1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-11-13 12:55:05 -05:00
SickRage/lib/html5lib/treebuilders/soup.py

237 lines
8.5 KiB
Python
Raw Normal View History

import warnings
warnings.warn("BeautifulSoup 3.x (as of 3.1) is not fully compatible with html5lib and support will be removed in the future", DeprecationWarning)
from BeautifulSoup import BeautifulSoup, Tag, NavigableString, Comment, Declaration
import _base
from html5lib.constants import namespaces, DataLossWarning
class AttrList(object):
def __init__(self, element):
self.element = element
self.attrs = dict(self.element.attrs)
def __iter__(self):
return self.attrs.items().__iter__()
def __setitem__(self, name, value):
"set attr", name, value
self.element[name] = value
def items(self):
return self.attrs.items()
def keys(self):
return self.attrs.keys()
def __getitem__(self, name):
return self.attrs[name]
def __contains__(self, name):
return name in self.attrs.keys()
def __eq__(self, other):
if len(self.keys()) != len(other.keys()):
return False
for item in self.keys():
if item not in other:
return False
if self[item] != other[item]:
return False
return True
class Element(_base.Node):
def __init__(self, element, soup, namespace):
_base.Node.__init__(self, element.name)
self.element = element
self.soup = soup
self.namespace = namespace
def _nodeIndex(self, node, refNode):
# Finds a node by identity rather than equality
for index in range(len(self.element.contents)):
if id(self.element.contents[index]) == id(refNode.element):
return index
return None
def appendChild(self, node):
if (node.element.__class__ == NavigableString and self.element.contents
and self.element.contents[-1].__class__ == NavigableString):
# Concatenate new text onto old text node
# (TODO: This has O(n^2) performance, for input like "a</a>a</a>a</a>...")
newStr = NavigableString(self.element.contents[-1]+node.element)
# Remove the old text node
# (Can't simply use .extract() by itself, because it fails if
# an equal text node exists within the parent node)
oldElement = self.element.contents[-1]
del self.element.contents[-1]
oldElement.parent = None
oldElement.extract()
self.element.insert(len(self.element.contents), newStr)
else:
self.element.insert(len(self.element.contents), node.element)
node.parent = self
def getAttributes(self):
return AttrList(self.element)
def setAttributes(self, attributes):
if attributes:
for name, value in attributes.items():
self.element[name] = value
attributes = property(getAttributes, setAttributes)
def insertText(self, data, insertBefore=None):
text = TextNode(NavigableString(data), self.soup)
if insertBefore:
self.insertBefore(text, insertBefore)
else:
self.appendChild(text)
def insertBefore(self, node, refNode):
index = self._nodeIndex(node, refNode)
if (node.element.__class__ == NavigableString and self.element.contents
and self.element.contents[index-1].__class__ == NavigableString):
# (See comments in appendChild)
newStr = NavigableString(self.element.contents[index-1]+node.element)
oldNode = self.element.contents[index-1]
del self.element.contents[index-1]
oldNode.parent = None
oldNode.extract()
self.element.insert(index-1, newStr)
else:
self.element.insert(index, node.element)
node.parent = self
def removeChild(self, node):
index = self._nodeIndex(node.parent, node)
del node.parent.element.contents[index]
node.element.parent = None
node.element.extract()
node.parent = None
def reparentChildren(self, newParent):
while self.element.contents:
child = self.element.contents[0]
child.extract()
if isinstance(child, Tag):
newParent.appendChild(Element(child, self.soup, namespaces["html"]))
else:
newParent.appendChild(TextNode(child, self.soup))
def cloneNode(self):
node = Element(Tag(self.soup, self.element.name), self.soup, self.namespace)
for key,value in self.attributes:
node.attributes[key] = value
return node
def hasContent(self):
return self.element.contents
def getNameTuple(self):
if self.namespace == None:
return namespaces["html"], self.name
else:
return self.namespace, self.name
nameTuple = property(getNameTuple)
class TextNode(Element):
def __init__(self, element, soup):
_base.Node.__init__(self, None)
self.element = element
self.soup = soup
def cloneNode(self):
raise NotImplementedError
class TreeBuilder(_base.TreeBuilder):
def __init__(self, namespaceHTMLElements):
if namespaceHTMLElements:
warnings.warn("BeautifulSoup cannot represent elements in any namespace", DataLossWarning)
_base.TreeBuilder.__init__(self, namespaceHTMLElements)
def documentClass(self):
self.soup = BeautifulSoup("")
return Element(self.soup, self.soup, None)
def insertDoctype(self, token):
name = token["name"]
publicId = token["publicId"]
systemId = token["systemId"]
if publicId:
self.soup.insert(0, Declaration("DOCTYPE %s PUBLIC \"%s\" \"%s\""%(name, publicId, systemId or "")))
elif systemId:
self.soup.insert(0, Declaration("DOCTYPE %s SYSTEM \"%s\""%
(name, systemId)))
else:
self.soup.insert(0, Declaration("DOCTYPE %s"%name))
def elementClass(self, name, namespace):
if namespace is not None:
warnings.warn("BeautifulSoup cannot represent elements in any namespace", DataLossWarning)
return Element(Tag(self.soup, name), self.soup, namespace)
def commentClass(self, data):
return TextNode(Comment(data), self.soup)
def fragmentClass(self):
self.soup = BeautifulSoup("")
self.soup.name = "[document_fragment]"
return Element(self.soup, self.soup, None)
def appendChild(self, node):
self.soup.insert(len(self.soup.contents), node.element)
def testSerializer(self, element):
return testSerializer(element)
def getDocument(self):
return self.soup
def getFragment(self):
return _base.TreeBuilder.getFragment(self).element
def testSerializer(element):
import re
rv = []
def serializeElement(element, indent=0):
if isinstance(element, Declaration):
doctype_regexp = r'DOCTYPE\s+(?P<name>[^\s]*)( PUBLIC "(?P<publicId>.*)" "(?P<systemId1>.*)"| SYSTEM "(?P<systemId2>.*)")?'
m = re.compile(doctype_regexp).match(element.string)
assert m is not None, "DOCTYPE did not match expected format"
name = m.group('name')
publicId = m.group('publicId')
if publicId is not None:
systemId = m.group('systemId1') or ""
else:
systemId = m.group('systemId2')
if publicId is not None or systemId is not None:
rv.append("""|%s<!DOCTYPE %s "%s" "%s">"""%
(' '*indent, name, publicId or "", systemId or ""))
else:
rv.append("|%s<!DOCTYPE %s>"%(' '*indent, name))
elif isinstance(element, BeautifulSoup):
if element.name == "[document_fragment]":
rv.append("#document-fragment")
else:
rv.append("#document")
elif isinstance(element, Comment):
rv.append("|%s<!-- %s -->"%(' '*indent, element.string))
elif isinstance(element, unicode):
rv.append("|%s\"%s\"" %(' '*indent, element))
else:
rv.append("|%s<%s>"%(' '*indent, element.name))
if element.attrs:
for name, value in sorted(element.attrs):
rv.append('|%s%s="%s"' % (' '*(indent+2), name, value))
indent += 2
if hasattr(element, "contents"):
for child in element.contents:
serializeElement(child, indent)
serializeElement(element, 0)
return "\n".join(rv)