Mercurial > genshi > genshi-test
annotate markup/filters.py @ 12:87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
| author | cmlenz |
|---|---|
| date | Sun, 04 Jun 2006 20:37:59 +0000 |
| parents | c5890ef863ba |
| children | bf9de5a4c896 |
| rev | line source |
|---|---|
| 1 | 1 # -*- coding: utf-8 -*- |
| 2 # | |
| 3 # Copyright (C) 2006 Christopher Lenz | |
| 4 # All rights reserved. | |
| 5 # | |
| 6 # This software is licensed as described in the file COPYING, which | |
| 7 # you should have received as part of this distribution. The terms | |
| 8 # are also available at http://trac.edgewall.com/license.html. | |
| 9 # | |
| 10 # This software consists of voluntary contributions made by many | |
| 11 # individuals. For the exact contribution history, see the revision | |
| 12 # history and logs, available at http://projects.edgewall.com/trac/. | |
| 13 | |
| 14 """Implementation of a number of stream filters.""" | |
| 15 | |
| 16 try: | |
| 17 frozenset | |
| 18 except NameError: | |
| 19 from sets import ImmutableSet as frozenset | |
| 20 import re | |
| 21 | |
| 22 from markup.core import Attributes, Markup, Stream | |
| 23 from markup.path import Path | |
| 24 | |
| 25 __all__ = ['EvalFilter', 'IncludeFilter', 'MatchFilter', 'WhitespaceFilter', | |
| 26 'HTMLSanitizer'] | |
| 27 | |
| 28 | |
| 29 class EvalFilter(object): | |
| 30 """Responsible for evaluating expressions in a template.""" | |
| 31 | |
| 32 def __call__(self, stream, ctxt=None): | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
33 from markup.template import Template |
|
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
34 |
| 1 | 35 for kind, data, pos in stream: |
| 36 | |
| 37 if kind is Stream.START: | |
| 38 # Attributes may still contain expressions in start tags at | |
| 39 # this point, so do some evaluation | |
| 40 tag, attrib = data | |
| 41 new_attrib = [] | |
| 42 for name, substream in attrib: | |
| 43 if isinstance(substream, basestring): | |
| 44 value = substream | |
| 45 else: | |
| 46 values = [] | |
| 47 for subkind, subdata, subpos in substream: | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
48 if subkind is Template.EXPR: |
| 1 | 49 values.append(subdata.evaluate(ctxt)) |
| 50 else: | |
| 51 values.append(subdata) | |
| 52 value = filter(lambda x: x is not None, values) | |
| 53 if not value: | |
| 54 continue | |
| 55 new_attrib.append((name, ''.join(value))) | |
| 56 yield kind, (tag, Attributes(new_attrib)), pos | |
| 57 | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
58 elif kind is Template.EXPR: |
| 1 | 59 result = data.evaluate(ctxt) |
| 60 if result is None: | |
| 61 continue | |
| 62 | |
| 63 # First check for a string, otherwise the iterable | |
| 64 # test below succeeds, and the string will be | |
| 65 # chopped up into characters | |
| 66 if isinstance(result, basestring): | |
| 67 yield Stream.TEXT, result, pos | |
| 68 else: | |
| 69 # Test if the expression evaluated to an | |
| 70 # iterable, in which case we yield the | |
| 71 # individual items | |
| 72 try: | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
73 yield Template.SUB, ([], iter(result)), pos |
| 1 | 74 except TypeError: |
| 75 # Neither a string nor an iterable, so just | |
| 76 # pass it through | |
| 77 yield Stream.TEXT, unicode(result), pos | |
| 78 | |
| 79 else: | |
| 80 yield kind, data, pos | |
| 81 | |
| 82 | |
| 83 class IncludeFilter(object): | |
| 84 """Template filter providing (very) basic XInclude support | |
| 85 (see http://www.w3.org/TR/xinclude/) in templates. | |
| 86 """ | |
| 87 | |
| 88 _NAMESPACE = 'http://www.w3.org/2001/XInclude' | |
| 89 | |
| 90 def __init__(self, loader): | |
| 91 """Initialize the filter. | |
| 92 | |
| 93 @param loader: the `TemplateLoader` to use for resolving references to | |
| 94 external template files | |
| 95 """ | |
| 96 self.loader = loader | |
| 97 | |
|
12
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
98 def __call__(self, stream, ctxt=None, ns_prefixes=None): |
| 1 | 99 """Filter the stream, processing any XInclude directives it may |
| 100 contain. | |
| 101 | |
| 102 @param ctxt: the template context | |
| 103 @param stream: the markup event stream to filter | |
| 104 """ | |
|
12
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
105 from markup.template import Template, TemplateError, TemplateNotFound |
| 1 | 106 |
|
12
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
107 if ns_prefixes is None: |
|
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
108 ns_prefixes = [] |
| 1 | 109 in_fallback = False |
| 110 include_href, fallback_stream = None, None | |
| 111 indent = 0 | |
| 112 | |
| 113 for kind, data, pos in stream: | |
| 114 | |
| 115 if kind is Stream.START and data[0].namespace == self._NAMESPACE \ | |
| 116 and not in_fallback: | |
| 117 tag, attrib = data | |
| 118 if tag.localname == 'include': | |
| 119 include_href = attrib.get('href') | |
| 120 indent = pos[1] | |
| 121 elif tag.localname == 'fallback': | |
| 122 in_fallback = True | |
| 123 fallback_stream = [] | |
| 124 | |
| 125 elif kind is Stream.END and data.namespace == self._NAMESPACE: | |
| 126 if data.localname == 'include': | |
| 127 try: | |
| 128 if not include_href: | |
| 129 raise TemplateError('Include misses required ' | |
| 130 'attribute "href"') | |
| 131 template = self.loader.load(include_href) | |
| 132 for ikind, idata, ipos in template.generate(ctxt): | |
| 133 # Fixup indentation of included markup | |
| 134 if ikind is Stream.TEXT: | |
| 135 idata = idata.replace('\n', '\n' + ' ' * indent) | |
| 136 yield ikind, idata, ipos | |
| 137 | |
| 138 # If the included template defines any filters added at | |
| 139 # runtime (such as py:match templates), those need to be | |
| 140 # applied to the including template, too. | |
| 141 for filter_ in template.filters: | |
| 142 stream = filter_(stream, ctxt) | |
| 143 | |
| 144 except TemplateNotFound: | |
| 145 if fallback_stream is None: | |
| 146 raise | |
| 147 for event in fallback_stream: | |
| 148 yield event | |
| 149 | |
| 150 include_href = None | |
| 151 fallback_stream = None | |
| 152 indent = 0 | |
| 153 break | |
| 154 elif data.localname == 'fallback': | |
| 155 in_fallback = False | |
| 156 | |
| 157 elif in_fallback: | |
| 158 fallback_stream.append((kind, data, pos)) | |
| 159 | |
| 160 elif kind is Stream.START_NS and data[1] == self._NAMESPACE: | |
|
12
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
161 ns_prefixes.append(data[0]) |
|
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
162 continue |
|
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
163 |
|
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
164 elif kind is Stream.END_NS and data in ns_prefixes: |
|
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
165 ns_prefixes.pop() |
| 1 | 166 continue |
| 167 | |
| 168 else: | |
| 169 yield kind, data, pos | |
| 170 else: | |
| 171 # The loop exited normally, so there shouldn't be further events to | |
| 172 # process | |
| 173 return | |
| 174 | |
|
12
87238328a71d
Make the XInclude filter track namespace context, to enable it to omit `END_NS` events for the XInclude namespace.
cmlenz
parents:
10
diff
changeset
|
175 for event in self(stream, ctxt, ns_prefixes=ns_prefixes): |
| 1 | 176 yield event |
| 177 | |
| 178 | |
| 179 class MatchFilter(object): | |
| 180 """A filter that delegates to a given handler function when the input stream | |
| 181 matches some path expression. | |
| 182 """ | |
| 183 | |
| 184 def __init__(self, path, handler): | |
| 185 self.path = Path(path) | |
| 186 self.handler = handler | |
| 187 | |
| 188 def __call__(self, stream, ctxt=None): | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
189 from markup.template import Template |
|
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
190 |
| 1 | 191 test = self.path.test() |
| 192 for kind, data, pos in stream: | |
| 193 result = test(kind, data, pos) | |
| 194 if result is True: | |
| 195 content = [(kind, data, pos)] | |
| 196 depth = 1 | |
| 197 while depth > 0: | |
| 198 ev = stream.next() | |
| 199 if ev[0] is Stream.START: | |
| 200 depth += 1 | |
| 201 elif ev[0] is Stream.END: | |
| 202 depth -= 1 | |
| 203 content.append(ev) | |
| 204 test(*ev) | |
| 205 | |
|
10
c5890ef863ba
Moved the template-specific stream event kinds into the template module.
cmlenz
parents:
1
diff
changeset
|
206 yield (Template.SUB, |
| 1 | 207 ([lambda stream, ctxt: self.handler(content, ctxt)], []), |
| 208 pos) | |
| 209 else: | |
| 210 yield kind, data, pos | |
| 211 | |
| 212 | |
| 213 class WhitespaceFilter(object): | |
| 214 """A filter that removes extraneous white space from the stream. | |
| 215 | |
| 216 Todo: | |
| 217 * Support for xml:space | |
| 218 """ | |
| 219 | |
| 220 _TRAILING_SPACE = re.compile('[ \t]+(?=\n)') | |
| 221 _LINE_COLLAPSE = re.compile('\n{2,}') | |
| 222 | |
| 223 def __call__(self, stream, ctxt=None): | |
| 224 textbuf = [] | |
| 225 prev_kind = None | |
| 226 for kind, data, pos in stream: | |
| 227 if kind is Stream.TEXT: | |
| 228 textbuf.append(data) | |
| 229 elif prev_kind is Stream.TEXT: | |
| 230 text = ''.join(textbuf) | |
| 231 text = self._TRAILING_SPACE.sub('', text) | |
| 232 text = self._LINE_COLLAPSE.sub('\n', text) | |
| 233 yield Stream.TEXT, text, pos | |
| 234 del textbuf[:] | |
| 235 prev_kind = kind | |
| 236 if kind is not Stream.TEXT: | |
| 237 yield kind, data, pos | |
| 238 | |
| 239 if textbuf: | |
| 240 text = self._LINE_COLLAPSE.sub('\n', ''.join(textbuf)) | |
| 241 yield Stream.TEXT, text, pos | |
| 242 | |
| 243 | |
| 244 class HTMLSanitizer(object): | |
| 245 """A filter that removes potentially dangerous HTML tags and attributes | |
| 246 from the stream. | |
| 247 """ | |
| 248 | |
| 249 _SAFE_TAGS = frozenset(['a', 'abbr', 'acronym', 'address', 'area', 'b', | |
| 250 'big', 'blockquote', 'br', 'button', 'caption', 'center', 'cite', | |
| 251 'code', 'col', 'colgroup', 'dd', 'del', 'dfn', 'dir', 'div', 'dl', 'dt', | |
| 252 'em', 'fieldset', 'font', 'form', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', | |
| 253 'hr', 'i', 'img', 'input', 'ins', 'kbd', 'label', 'legend', 'li', 'map', | |
| 254 'menu', 'ol', 'optgroup', 'option', 'p', 'pre', 'q', 's', 'samp', | |
| 255 'select', 'small', 'span', 'strike', 'strong', 'sub', 'sup', 'table', | |
| 256 'tbody', 'td', 'textarea', 'tfoot', 'th', 'thead', 'tr', 'tt', 'u', | |
| 257 'ul', 'var']) | |
| 258 | |
| 259 _SAFE_ATTRS = frozenset(['abbr', 'accept', 'accept-charset', 'accesskey', | |
| 260 'action', 'align', 'alt', 'axis', 'border', 'cellpadding', | |
| 261 'cellspacing', 'char', 'charoff', 'charset', 'checked', 'cite', 'class', | |
| 262 'clear', 'cols', 'colspan', 'color', 'compact', 'coords', 'datetime', | |
| 263 'dir', 'disabled', 'enctype', 'for', 'frame', 'headers', 'height', | |
| 264 'href', 'hreflang', 'hspace', 'id', 'ismap', 'label', 'lang', | |
| 265 'longdesc', 'maxlength', 'media', 'method', 'multiple', 'name', | |
| 266 'nohref', 'noshade', 'nowrap', 'prompt', 'readonly', 'rel', 'rev', | |
| 267 'rows', 'rowspan', 'rules', 'scope', 'selected', 'shape', 'size', | |
| 268 'span', 'src', 'start', 'style', 'summary', 'tabindex', 'target', | |
| 269 'title', 'type', 'usemap', 'valign', 'value', 'vspace', 'width']) | |
| 270 _URI_ATTRS = frozenset(['action', 'background', 'dynsrc', 'href', 'lowsrc', | |
| 271 'src']) | |
| 272 _SAFE_SCHEMES = frozenset(['file', 'ftp', 'http', 'https', 'mailto', None]) | |
| 273 | |
| 274 def __call__(self, stream, ctxt=None): | |
| 275 waiting_for = None | |
| 276 | |
| 277 for kind, data, pos in stream: | |
| 278 if kind is Stream.START: | |
| 279 if waiting_for: | |
| 280 continue | |
| 281 tag, attrib = data | |
| 282 if tag not in self._SAFE_TAGS: | |
| 283 waiting_for = tag | |
| 284 continue | |
| 285 | |
| 286 new_attrib = [] | |
| 287 for attr, value in attrib: | |
| 288 if attr not in self._SAFE_ATTRS: | |
| 289 continue | |
| 290 elif attr in self._URI_ATTRS: | |
| 291 # Don't allow URI schemes such as "javascript:" | |
| 292 if self._get_scheme(value) not in self._SAFE_SCHEMES: | |
| 293 continue | |
| 294 elif attr == 'style': | |
| 295 # Remove dangerous CSS declarations from inline styles | |
| 296 decls = [] | |
| 297 for decl in filter(None, value.split(';')): | |
| 298 is_evil = False | |
| 299 if 'expression' in decl: | |
| 300 is_evil = True | |
| 301 for m in re.finditer(r'url\s*\(([^)]+)', decl): | |
| 302 if self._get_scheme(m.group(1)) not in self._SAFE_SCHEMES: | |
| 303 is_evil = True | |
| 304 break | |
| 305 if not is_evil: | |
| 306 decls.append(decl.strip()) | |
| 307 if not decls: | |
| 308 continue | |
| 309 value = '; '.join(decls) | |
| 310 new_attrib.append((attr, value)) | |
| 311 | |
| 312 yield kind, (tag, new_attrib), pos | |
| 313 | |
| 314 elif kind is Stream.END: | |
| 315 tag = data | |
| 316 if waiting_for: | |
| 317 if waiting_for == tag: | |
| 318 waiting_for = None | |
| 319 else: | |
| 320 yield kind, data, pos | |
| 321 | |
| 322 else: | |
| 323 if not waiting_for: | |
| 324 yield kind, data, pos | |
| 325 | |
| 326 def _get_scheme(self, text): | |
| 327 if ':' not in text: | |
| 328 return None | |
| 329 chars = [char for char in text.split(':', 1)[0] if char.isalnum()] | |
| 330 return ''.join(chars).lower() |