cmlenz@336: # -*- coding: utf-8 -*-
cmlenz@336: #
cmlenz@408: # Copyright (C) 2006-2007 Edgewall Software
cmlenz@336: # All rights reserved.
cmlenz@336: #
cmlenz@336: # This software is licensed as described in the file COPYING, which
cmlenz@336: # you should have received as part of this distribution. The terms
cmlenz@336: # are also available at http://genshi.edgewall.org/wiki/License.
cmlenz@336: #
cmlenz@336: # This software consists of voluntary contributions made by many
cmlenz@336: # individuals. For the exact contribution history, see the revision
cmlenz@336: # history and logs, available at http://genshi.edgewall.org/log/.
cmlenz@336: 
cmlenz@336: """Markup templating engine."""
cmlenz@336: 
cmlenz@336: from itertools import chain
cmlenz@336: 
cmlenz@636: from genshi.core import Attrs, Markup, Namespace, Stream, StreamEventKind
cmlenz@405: from genshi.core import START, END, START_NS, END_NS, TEXT, PI, COMMENT
cmlenz@336: from genshi.input import XMLParser
cmlenz@400: from genshi.template.base import BadDirectiveError, Template, \
cmlenz@475:                                  TemplateSyntaxError, _apply_directives, \
cmlenz@609:                                  EXEC, INCLUDE, SUB
cmlenz@405: from genshi.template.eval import Suite
cmlenz@407: from genshi.template.interpolation import interpolate
cmlenz@336: from genshi.template.directives import *
cmlenz@610: from genshi.template.text import NewTextTemplate
cmlenz@336: 
cmlenz@425: __all__ = ['MarkupTemplate']
cmlenz@425: __docformat__ = 'restructuredtext en'
cmlenz@425: 
cmlenz@336: 
cmlenz@336: class MarkupTemplate(Template):
cmlenz@336:     """Implementation of the template language for XML-based templates.
cmlenz@336:     
cmlenz@336:     >>> tmpl = MarkupTemplate('''<ul xmlns:py="http://genshi.edgewall.org/">
cmlenz@336:     ...   <li py:for="item in items">${item}</li>
cmlenz@336:     ... </ul>''')
cmlenz@336:     >>> print tmpl.generate(items=[1, 2, 3])
cmlenz@336:     <ul>
cmlenz@336:       <li>1</li><li>2</li><li>3</li>
cmlenz@336:     </ul>
cmlenz@336:     """
cmlenz@427: 
cmlenz@363:     DIRECTIVE_NAMESPACE = Namespace('http://genshi.edgewall.org/')
cmlenz@363:     XINCLUDE_NAMESPACE = Namespace('http://www.w3.org/2001/XInclude')
cmlenz@336: 
cmlenz@336:     directives = [('def', DefDirective),
cmlenz@336:                   ('match', MatchDirective),
cmlenz@336:                   ('when', WhenDirective),
cmlenz@336:                   ('otherwise', OtherwiseDirective),
cmlenz@336:                   ('for', ForDirective),
cmlenz@336:                   ('if', IfDirective),
cmlenz@336:                   ('choose', ChooseDirective),
cmlenz@336:                   ('with', WithDirective),
cmlenz@336:                   ('replace', ReplaceDirective),
cmlenz@336:                   ('content', ContentDirective),
cmlenz@336:                   ('attrs', AttrsDirective),
cmlenz@336:                   ('strip', StripDirective)]
cmlenz@605:     serializer = 'xml'
cmlenz@636:     _number_conv = Markup
cmlenz@336: 
cmlenz@336:     def __init__(self, source, basedir=None, filename=None, loader=None,
aronacher@642:                  encoding=None, lookup='strict', allow_exec=True,
aronacher@647:                  restricted=False):
cmlenz@336:         Template.__init__(self, source, basedir=basedir, filename=filename,
cmlenz@545:                           loader=loader, encoding=encoding, lookup=lookup,
aronacher@647:                           allow_exec=allow_exec, restricted=restricted)
cmlenz@496:         # Make sure the include filter comes after the match filter
cmlenz@496:         if loader:
cmlenz@496:             self.filters.remove(self._include)
cmlenz@609:         self.filters += [self._match]
cmlenz@496:         if loader:
cmlenz@496:             self.filters.append(self._include)
cmlenz@336: 
cmlenz@374:     def _parse(self, source, encoding):
cmlenz@381:         streams = [[]] # stacked lists of events of the "compiled" template
cmlenz@336:         dirmap = {} # temporary mapping of directives to elements
cmlenz@336:         ns_prefix = {}
cmlenz@336:         depth = 0
cmlenz@590:         fallbacks = []
cmlenz@590:         includes = []
cmlenz@336: 
cmlenz@374:         if not isinstance(source, Stream):
cmlenz@434:             source = XMLParser(source, filename=self.filename,
cmlenz@374:                                encoding=encoding)
cmlenz@374: 
cmlenz@374:         for kind, data, pos in source:
cmlenz@381:             stream = streams[-1]
cmlenz@336: 
cmlenz@336:             if kind is START_NS:
cmlenz@336:                 # Strip out the namespace declaration for template directives
cmlenz@336:                 prefix, uri = data
cmlenz@336:                 ns_prefix[prefix] = uri
cmlenz@363:                 if uri not in (self.DIRECTIVE_NAMESPACE,
cmlenz@363:                                self.XINCLUDE_NAMESPACE):
cmlenz@336:                     stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@336:             elif kind is END_NS:
cmlenz@336:                 uri = ns_prefix.pop(data, None)
cmlenz@363:                 if uri and uri not in (self.DIRECTIVE_NAMESPACE,
cmlenz@363:                                        self.XINCLUDE_NAMESPACE):
cmlenz@336:                     stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@336:             elif kind is START:
cmlenz@336:                 # Record any directive attributes in start tags
cmlenz@363:                 tag, attrs = data
cmlenz@336:                 directives = []
cmlenz@336:                 strip = False
cmlenz@336: 
cmlenz@363:                 if tag in self.DIRECTIVE_NAMESPACE:
cmlenz@336:                     cls = self._dir_by_name.get(tag.localname)
cmlenz@336:                     if cls is None:
cmlenz@336:                         raise BadDirectiveError(tag.localname, self.filepath,
cmlenz@336:                                                 pos[1])
cmlenz@552:                     args = dict([(name.localname, value) for name, value
cmlenz@552:                                  in attrs if not name.namespace])
cmlenz@552:                     directives.append((cls, args, ns_prefix.copy(), pos))
cmlenz@336:                     strip = True
cmlenz@336: 
cmlenz@363:                 new_attrs = []
cmlenz@363:                 for name, value in attrs:
cmlenz@363:                     if name in self.DIRECTIVE_NAMESPACE:
cmlenz@336:                         cls = self._dir_by_name.get(name.localname)
cmlenz@336:                         if cls is None:
cmlenz@336:                             raise BadDirectiveError(name.localname,
cmlenz@336:                                                     self.filepath, pos[1])
cmlenz@362:                         directives.append((cls, value, ns_prefix.copy(), pos))
cmlenz@336:                     else:
cmlenz@336:                         if value:
cmlenz@442:                             value = list(interpolate(value, self.basedir,
cmlenz@442:                                                      pos[0], pos[1], pos[2],
aronacher@647:                                                      lookup=self.lookup,
aronacher@647:                                                      restricted=self.restricted))
cmlenz@336:                             if len(value) == 1 and value[0][0] is TEXT:
cmlenz@336:                                 value = value[0][1]
cmlenz@336:                         else:
cmlenz@336:                             value = [(TEXT, u'', pos)]
cmlenz@363:                         new_attrs.append((name, value))
cmlenz@363:                 new_attrs = Attrs(new_attrs)
cmlenz@336: 
cmlenz@336:                 if directives:
cmlenz@336:                     index = self._dir_order.index
cmlenz@362:                     directives.sort(lambda a, b: cmp(index(a[0]), index(b[0])))
cmlenz@336:                     dirmap[(depth, tag)] = (directives, len(stream), strip)
cmlenz@336: 
cmlenz@363:                 if tag in self.XINCLUDE_NAMESPACE:
cmlenz@363:                     if tag.localname == 'include':
cmlenz@363:                         include_href = new_attrs.get('href')
cmlenz@363:                         if not include_href:
cmlenz@363:                             raise TemplateSyntaxError('Include misses required '
cmlenz@422:                                                       'attribute "href"',
cmlenz@422:                                                       self.filepath, *pos[1:])
cmlenz@610:                         includes.append((include_href, new_attrs.get('parse')))
cmlenz@381:                         streams.append([])
cmlenz@363:                     elif tag.localname == 'fallback':
cmlenz@590:                         streams.append([])
cmlenz@590:                         fallbacks.append(streams[-1])
cmlenz@363: 
cmlenz@363:                 else:
cmlenz@363:                     stream.append((kind, (tag, new_attrs), pos))
cmlenz@363: 
cmlenz@336:                 depth += 1
cmlenz@336: 
cmlenz@336:             elif kind is END:
cmlenz@336:                 depth -= 1
cmlenz@363: 
cmlenz@590:                 if fallbacks and data == self.XINCLUDE_NAMESPACE['fallback']:
cmlenz@590:                     assert streams.pop() is fallbacks[-1]
cmlenz@363:                 elif data == self.XINCLUDE_NAMESPACE['include']:
cmlenz@590:                     fallback = None
cmlenz@590:                     if len(fallbacks) == len(includes):
cmlenz@590:                         fallback = fallbacks.pop()
cmlenz@590:                     streams.pop() # discard anything between the include tags
cmlenz@590:                                   # and the fallback element
cmlenz@381:                     stream = streams[-1]
cmlenz@610:                     href, parse = includes.pop()
cmlenz@610:                     try:
cmlenz@610:                         cls = {
cmlenz@610:                             'xml': MarkupTemplate,
cmlenz@610:                             'text': NewTextTemplate
cmlenz@610:                         }[parse or 'xml']
cmlenz@610:                     except KeyError:
cmlenz@610:                         raise TemplateSyntaxError('Invalid value for "parse" '
cmlenz@610:                                                   'attribute of include',
cmlenz@610:                                                   self.filepath, *pos[1:])
cmlenz@610:                     stream.append((INCLUDE, (href, cls, fallback), pos))
cmlenz@363:                 else:
cmlenz@363:                     stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@336:                 # If there have have directive attributes with the corresponding
cmlenz@336:                 # start tag, move the events inbetween into a "subprogram"
cmlenz@336:                 if (depth, data) in dirmap:
cmlenz@336:                     directives, start_offset, strip = dirmap.pop((depth, data))
cmlenz@336:                     substream = stream[start_offset:]
cmlenz@336:                     if strip:
cmlenz@336:                         substream = substream[1:-1]
cmlenz@336:                     stream[start_offset:] = [(SUB, (directives, substream),
cmlenz@336:                                               pos)]
cmlenz@336: 
cmlenz@405:             elif kind is PI and data[0] == 'python':
cmlenz@545:                 if not self.allow_exec:
cmlenz@545:                     raise TemplateSyntaxError('Python code blocks not allowed',
cmlenz@545:                                               self.filepath, *pos[1:])
cmlenz@405:                 try:
cmlenz@601:                     suite = Suite(data[1], self.filepath, pos[1],
cmlenz@442:                                   lookup=self.lookup)
cmlenz@405:                 except SyntaxError, err:
cmlenz@405:                     raise TemplateSyntaxError(err, self.filepath,
cmlenz@405:                                               pos[1] + (err.lineno or 1) - 1,
cmlenz@405:                                               pos[2] + (err.offset or 0))
cmlenz@405:                 stream.append((EXEC, suite, pos))
cmlenz@405: 
cmlenz@336:             elif kind is TEXT:
cmlenz@442:                 for kind, data, pos in interpolate(data, self.basedir, pos[0],
cmlenz@442:                                                    pos[1], pos[2],
aronacher@647:                                                    lookup=self.lookup,
aronacher@647:                                                    restricted=self.restricted):
cmlenz@336:                     stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@336:             elif kind is COMMENT:
cmlenz@336:                 if not data.lstrip().startswith('!'):
cmlenz@336:                     stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@336:             else:
cmlenz@336:                 stream.append((kind, data, pos))
cmlenz@336: 
cmlenz@381:         assert len(streams) == 1
cmlenz@381:         return streams[0]
cmlenz@381: 
cmlenz@336:     def _match(self, stream, ctxt, match_templates=None):
cmlenz@336:         """Internal stream filter that applies any defined match templates
cmlenz@336:         to the stream.
cmlenz@336:         """
cmlenz@336:         if match_templates is None:
cmlenz@336:             match_templates = ctxt._match_templates
cmlenz@336: 
cmlenz@336:         tail = []
cmlenz@336:         def _strip(stream):
cmlenz@336:             depth = 1
cmlenz@336:             while 1:
cmlenz@336:                 event = stream.next()
cmlenz@336:                 if event[0] is START:
cmlenz@336:                     depth += 1
cmlenz@336:                 elif event[0] is END:
cmlenz@336:                     depth -= 1
cmlenz@336:                 if depth > 0:
cmlenz@336:                     yield event
cmlenz@336:                 else:
cmlenz@336:                     tail[:] = [event]
cmlenz@336:                     break
cmlenz@336: 
cmlenz@336:         for event in stream:
cmlenz@336: 
cmlenz@336:             # We (currently) only care about start and end events for matching
cmlenz@336:             # We might care about namespace events in the future, though
cmlenz@336:             if not match_templates or (event[0] is not START and
cmlenz@336:                                        event[0] is not END):
cmlenz@336:                 yield event
cmlenz@336:                 continue
cmlenz@336: 
cmlenz@602:             for idx, (test, path, template, hints, namespaces, directives) \
cmlenz@602:                     in enumerate(match_templates):
cmlenz@336: 
cmlenz@336:                 if test(event, namespaces, ctxt) is True:
cmlenz@602:                     if 'match_once' in hints:
cmlenz@602:                         del match_templates[idx]
cmlenz@602:                         idx -= 1
cmlenz@336: 
cmlenz@336:                     # Let the remaining match templates know about the event so
cmlenz@336:                     # they get a chance to update their internal state
cmlenz@336:                     for test in [mt[0] for mt in match_templates[idx + 1:]]:
cmlenz@336:                         test(event, namespaces, ctxt, updateonly=True)
cmlenz@336: 
cmlenz@336:                     # Consume and store all events until an end event
cmlenz@336:                     # corresponding to this start event is encountered
cmlenz@602:                     inner = _strip(stream)
cmlenz@602:                     if 'match_once' not in hints \
cmlenz@602:                             and 'not_recursive' not in hints:
cmlenz@602:                         inner = self._match(inner, ctxt, [match_templates[idx]])
cmlenz@602:                     content = list(self._include(chain([event], inner, tail),
cmlenz@602:                                                  ctxt))
cmlenz@336: 
cmlenz@336:                     for test in [mt[0] for mt in match_templates]:
cmlenz@336:                         test(tail[0], namespaces, ctxt, updateonly=True)
cmlenz@336: 
cmlenz@336:                     # Make the select() function available in the body of the
cmlenz@336:                     # match template
cmlenz@336:                     def select(path):
cmlenz@336:                         return Stream(content).select(path, namespaces, ctxt)
cmlenz@336:                     ctxt.push(dict(select=select))
cmlenz@336: 
cmlenz@336:                     # Recursively process the output
cmlenz@336:                     template = _apply_directives(template, ctxt, directives)
cmlenz@602:                     remaining = match_templates
cmlenz@602:                     if 'match_once' not in hints:
cmlenz@602:                         remaining = remaining[:idx] + remaining[idx + 1:]
cmlenz@336:                     for event in self._match(self._eval(self._flatten(template,
cmlenz@336:                                                                       ctxt),
cmlenz@602:                                                         ctxt), ctxt, remaining):
cmlenz@336:                         yield event
cmlenz@336: 
cmlenz@336:                     ctxt.pop()
cmlenz@336:                     break
cmlenz@336: 
cmlenz@336:             else: # no matches
cmlenz@336:                 yield event