1__all__ = [ 2 'HTML5TreeBuilder', 3 ] 4 5import warnings 6from bs4.builder import ( 7 PERMISSIVE, 8 HTML, 9 HTML_5, 10 HTMLTreeBuilder, 11 ) 12from bs4.element import ( 13 NamespacedAttribute, 14 whitespace_re, 15) 16import html5lib 17try: 18 # html5lib >= 0.99999999/1.0b9 19 from html5lib.treebuilders import base as treebuildersbase 20except ImportError: 21 # html5lib <= 0.9999999/1.0b8 22 from html5lib.treebuilders import _base as treebuildersbase 23from html5lib.constants import namespaces 24 25from bs4.element import ( 26 Comment, 27 Doctype, 28 NavigableString, 29 Tag, 30 ) 31 32class HTML5TreeBuilder(HTMLTreeBuilder): 33 """Use html5lib to build a tree.""" 34 35 NAME = "html5lib" 36 37 features = [NAME, PERMISSIVE, HTML_5, HTML] 38 39 def prepare_markup(self, markup, user_specified_encoding, 40 document_declared_encoding=None, exclude_encodings=None): 41 # Store the user-specified encoding for use later on. 42 self.user_specified_encoding = user_specified_encoding 43 44 # document_declared_encoding and exclude_encodings aren't used 45 # ATM because the html5lib TreeBuilder doesn't use 46 # UnicodeDammit. 47 if exclude_encodings: 48 warnings.warn("You provided a value for exclude_encoding, but the html5lib tree builder doesn't support exclude_encoding.") 49 yield (markup, None, None, False) 50 51 # These methods are defined by Beautiful Soup. 52 def feed(self, markup): 53 if self.soup.parse_only is not None: 54 warnings.warn("You provided a value for parse_only, but the html5lib tree builder doesn't support parse_only. The entire document will be parsed.") 55 parser = html5lib.HTMLParser(tree=self.create_treebuilder) 56 doc = parser.parse(markup, encoding=self.user_specified_encoding) 57 58 # Set the character encoding detected by the tokenizer. 59 if isinstance(markup, str): 60 # We need to special-case this because html5lib sets 61 # charEncoding to UTF-8 if it gets Unicode input. 62 doc.original_encoding = None 63 else: 64 doc.original_encoding = parser.tokenizer.stream.charEncoding[0] 65 66 def create_treebuilder(self, namespaceHTMLElements): 67 self.underlying_builder = TreeBuilderForHtml5lib( 68 self.soup, namespaceHTMLElements) 69 return self.underlying_builder 70 71 def test_fragment_to_document(self, fragment): 72 """See `TreeBuilder`.""" 73 return '<html><head></head><body>%s</body></html>' % fragment 74 75 76class TreeBuilderForHtml5lib(treebuildersbase.TreeBuilder): 77 78 def __init__(self, soup, namespaceHTMLElements): 79 self.soup = soup 80 super(TreeBuilderForHtml5lib, self).__init__(namespaceHTMLElements) 81 82 def documentClass(self): 83 self.soup.reset() 84 return Element(self.soup, self.soup, None) 85 86 def insertDoctype(self, token): 87 name = token["name"] 88 publicId = token["publicId"] 89 systemId = token["systemId"] 90 91 doctype = Doctype.for_name_and_ids(name, publicId, systemId) 92 self.soup.object_was_parsed(doctype) 93 94 def elementClass(self, name, namespace): 95 tag = self.soup.new_tag(name, namespace) 96 return Element(tag, self.soup, namespace) 97 98 def commentClass(self, data): 99 return TextNode(Comment(data), self.soup) 100 101 def fragmentClass(self): 102 self.soup = BeautifulSoup("") 103 self.soup.name = "[document_fragment]" 104 return Element(self.soup, self.soup, None) 105 106 def appendChild(self, node): 107 # XXX This code is not covered by the BS4 tests. 108 self.soup.append(node.element) 109 110 def getDocument(self): 111 return self.soup 112 113 def getFragment(self): 114 return treebuildersbase.TreeBuilder.getFragment(self).element 115 116class AttrList(object): 117 def __init__(self, element): 118 self.element = element 119 self.attrs = dict(self.element.attrs) 120 def __iter__(self): 121 return list(self.attrs.items()).__iter__() 122 def __setitem__(self, name, value): 123 # If this attribute is a multi-valued attribute for this element, 124 # turn its value into a list. 125 list_attr = HTML5TreeBuilder.cdata_list_attributes 126 if (name in list_attr['*'] 127 or (self.element.name in list_attr 128 and name in list_attr[self.element.name])): 129 # A node that is being cloned may have already undergone 130 # this procedure. 131 if not isinstance(value, list): 132 value = whitespace_re.split(value) 133 self.element[name] = value 134 def items(self): 135 return list(self.attrs.items()) 136 def keys(self): 137 return list(self.attrs.keys()) 138 def __len__(self): 139 return len(self.attrs) 140 def __getitem__(self, name): 141 return self.attrs[name] 142 def __contains__(self, name): 143 return name in list(self.attrs.keys()) 144 145 146class Element(treebuildersbase.Node): 147 def __init__(self, element, soup, namespace): 148 treebuildersbase.Node.__init__(self, element.name) 149 self.element = element 150 self.soup = soup 151 self.namespace = namespace 152 153 def appendChild(self, node): 154 string_child = child = None 155 if isinstance(node, str): 156 # Some other piece of code decided to pass in a string 157 # instead of creating a TextElement object to contain the 158 # string. 159 string_child = child = node 160 elif isinstance(node, Tag): 161 # Some other piece of code decided to pass in a Tag 162 # instead of creating an Element object to contain the 163 # Tag. 164 child = node 165 elif node.element.__class__ == NavigableString: 166 string_child = child = node.element 167 else: 168 child = node.element 169 170 if not isinstance(child, str) and child.parent is not None: 171 node.element.extract() 172 173 if (string_child and self.element.contents 174 and self.element.contents[-1].__class__ == NavigableString): 175 # We are appending a string onto another string. 176 # TODO This has O(n^2) performance, for input like 177 # "a</a>a</a>a</a>..." 178 old_element = self.element.contents[-1] 179 new_element = self.soup.new_string(old_element + string_child) 180 old_element.replace_with(new_element) 181 self.soup._most_recent_element = new_element 182 else: 183 if isinstance(node, str): 184 # Create a brand new NavigableString from this string. 185 child = self.soup.new_string(node) 186 187 # Tell Beautiful Soup to act as if it parsed this element 188 # immediately after the parent's last descendant. (Or 189 # immediately after the parent, if it has no children.) 190 if self.element.contents: 191 most_recent_element = self.element._last_descendant(False) 192 elif self.element.next_element is not None: 193 # Something from further ahead in the parse tree is 194 # being inserted into this earlier element. This is 195 # very annoying because it means an expensive search 196 # for the last element in the tree. 197 most_recent_element = self.soup._last_descendant() 198 else: 199 most_recent_element = self.element 200 201 self.soup.object_was_parsed( 202 child, parent=self.element, 203 most_recent_element=most_recent_element) 204 205 def getAttributes(self): 206 return AttrList(self.element) 207 208 def setAttributes(self, attributes): 209 210 if attributes is not None and len(attributes) > 0: 211 212 converted_attributes = [] 213 for name, value in list(attributes.items()): 214 if isinstance(name, tuple): 215 new_name = NamespacedAttribute(*name) 216 del attributes[name] 217 attributes[new_name] = value 218 219 self.soup.builder._replace_cdata_list_attribute_values( 220 self.name, attributes) 221 for name, value in list(attributes.items()): 222 self.element[name] = value 223 224 # The attributes may contain variables that need substitution. 225 # Call set_up_substitutions manually. 226 # 227 # The Tag constructor called this method when the Tag was created, 228 # but we just set/changed the attributes, so call it again. 229 self.soup.builder.set_up_substitutions(self.element) 230 attributes = property(getAttributes, setAttributes) 231 232 def insertText(self, data, insertBefore=None): 233 if insertBefore: 234 text = TextNode(self.soup.new_string(data), self.soup) 235 self.insertBefore(data, insertBefore) 236 else: 237 self.appendChild(data) 238 239 def insertBefore(self, node, refNode): 240 index = self.element.index(refNode.element) 241 if (node.element.__class__ == NavigableString and self.element.contents 242 and self.element.contents[index-1].__class__ == NavigableString): 243 # (See comments in appendChild) 244 old_node = self.element.contents[index-1] 245 new_str = self.soup.new_string(old_node + node.element) 246 old_node.replace_with(new_str) 247 else: 248 self.element.insert(index, node.element) 249 node.parent = self 250 251 def removeChild(self, node): 252 node.element.extract() 253 254 def reparentChildren(self, new_parent): 255 """Move all of this tag's children into another tag.""" 256 # print "MOVE", self.element.contents 257 # print "FROM", self.element 258 # print "TO", new_parent.element 259 element = self.element 260 new_parent_element = new_parent.element 261 # Determine what this tag's next_element will be once all the children 262 # are removed. 263 final_next_element = element.next_sibling 264 265 new_parents_last_descendant = new_parent_element._last_descendant(False, False) 266 if len(new_parent_element.contents) > 0: 267 # The new parent already contains children. We will be 268 # appending this tag's children to the end. 269 new_parents_last_child = new_parent_element.contents[-1] 270 new_parents_last_descendant_next_element = new_parents_last_descendant.next_element 271 else: 272 # The new parent contains no children. 273 new_parents_last_child = None 274 new_parents_last_descendant_next_element = new_parent_element.next_element 275 276 to_append = element.contents 277 append_after = new_parent_element.contents 278 if len(to_append) > 0: 279 # Set the first child's previous_element and previous_sibling 280 # to elements within the new parent 281 first_child = to_append[0] 282 if new_parents_last_descendant: 283 first_child.previous_element = new_parents_last_descendant 284 else: 285 first_child.previous_element = new_parent_element 286 first_child.previous_sibling = new_parents_last_child 287 if new_parents_last_descendant: 288 new_parents_last_descendant.next_element = first_child 289 else: 290 new_parent_element.next_element = first_child 291 if new_parents_last_child: 292 new_parents_last_child.next_sibling = first_child 293 294 # Fix the last child's next_element and next_sibling 295 last_child = to_append[-1] 296 last_child.next_element = new_parents_last_descendant_next_element 297 if new_parents_last_descendant_next_element: 298 new_parents_last_descendant_next_element.previous_element = last_child 299 last_child.next_sibling = None 300 301 for child in to_append: 302 child.parent = new_parent_element 303 new_parent_element.contents.append(child) 304 305 # Now that this element has no children, change its .next_element. 306 element.contents = [] 307 element.next_element = final_next_element 308 309 # print "DONE WITH MOVE" 310 # print "FROM", self.element 311 # print "TO", new_parent_element 312 313 def cloneNode(self): 314 tag = self.soup.new_tag(self.element.name, self.namespace) 315 node = Element(tag, self.soup, self.namespace) 316 for key,value in self.attributes: 317 node.attributes[key] = value 318 return node 319 320 def hasContent(self): 321 return self.element.contents 322 323 def getNameTuple(self): 324 if self.namespace is None: 325 return namespaces["html"], self.name 326 else: 327 return self.namespace, self.name 328 329 nameTuple = property(getNameTuple) 330 331class TextNode(Element): 332 def __init__(self, element, soup): 333 treebuildersbase.Node.__init__(self, None) 334 self.element = element 335 self.soup = soup 336 337 def cloneNode(self): 338 raise NotImplementedError 339