             self.body.append(
1502                 self.starttag(node, tag, '', CLASS='section-subtitle') +
1503                 self.starttag({}, 'span', '', CLASS='section-subtitle'))
1504             self.context.append('</span></%s>\n' % tag)
1505
1506     def depart_subtitle(self, node):
1507         self.body.append(self.context.pop())
1508         if self.in_document_title:
1509             self.subtitle = self.body[self.in_document_title:-1]
1510             self.in_document_title = 0
1511             self.body_pre_docinfo.extend(self.body)
1512             self.html_subtitle.extend(self.body)
1513             del self.body[:]
1514
1515     def visit_superscript(self, node):
1516         self.body.append(self.starttag(node, 'sup', ''))
1517
1518     def depart_superscript(self, node):
1519         self.body.append('</sup>')
1520
1521     def visit_system_message(self, node):
1522         self.body.append(self.starttag(node, 'div', CLASS='system-message'))
1523         self.body.append('<p class="system-message-title">')
1524         backref_text = ''
1525         if len(node['backrefs']):
1526             backrefs = node['backrefs']
1527             if len(backrefs) == 1:
1528                 backref_text = ('; <em><a href="#%s">backlink</a></em>'
1529                                 % backrefs[0])
1530             else:
1531                 i = 1
1532                 backlinks = []
1533                 for backref in backrefs:
1534                     backlinks.append('<a href="#%s">%s</a>' % (backref, i))
1535                     i += 1
1536                 backref_text = ('; <em>backlinks: %s</em>'
1537                                 % ', '.join(backlinks))
1538         if node.hasattr('line'):
1539             line = ', line %s' % node['line']
1540         else:
1541             line = ''
1542         self.body.append('System Message: %s/%s '
1543                          '(<tt class="docutils">%s</tt>%s)%s</p>\n'
1544                          % (node['type'], node['level'],
1545                             self.encode(node['source']), line, backref_text))
1546
1547     def depart_system_message(self, node):
1548         self.body.append('</div>\n')
1549
1550     def visit_table(self, node):
1551         self.context.append(self.compact_p)
1552         self.compact_p = True
1553         classes = ' '.join(['docutils', self.settings.table_style]).strip()
1554         self.body.append(
1555             self.starttag(node, 'table', CLASS=classes, border="1"))
1556
1557     def depart_table(self, node):
1558         self.compact_p = self.context.pop()
1559         self.body.append('</table>\n')
1560
1561     def visit_target(self, node):
1562         if not ('refuri' in node or 'refid' in node
1563                 or 'refname' in node):
1564             self.body.append(self.starttag(node, 'span', '', CLASS='target'))
1565             self.context.append('</span>')
1566         else:
1567             self.context.append('')
1568
1569     def depart_target(self, node):
1570         self.body.append(self.context.pop())
1571
1572     def visit_tbody(self, node):
1573         self.write_colspecs()
1574         self.body.append(self.context.pop()) # '</colgroup>\n' or ''
1575         self.body.append(self.starttag(node, 'tbody', valign='top'))
1576
1577     def depart_tbody(self, node):
1578         self.body.append('</tbody>\n')
1579
1580     def visit_term(self, node):
1581         self.body.append(self.starttag(node, 'dt', ''))
1582
1583     def depart_term(self, node):
1584         """
1585         Leave the end tag to `self.visit_definition()`, in case there's a
1586         classifier.
1587         """
1588         pass
1589
1590     def visit_tgroup(self, node):
1591         # Mozilla needs <colgroup>:
1592         self.body.append(self.starttag(node, 'colgroup'))
1593         # Appended by thead or tbody:
1594         self.context.append('</colgroup>\n')
1595         node.stubs = []
1596
1597     def depart_tgroup(self, node):
1598         pass
1599
1600     def visit_thead(self, node):
1601         self.write_colspecs()
1602         self.body.append(self.context.pop()) # '</colgroup>\n'
1603         # There may or may not be a <thead>; this is for <tbody> to use:
1604         self.context.append('')
1605         self.body.append(self.starttag(node, 'thead', valign='bottom'))
1606
1607     def depart_thead(self, node):
1608         self.body.append('</thead>\n')
1609
1610     def visit_title(self, node):
1611         """Only 6 section levels are supported by HTML."""
1612         check_id = 0  # TODO: is this a bool (False) or a counter?
1613         close_tag = '</p>\n'
1614         if isinstance(node.parent, nodes.topic):
1615             self.body.append(
1616                   self.starttag(node, 'p', '', CLASS='topic-title first'))
1617         elif isinstance(node.parent, nodes.sidebar):
1618             self.body.append(
1619                   self.starttag(node, 'p', '', CLASS='sidebar-title'))
1620         elif isinstance(node.parent, nodes.Admonition):
1621             self.body.append(
1622                   self.starttag(node, 'p', '', CLASS='admonition-title'))
1623         elif isinstance(node.parent, nodes.table):
1624             self.body.append(
1625                   self.starttag(node, 'caption', ''))
1626             close_tag = '</caption>\n'
1627         elif isinstance(node.parent, nodes.document):
1628             self.body.append(self.starttag(node, 'h1', '', CLASS='title'))
1629             close_tag = '</h1>\n'
1630             self.in_document_title = len(self.body)
1631         else:
1632             assert isinstance(node.parent, nodes.section)
1633             h_level = self.section_level + self.initial_header_level - 1
1634             atts = {}
1635             if (len(node.parent) >= 2 and
1636                 isinstance(node.parent[1], nodes.subtitle)):
1637                 atts['CLASS'] = 'with-subtitle'
1638             self.body.append(
1639                   self.starttag(node, 'h%s' % h_level, '', **atts))
1640             atts = {}
1641             if node.hasattr('refid'):
1642                 atts['class'] = 'toc-backref'
1643                 atts['href'] = '#' + node['refid']
1644             if atts:
1645                 self.body.append(self.starttag({}, 'a', '', **atts))
1646                 close_tag = '</a></h%s>\n' % (h_level)
1647             else:
1648                 close_tag = '</h%s>\n' % (h_level)
1649         self.context.append(close_tag)
1650
1651     def depart_title(self, node):
1652         self.body.append(self.context.pop())
1653         if self.in_document_title:
1654             self.title = self.body[self.in_document_title:-1]
1655             self.in_document_title = 0
1656             self.body_pre_docinfo.extend(self.body)
1657             self.html_title.extend(self.body)
1658             del self.body[:]
1659
1660     def visit_title_reference(self, node):
1661         self.body.append(self.starttag(node, 'cite', ''))
1662
1663     def depart_title_reference(self, node):
1664         self.body.append('</cite>')
1665
1666     def visit_topic(self, node):
1667         self.body.append(self.starttag(node, 'div', CLASS='topic'))
1668         self.topic_classes = node['classes']
1669
1670     def depart_topic(self, node):
1671         self.body.append('</div>\n')
1672         self.topic_classes = []
1673
1674     def visit_transition(self, node):
1675         self.body.append(self.emptytag(node, 'hr', CLASS='docutils'))
1676
1677     def depart_transition(self, node):
1678         pass
1679
1680     def visit_version(self, node):
1681         self.visit_docinfo_item(node, 'version', meta=False)
1682
1683     def depart_version(self, node):
1684         self.depart_docinfo_item()
1685
1686     def unimplemented_visit(self, node):
1687         raise NotImplementedError('visiting unimplemented node type: %s'
1688                                   % node.__class__.__name__)
1689
1690
1691 class SimpleListChecker(nodes.GenericNodeVisitor):
1692
1693     """
1694     Raise `nodes.NodeFound` if non-simple list item is encountered.
1695
1696     Here "simple" means a list item containing nothing other than a single
1697     paragraph, a simple list, or a paragraph followed by a simple list.
1698     """
1699
1700     def default_visit(self, node):
1701         raise nodes.NodeFound
1702
1703     def visit_bullet_list(self, node):
1704         pass
1705
1706     def visit_enumerated_list(self, node):
1707         pass
1708
1709     def visit_list_item(self, node):
1710         children = []
1711         for child in node.children:
1712             if not isinstance(child, nodes.Invisible):
1713                 children.append(child)
1714         if (children and isinstance(children[0], nodes.paragraph)
1715             and (isinstance(children[-1], nodes.bullet_list)
1716                  or isinstance(children[-1], nodes.enumerated_list))):
1717             children.pop()
1718         if len(children) <= 1:
1719             return
1720         else:
1721             raise nodes.NodeFound
1722
1723     def visit_paragraph(self, node):
1724         raise nodes.SkipNode
1725
1726     def invisible_visit(self, node):
1727         """Invisible nodes should be ignored."""
1728         raise nodes.SkipNode
1729
1730     visit_comment = invisible_visit
1731     visit_substitution_definition = invisible_visit
1732     visit_target = invisible_visit
1733     visit_pending = invisible_visit