Initial Commit

lib/html5lib/__init__.py

@@ -0,0 +1,23 @@
+"""
+HTML parsing library based on the WHATWG "HTML5"
+specification. The parser is designed to be compatible with existing
+HTML found in the wild and implements well-defined error recovery that
+is largely compatible with modern desktop web browsers.
+
+Example usage:
+
+import html5lib
+f = open("my_document.html")
+tree = html5lib.parse(f)
+"""
+
+from __future__ import absolute_import, division, unicode_literals
+
+from .html5parser import HTMLParser, parse, parseFragment
+from .treebuilders import getTreeBuilder
+from .treewalkers import getTreeWalker
+from .serializer import serialize
+
+__all__ = ["HTMLParser", "parse", "parseFragment", "getTreeBuilder",
+           "getTreeWalker", "serialize"]
+__version__ = "0.999"

lib/html5lib/filters/_base.py

@@ -0,0 +1,12 @@
+from __future__ import absolute_import, division, unicode_literals
+
+
+class Filter(object):
+    def __init__(self, source):
+        self.source = source
+
+    def __iter__(self):
+        return iter(self.source)
+
+    def __getattr__(self, name):
+        return getattr(self.source, name)

lib/html5lib/filters/alphabeticalattributes.py

@@ -0,0 +1,20 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from . import _base
+
+try:
+    from collections import OrderedDict
+except ImportError:
+    from ordereddict import OrderedDict
+
+
+class Filter(_base.Filter):
+    def __iter__(self):
+        for token in _base.Filter.__iter__(self):
+            if token["type"] in ("StartTag", "EmptyTag"):
+                attrs = OrderedDict()
+                for name, value in sorted(token["data"].items(),
+                                          key=lambda x: x[0]):
+                    attrs[name] = value
+                token["data"] = attrs
+            yield token

lib/html5lib/filters/inject_meta_charset.py

@@ -0,0 +1,65 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from . import _base
+
+
+class Filter(_base.Filter):
+    def __init__(self, source, encoding):
+        _base.Filter.__init__(self, source)
+        self.encoding = encoding
+
+    def __iter__(self):
+        state = "pre_head"
+        meta_found = (self.encoding is None)
+        pending = []
+
+        for token in _base.Filter.__iter__(self):
+            type = token["type"]
+            if type == "StartTag":
+                if token["name"].lower() == "head":
+                    state = "in_head"
+
+            elif type == "EmptyTag":
+                if token["name"].lower() == "meta":
+                    # replace charset with actual encoding
+                    has_http_equiv_content_type = False
+                    for (namespace, name), value in token["data"].items():
+                        if namespace is not None:
+                            continue
+                        elif name.lower() == 'charset':
+                            token["data"][(namespace, name)] = self.encoding
+                            meta_found = True
+                            break
+                        elif name == 'http-equiv' and value.lower() == 'content-type':
+                            has_http_equiv_content_type = True
+                    else:
+                        if has_http_equiv_content_type and (None, "content") in token["data"]:
+                            token["data"][(None, "content")] = 'text/html; charset=%s' % self.encoding
+                            meta_found = True
+
+                elif token["name"].lower() == "head" and not meta_found:
+                    # insert meta into empty head
+                    yield {"type": "StartTag", "name": "head",
+                           "data": token["data"]}
+                    yield {"type": "EmptyTag", "name": "meta",
+                           "data": {(None, "charset"): self.encoding}}
+                    yield {"type": "EndTag", "name": "head"}
+                    meta_found = True
+                    continue
+
+            elif type == "EndTag":
+                if token["name"].lower() == "head" and pending:
+                    # insert meta into head (if necessary) and flush pending queue
+                    yield pending.pop(0)
+                    if not meta_found:
+                        yield {"type": "EmptyTag", "name": "meta",
+                               "data": {(None, "charset"): self.encoding}}
+                    while pending:
+                        yield pending.pop(0)
+                    meta_found = True
+                    state = "post_head"
+
+            if state == "in_head":
+                pending.append(token)
+            else:
+                yield token

lib/html5lib/filters/lint.py

@@ -0,0 +1,93 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from gettext import gettext
+_ = gettext
+
+from . import _base
+from ..constants import cdataElements, rcdataElements, voidElements
+
+from ..constants import spaceCharacters
+spaceCharacters = "".join(spaceCharacters)
+
+
+class LintError(Exception):
+    pass
+
+
+class Filter(_base.Filter):
+    def __iter__(self):
+        open_elements = []
+        contentModelFlag = "PCDATA"
+        for token in _base.Filter.__iter__(self):
+            type = token["type"]
+            if type in ("StartTag", "EmptyTag"):
+                name = token["name"]
+                if contentModelFlag != "PCDATA":
+                    raise LintError(_("StartTag not in PCDATA content model flag: %(tag)s") % {"tag": name})
+                if not isinstance(name, str):
+                    raise LintError(_("Tag name is not a string: %(tag)r") % {"tag": name})
+                if not name:
+                    raise LintError(_("Empty tag name"))
+                if type == "StartTag" and name in voidElements:
+                    raise LintError(_("Void element reported as StartTag token: %(tag)s") % {"tag": name})
+                elif type == "EmptyTag" and name not in voidElements:
+                    raise LintError(_("Non-void element reported as EmptyTag token: %(tag)s") % {"tag": token["name"]})
+                if type == "StartTag":
+                    open_elements.append(name)
+                for name, value in token["data"]:
+                    if not isinstance(name, str):
+                        raise LintError(_("Attribute name is not a string: %(name)r") % {"name": name})
+                    if not name:
+                        raise LintError(_("Empty attribute name"))
+                    if not isinstance(value, str):
+                        raise LintError(_("Attribute value is not a string: %(value)r") % {"value": value})
+                if name in cdataElements:
+                    contentModelFlag = "CDATA"
+                elif name in rcdataElements:
+                    contentModelFlag = "RCDATA"
+                elif name == "plaintext":
+                    contentModelFlag = "PLAINTEXT"
+
+            elif type == "EndTag":
+                name = token["name"]
+                if not isinstance(name, str):
+                    raise LintError(_("Tag name is not a string: %(tag)r") % {"tag": name})
+                if not name:
+                    raise LintError(_("Empty tag name"))
+                if name in voidElements:
+                    raise LintError(_("Void element reported as EndTag token: %(tag)s") % {"tag": name})
+                start_name = open_elements.pop()
+                if start_name != name:
+                    raise LintError(_("EndTag (%(end)s) does not match StartTag (%(start)s)") % {"end": name, "start": start_name})
+                contentModelFlag = "PCDATA"
+
+            elif type == "Comment":
+                if contentModelFlag != "PCDATA":
+                    raise LintError(_("Comment not in PCDATA content model flag"))
+
+            elif type in ("Characters", "SpaceCharacters"):
+                data = token["data"]
+                if not isinstance(data, str):
+                    raise LintError(_("Attribute name is not a string: %(name)r") % {"name": data})
+                if not data:
+                    raise LintError(_("%(type)s token with empty data") % {"type": type})
+                if type == "SpaceCharacters":
+                    data = data.strip(spaceCharacters)
+                    if data:
+                        raise LintError(_("Non-space character(s) found in SpaceCharacters token: %(token)r") % {"token": data})
+
+            elif type == "Doctype":
+                name = token["name"]
+                if contentModelFlag != "PCDATA":
+                    raise LintError(_("Doctype not in PCDATA content model flag: %(name)s") % {"name": name})
+                if not isinstance(name, str):
+                    raise LintError(_("Tag name is not a string: %(tag)r") % {"tag": name})
+                # XXX: what to do with token["data"] ?
+
+            elif type in ("ParseError", "SerializeError"):
+                pass
+
+            else:
+                raise LintError(_("Unknown token type: %(type)s") % {"type": type})
+
+            yield token

lib/html5lib/filters/optionaltags.py

@@ -0,0 +1,205 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from . import _base
+
+
+class Filter(_base.Filter):
+    def slider(self):
+        previous1 = previous2 = None
+        for token in self.source:
+            if previous1 is not None:
+                yield previous2, previous1, token
+            previous2 = previous1
+            previous1 = token
+        yield previous2, previous1, None
+
+    def __iter__(self):
+        for previous, token, next in self.slider():
+            type = token["type"]
+            if type == "StartTag":
+                if (token["data"] or
+                        not self.is_optional_start(token["name"], previous, next)):
+                    yield token
+            elif type == "EndTag":
+                if not self.is_optional_end(token["name"], next):
+                    yield token
+            else:
+                yield token
+
+    def is_optional_start(self, tagname, previous, next):
+        type = next and next["type"] or None
+        if tagname in 'html':
+            # An html element's start tag may be omitted if the first thing
+            # inside the html element is not a space character or a comment.
+            return type not in ("Comment", "SpaceCharacters")
+        elif tagname == 'head':
+            # A head element's start tag may be omitted if the first thing
+            # inside the head element is an element.
+            # XXX: we also omit the start tag if the head element is empty
+            if type in ("StartTag", "EmptyTag"):
+                return True
+            elif type == "EndTag":
+                return next["name"] == "head"
+        elif tagname == 'body':
+            # A body element's start tag may be omitted if the first thing
+            # inside the body element is not a space character or a comment,
+            # except if the first thing inside the body element is a script
+            # or style element and the node immediately preceding the body
+            # element is a head element whose end tag has been omitted.
+            if type in ("Comment", "SpaceCharacters"):
+                return False
+            elif type == "StartTag":
+                # XXX: we do not look at the preceding event, so we never omit
+                # the body element's start tag if it's followed by a script or
+                # a style element.
+                return next["name"] not in ('script', 'style')
+            else:
+                return True
+        elif tagname == 'colgroup':
+            # A colgroup element's start tag may be omitted if the first thing
+            # inside the colgroup element is a col element, and if the element
+            # is not immediately preceeded by another colgroup element whose
+            # end tag has been omitted.
+            if type in ("StartTag", "EmptyTag"):
+                # XXX: we do not look at the preceding event, so instead we never
+                # omit the colgroup element's end tag when it is immediately
+                # followed by another colgroup element. See is_optional_end.
+                return next["name"] == "col"
+            else:
+                return False
+        elif tagname == 'tbody':
+            # A tbody element's start tag may be omitted if the first thing
+            # inside the tbody element is a tr element, and if the element is
+            # not immediately preceeded by a tbody, thead, or tfoot element
+            # whose end tag has been omitted.
+            if type == "StartTag":
+                # omit the thead and tfoot elements' end tag when they are
+                # immediately followed by a tbody element. See is_optional_end.
+                if previous and previous['type'] == 'EndTag' and \
+                        previous['name'] in ('tbody', 'thead', 'tfoot'):
+                    return False
+                return next["name"] == 'tr'
+            else:
+                return False
+        return False
+
+    def is_optional_end(self, tagname, next):
+        type = next and next["type"] or None
+        if tagname in ('html', 'head', 'body'):
+            # An html element's end tag may be omitted if the html element
+            # is not immediately followed by a space character or a comment.
+            return type not in ("Comment", "SpaceCharacters")
+        elif tagname in ('li', 'optgroup', 'tr'):
+            # A li element's end tag may be omitted if the li element is
+            # immediately followed by another li element or if there is
+            # no more content in the parent element.
+            # An optgroup element's end tag may be omitted if the optgroup
+            # element is immediately followed by another optgroup element,
+            # or if there is no more content in the parent element.
+            # A tr element's end tag may be omitted if the tr element is
+            # immediately followed by another tr element, or if there is
+            # no more content in the parent element.
+            if type == "StartTag":
+                return next["name"] == tagname
+            else:
+                return type == "EndTag" or type is None
+        elif tagname in ('dt', 'dd'):
+            # A dt element's end tag may be omitted if the dt element is
+            # immediately followed by another dt element or a dd element.
+            # A dd element's end tag may be omitted if the dd element is
+            # immediately followed by another dd element or a dt element,
+            # or if there is no more content in the parent element.
+            if type == "StartTag":
+                return next["name"] in ('dt', 'dd')
+            elif tagname == 'dd':
+                return type == "EndTag" or type is None
+            else:
+                return False
+        elif tagname == 'p':
+            # A p element's end tag may be omitted if the p element is
+            # immediately followed by an address, article, aside,
+            # blockquote, datagrid, dialog, dir, div, dl, fieldset,
+            # footer, form, h1, h2, h3, h4, h5, h6, header, hr, menu,
+            # nav, ol, p, pre, section, table, or ul, element, or if
+            # there is no more content in the parent element.
+            if type in ("StartTag", "EmptyTag"):
+                return next["name"] in ('address', 'article', 'aside',
+                                        'blockquote', 'datagrid', 'dialog',
+                                        'dir', 'div', 'dl', 'fieldset', 'footer',
+                                        'form', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6',
+                                        'header', 'hr', 'menu', 'nav', 'ol',
+                                        'p', 'pre', 'section', 'table', 'ul')
+            else:
+                return type == "EndTag" or type is None
+        elif tagname == 'option':
+            # An option element's end tag may be omitted if the option
+            # element is immediately followed by another option element,
+            # or if it is immediately followed by an <code>optgroup</code>
+            # element, or if there is no more content in the parent
+            # element.
+            if type == "StartTag":
+                return next["name"] in ('option', 'optgroup')
+            else:
+                return type == "EndTag" or type is None
+        elif tagname in ('rt', 'rp'):
+            # An rt element's end tag may be omitted if the rt element is
+            # immediately followed by an rt or rp element, or if there is
+            # no more content in the parent element.
+            # An rp element's end tag may be omitted if the rp element is
+            # immediately followed by an rt or rp element, or if there is
+            # no more content in the parent element.
+            if type == "StartTag":
+                return next["name"] in ('rt', 'rp')
+            else:
+                return type == "EndTag" or type is None
+        elif tagname == 'colgroup':
+            # A colgroup element's end tag may be omitted if the colgroup
+            # element is not immediately followed by a space character or
+            # a comment.
+            if type in ("Comment", "SpaceCharacters"):
+                return False
+            elif type == "StartTag":
+                # XXX: we also look for an immediately following colgroup
+                # element. See is_optional_start.
+                return next["name"] != 'colgroup'
+            else:
+                return True
+        elif tagname in ('thead', 'tbody'):
+            # A thead element's end tag may be omitted if the thead element
+            # is immediately followed by a tbody or tfoot element.
+            # A tbody element's end tag may be omitted if the tbody element
+            # is immediately followed by a tbody or tfoot element, or if
+            # there is no more content in the parent element.
+            # A tfoot element's end tag may be omitted if the tfoot element
+            # is immediately followed by a tbody element, or if there is no
+            # more content in the parent element.
+            # XXX: we never omit the end tag when the following element is
+            # a tbody. See is_optional_start.
+            if type == "StartTag":
+                return next["name"] in ['tbody', 'tfoot']
+            elif tagname == 'tbody':
+                return type == "EndTag" or type is None
+            else:
+                return False
+        elif tagname == 'tfoot':
+            # A tfoot element's end tag may be omitted if the tfoot element
+            # is immediately followed by a tbody element, or if there is no
+            # more content in the parent element.
+            # XXX: we never omit the end tag when the following element is
+            # a tbody. See is_optional_start.
+            if type == "StartTag":
+                return next["name"] == 'tbody'
+            else:
+                return type == "EndTag" or type is None
+        elif tagname in ('td', 'th'):
+            # A td element's end tag may be omitted if the td element is
+            # immediately followed by a td or th element, or if there is
+            # no more content in the parent element.
+            # A th element's end tag may be omitted if the th element is
+            # immediately followed by a td or th element, or if there is
+            # no more content in the parent element.
+            if type == "StartTag":
+                return next["name"] in ('td', 'th')
+            else:
+                return type == "EndTag" or type is None
+        return False

lib/html5lib/filters/sanitizer.py

@@ -0,0 +1,12 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from . import _base
+from ..sanitizer import HTMLSanitizerMixin
+
+
+class Filter(_base.Filter, HTMLSanitizerMixin):
+    def __iter__(self):
+        for token in _base.Filter.__iter__(self):
+            token = self.sanitize_token(token)
+            if token:
+                yield token

lib/html5lib/filters/whitespace.py

@@ -0,0 +1,38 @@
+from __future__ import absolute_import, division, unicode_literals
+
+import re
+
+from . import _base
+from ..constants import rcdataElements, spaceCharacters
+spaceCharacters = "".join(spaceCharacters)
+
+SPACES_REGEX = re.compile("[%s]+" % spaceCharacters)
+
+
+class Filter(_base.Filter):
+
+    spacePreserveElements = frozenset(["pre", "textarea"] + list(rcdataElements))
+
+    def __iter__(self):
+        preserve = 0
+        for token in _base.Filter.__iter__(self):
+            type = token["type"]
+            if type == "StartTag" \
+                    and (preserve or token["name"] in self.spacePreserveElements):
+                preserve += 1
+
+            elif type == "EndTag" and preserve:
+                preserve -= 1
+
+            elif not preserve and type == "SpaceCharacters" and token["data"]:
+                # Test on token["data"] above to not introduce spaces where there were not
+                token["data"] = " "
+
+            elif not preserve and type == "Characters":
+                token["data"] = collapse_spaces(token["data"])
+
+            yield token
+
+
+def collapse_spaces(text):
+    return SPACES_REGEX.sub(' ', text)

lib/html5lib/ihatexml.py

@@ -0,0 +1,285 @@
+from __future__ import absolute_import, division, unicode_literals
+
+import re
+import warnings
+
+from .constants import DataLossWarning
+
+baseChar = """
+[#x0041-#x005A] | [#x0061-#x007A] | [#x00C0-#x00D6] | [#x00D8-#x00F6] |
+[#x00F8-#x00FF] | [#x0100-#x0131] | [#x0134-#x013E] | [#x0141-#x0148] |
+[#x014A-#x017E] | [#x0180-#x01C3] | [#x01CD-#x01F0] | [#x01F4-#x01F5] |
+[#x01FA-#x0217] | [#x0250-#x02A8] | [#x02BB-#x02C1] | #x0386 |
+[#x0388-#x038A] | #x038C | [#x038E-#x03A1] | [#x03A3-#x03CE] |
+[#x03D0-#x03D6] | #x03DA | #x03DC | #x03DE | #x03E0 | [#x03E2-#x03F3] |
+[#x0401-#x040C] | [#x040E-#x044F] | [#x0451-#x045C] | [#x045E-#x0481] |
+[#x0490-#x04C4] | [#x04C7-#x04C8] | [#x04CB-#x04CC] | [#x04D0-#x04EB] |
+[#x04EE-#x04F5] | [#x04F8-#x04F9] | [#x0531-#x0556] | #x0559 |
+[#x0561-#x0586] | [#x05D0-#x05EA] | [#x05F0-#x05F2] | [#x0621-#x063A] |
+[#x0641-#x064A] | [#x0671-#x06B7] | [#x06BA-#x06BE] | [#x06C0-#x06CE] |
+[#x06D0-#x06D3] | #x06D5 | [#x06E5-#x06E6] | [#x0905-#x0939] | #x093D |
+[#x0958-#x0961] | [#x0985-#x098C] | [#x098F-#x0990] | [#x0993-#x09A8] |
+[#x09AA-#x09B0] | #x09B2 | [#x09B6-#x09B9] | [#x09DC-#x09DD] |
+[#x09DF-#x09E1] | [#x09F0-#x09F1] | [#x0A05-#x0A0A] | [#x0A0F-#x0A10] |
+[#x0A13-#x0A28] | [#x0A2A-#x0A30] | [#x0A32-#x0A33] | [#x0A35-#x0A36] |
+[#x0A38-#x0A39] | [#x0A59-#x0A5C] | #x0A5E | [#x0A72-#x0A74] |
+[#x0A85-#x0A8B] | #x0A8D | [#x0A8F-#x0A91] | [#x0A93-#x0AA8] |
+[#x0AAA-#x0AB0] | [#x0AB2-#x0AB3] | [#x0AB5-#x0AB9] | #x0ABD | #x0AE0 |
+[#x0B05-#x0B0C] | [#x0B0F-#x0B10] | [#x0B13-#x0B28] | [#x0B2A-#x0B30] |
+[#x0B32-#x0B33] | [#x0B36-#x0B39] | #x0B3D | [#x0B5C-#x0B5D] |
+[#x0B5F-#x0B61] | [#x0B85-#x0B8A] | [#x0B8E-#x0B90] | [#x0B92-#x0B95] |
+[#x0B99-#x0B9A] | #x0B9C | [#x0B9E-#x0B9F] | [#x0BA3-#x0BA4] |
+[#x0BA8-#x0BAA] | [#x0BAE-#x0BB5] | [#x0BB7-#x0BB9] | [#x0C05-#x0C0C] |
+[#x0C0E-#x0C10] | [#x0C12-#x0C28] | [#x0C2A-#x0C33] | [#x0C35-#x0C39] |
+[#x0C60-#x0C61] | [#x0C85-#x0C8C] | [#x0C8E-#x0C90] | [#x0C92-#x0CA8] |
+[#x0CAA-#x0CB3] | [#x0CB5-#x0CB9] | #x0CDE | [#x0CE0-#x0CE1] |
+[#x0D05-#x0D0C] | [#x0D0E-#x0D10] | [#x0D12-#x0D28] | [#x0D2A-#x0D39] |
+[#x0D60-#x0D61] | [#x0E01-#x0E2E] | #x0E30 | [#x0E32-#x0E33] |
+[#x0E40-#x0E45] | [#x0E81-#x0E82] | #x0E84 | [#x0E87-#x0E88] | #x0E8A |
+#x0E8D | [#x0E94-#x0E97] | [#x0E99-#x0E9F] | [#x0EA1-#x0EA3] | #x0EA5 |
+#x0EA7 | [#x0EAA-#x0EAB] | [#x0EAD-#x0EAE] | #x0EB0 | [#x0EB2-#x0EB3] |
+#x0EBD | [#x0EC0-#x0EC4] | [#x0F40-#x0F47] | [#x0F49-#x0F69] |
+[#x10A0-#x10C5] | [#x10D0-#x10F6] | #x1100 | [#x1102-#x1103] |
+[#x1105-#x1107] | #x1109 | [#x110B-#x110C] | [#x110E-#x1112] | #x113C |
+#x113E | #x1140 | #x114C | #x114E | #x1150 | [#x1154-#x1155] | #x1159 |
+[#x115F-#x1161] | #x1163 | #x1165 | #x1167 | #x1169 | [#x116D-#x116E] |
+[#x1172-#x1173] | #x1175 | #x119E | #x11A8 | #x11AB | [#x11AE-#x11AF] |
+[#x11B7-#x11B8] | #x11BA | [#x11BC-#x11C2] | #x11EB | #x11F0 | #x11F9 |
+[#x1E00-#x1E9B] | [#x1EA0-#x1EF9] | [#x1F00-#x1F15] | [#x1F18-#x1F1D] |
+[#x1F20-#x1F45] | [#x1F48-#x1F4D] | [#x1F50-#x1F57] | #x1F59 | #x1F5B |
+#x1F5D | [#x1F5F-#x1F7D] | [#x1F80-#x1FB4] | [#x1FB6-#x1FBC] | #x1FBE |
+[#x1FC2-#x1FC4] | [#x1FC6-#x1FCC] | [#x1FD0-#x1FD3] | [#x1FD6-#x1FDB] |
+[#x1FE0-#x1FEC] | [#x1FF2-#x1FF4] | [#x1FF6-#x1FFC] | #x2126 |
+[#x212A-#x212B] | #x212E | [#x2180-#x2182] | [#x3041-#x3094] |
+[#x30A1-#x30FA] | [#x3105-#x312C] | [#xAC00-#xD7A3]"""
+
+ideographic = """[#x4E00-#x9FA5] | #x3007 | [#x3021-#x3029]"""
+
+combiningCharacter = """
+[#x0300-#x0345] | [#x0360-#x0361] | [#x0483-#x0486] | [#x0591-#x05A1] |
+[#x05A3-#x05B9] | [#x05BB-#x05BD] | #x05BF | [#x05C1-#x05C2] | #x05C4 |
+[#x064B-#x0652] | #x0670 | [#x06D6-#x06DC] | [#x06DD-#x06DF] |
+[#x06E0-#x06E4] | [#x06E7-#x06E8] | [#x06EA-#x06ED] | [#x0901-#x0903] |
+#x093C | [#x093E-#x094C] | #x094D | [#x0951-#x0954] | [#x0962-#x0963] |
+[#x0981-#x0983] | #x09BC | #x09BE | #x09BF | [#x09C0-#x09C4] |
+[#x09C7-#x09C8] | [#x09CB-#x09CD] | #x09D7 | [#x09E2-#x09E3] | #x0A02 |
+#x0A3C | #x0A3E | #x0A3F | [#x0A40-#x0A42] | [#x0A47-#x0A48] |
+[#x0A4B-#x0A4D] | [#x0A70-#x0A71] | [#x0A81-#x0A83] | #x0ABC |
+[#x0ABE-#x0AC5] | [#x0AC7-#x0AC9] | [#x0ACB-#x0ACD] | [#x0B01-#x0B03] |
+#x0B3C | [#x0B3E-#x0B43] | [#x0B47-#x0B48] | [#x0B4B-#x0B4D] |
+[#x0B56-#x0B57] | [#x0B82-#x0B83] | [#x0BBE-#x0BC2] | [#x0BC6-#x0BC8] |
+[#x0BCA-#x0BCD] | #x0BD7 | [#x0C01-#x0C03] | [#x0C3E-#x0C44] |
+[#x0C46-#x0C48] | [#x0C4A-#x0C4D] | [#x0C55-#x0C56] | [#x0C82-#x0C83] |
+[#x0CBE-#x0CC4] | [#x0CC6-#x0CC8] | [#x0CCA-#x0CCD] | [#x0CD5-#x0CD6] |
+[#x0D02-#x0D03] | [#x0D3E-#x0D43] | [#x0D46-#x0D48] | [#x0D4A-#x0D4D] |
+#x0D57 | #x0E31 | [#x0E34-#x0E3A] | [#x0E47-#x0E4E] | #x0EB1 |
+[#x0EB4-#x0EB9] | [#x0EBB-#x0EBC] | [#x0EC8-#x0ECD] | [#x0F18-#x0F19] |
+#x0F35 | #x0F37 | #x0F39 | #x0F3E | #x0F3F | [#x0F71-#x0F84] |
+[#x0F86-#x0F8B] | [#x0F90-#x0F95] | #x0F97 | [#x0F99-#x0FAD] |
+[#x0FB1-#x0FB7] | #x0FB9 | [#x20D0-#x20DC] | #x20E1 | [#x302A-#x302F] |
+#x3099 | #x309A"""
+
+digit = """
+[#x0030-#x0039] | [#x0660-#x0669] | [#x06F0-#x06F9] | [#x0966-#x096F] |
+[#x09E6-#x09EF] | [#x0A66-#x0A6F] | [#x0AE6-#x0AEF] | [#x0B66-#x0B6F] |
+[#x0BE7-#x0BEF] | [#x0C66-#x0C6F] | [#x0CE6-#x0CEF] | [#x0D66-#x0D6F] |
+[#x0E50-#x0E59] | [#x0ED0-#x0ED9] | [#x0F20-#x0F29]"""
+
+extender = """
+#x00B7 | #x02D0 | #x02D1 | #x0387 | #x0640 | #x0E46 | #x0EC6 | #x3005 |
+#[#x3031-#x3035] | [#x309D-#x309E] | [#x30FC-#x30FE]"""
+
+letter = " | ".join([baseChar, ideographic])
+
+# Without the
+name = " | ".join([letter, digit, ".", "-", "_", combiningCharacter,
+                   extender])
+nameFirst = " | ".join([letter, "_"])
+
+reChar = re.compile(r"#x([\d|A-F]{4,4})")
+reCharRange = re.compile(r"\[#x([\d|A-F]{4,4})-#x([\d|A-F]{4,4})\]")
+
+
+def charStringToList(chars):
+    charRanges = [item.strip() for item in chars.split(" | ")]
+    rv = []
+    for item in charRanges:
+        foundMatch = False
+        for regexp in (reChar, reCharRange):
+            match = regexp.match(item)
+            if match is not None:
+                rv.append([hexToInt(item) for item in match.groups()])
+                if len(rv[-1]) == 1:
+                    rv[-1] = rv[-1] * 2
+                foundMatch = True
+                break
+        if not foundMatch:
+            assert len(item) == 1
+
+            rv.append([ord(item)] * 2)
+    rv = normaliseCharList(rv)
+    return rv
+
+
+def normaliseCharList(charList):
+    charList = sorted(charList)
+    for item in charList:
+        assert item[1] >= item[0]
+    rv = []
+    i = 0
+    while i < len(charList):
+        j = 1
+        rv.append(charList[i])
+        while i + j < len(charList) and charList[i + j][0] <= rv[-1][1] + 1:
+            rv[-1][1] = charList[i + j][1]
+            j += 1
+        i += j
+    return rv
+
+# We don't really support characters above the BMP :(
+max_unicode = int("FFFF", 16)
+
+
+def missingRanges(charList):
+    rv = []
+    if charList[0] != 0:
+        rv.append([0, charList[0][0] - 1])
+    for i, item in enumerate(charList[:-1]):
+        rv.append([item[1] + 1, charList[i + 1][0] - 1])
+    if charList[-1][1] != max_unicode:
+        rv.append([charList[-1][1] + 1, max_unicode])
+    return rv
+
+
+def listToRegexpStr(charList):
+    rv = []
+    for item in charList:
+        if item[0] == item[1]:
+            rv.append(escapeRegexp(chr(item[0])))
+        else:
+            rv.append(escapeRegexp(chr(item[0])) + "-" +
+                      escapeRegexp(chr(item[1])))
+    return "[%s]" % "".join(rv)
+
+
+def hexToInt(hex_str):
+    return int(hex_str, 16)
+
+
+def escapeRegexp(string):
+    specialCharacters = (".", "^", "$", "*", "+", "?", "{", "}",
+                         "[", "]", "|", "(", ")", "-")
+    for char in specialCharacters:
+        string = string.replace(char, "\\" + char)
+
+    return string
+
+# output from the above
+nonXmlNameBMPRegexp = re.compile('[\x00-,/:-@\\[-\\^`\\{-\xb6\xb8-\xbf\xd7\xf7\u0132-\u0133\u013f-\u0140\u0149\u017f\u01c4-\u01cc\u01f1-\u01f3\u01f6-\u01f9\u0218-\u024f\u02a9-\u02ba\u02c2-\u02cf\u02d2-\u02ff\u0346-\u035f\u0362-\u0385\u038b\u038d\u03a2\u03cf\u03d7-\u03d9\u03db\u03dd\u03df\u03e1\u03f4-\u0400\u040d\u0450\u045d\u0482\u0487-\u048f\u04c5-\u04c6\u04c9-\u04ca\u04cd-\u04cf\u04ec-\u04ed\u04f6-\u04f7\u04fa-\u0530\u0557-\u0558\u055a-\u0560\u0587-\u0590\u05a2\u05ba\u05be\u05c0\u05c3\u05c5-\u05cf\u05eb-\u05ef\u05f3-\u0620\u063b-\u063f\u0653-\u065f\u066a-\u066f\u06b8-\u06b9\u06bf\u06cf\u06d4\u06e9\u06ee-\u06ef\u06fa-\u0900\u0904\u093a-\u093b\u094e-\u0950\u0955-\u0957\u0964-\u0965\u0970-\u0980\u0984\u098d-\u098e\u0991-\u0992\u09a9\u09b1\u09b3-\u09b5\u09ba-\u09bb\u09bd\u09c5-\u09c6\u09c9-\u09ca\u09ce-\u09d6\u09d8-\u09db\u09de\u09e4-\u09e5\u09f2-\u0a01\u0a03-\u0a04\u0a0b-\u0a0e\u0a11-\u0a12\u0a29\u0a31\u0a34\u0a37\u0a3a-\u0a3b\u0a3d\u0a43-\u0a46\u0a49-\u0a4a\u0a4e-\u0a58\u0a5d\u0a5f-\u0a65\u0a75-\u0a80\u0a84\u0a8c\u0a8e\u0a92\u0aa9\u0ab1\u0ab4\u0aba-\u0abb\u0ac6\u0aca\u0ace-\u0adf\u0ae1-\u0ae5\u0af0-\u0b00\u0b04\u0b0d-\u0b0e\u0b11-\u0b12\u0b29\u0b31\u0b34-\u0b35\u0b3a-\u0b3b\u0b44-\u0b46\u0b49-\u0b4a\u0b4e-\u0b55\u0b58-\u0b5b\u0b5e\u0b62-\u0b65\u0b70-\u0b81\u0b84\u0b8b-\u0b8d\u0b91\u0b96-\u0b98\u0b9b\u0b9d\u0ba0-\u0ba2\u0ba5-\u0ba7\u0bab-\u0bad\u0bb6\u0bba-\u0bbd\u0bc3-\u0bc5\u0bc9\u0bce-\u0bd6\u0bd8-\u0be6\u0bf0-\u0c00\u0c04\u0c0d\u0c11\u0c29\u0c34\u0c3a-\u0c3d\u0c45\u0c49\u0c4e-\u0c54\u0c57-\u0c5f\u0c62-\u0c65\u0c70-\u0c81\u0c84\u0c8d\u0c91\u0ca9\u0cb4\u0cba-\u0cbd\u0cc5\u0cc9\u0cce-\u0cd4\u0cd7-\u0cdd\u0cdf\u0ce2-\u0ce5\u0cf0-\u0d01\u0d04\u0d0d\u0d11\u0d29\u0d3a-\u0d3d\u0d44-\u0d45\u0d49\u0d4e-\u0d56\u0d58-\u0d5f\u0d62-\u0d65\u0d70-\u0e00\u0e2f\u0e3b-\u0e3f\u0e4f\u0e5a-\u0e80\u0e83\u0e85-\u0e86\u0e89\u0e8b-\u0e8c\u0e8e-\u0e93\u0e98\u0ea0\u0ea4\u0ea6\u0ea8-\u0ea9\u0eac\u0eaf\u0eba\u0ebe-\u0ebf\u0ec5\u0ec7\u0ece-\u0ecf\u0eda-\u0f17\u0f1a-\u0f1f\u0f2a-\u0f34\u0f36\u0f38\u0f3a-\u0f3d\u0f48\u0f6a-\u0f70\u0f85\u0f8c-\u0f8f\u0f96\u0f98\u0fae-\u0fb0\u0fb8\u0fba-\u109f\u10c6-\u10cf\u10f7-\u10ff\u1101\u1104\u1108\u110a\u110d\u1113-\u113b\u113d\u113f\u1141-\u114b\u114d\u114f\u1151-\u1153\u1156-\u1158\u115a-\u115e\u1162\u1164\u1166\u1168\u116a-\u116c\u116f-\u1171\u1174\u1176-\u119d\u119f-\u11a7\u11a9-\u11aa\u11ac-\u11ad\u11b0-\u11b6\u11b9\u11bb\u11c3-\u11ea\u11ec-\u11ef\u11f1-\u11f8\u11fa-\u1dff\u1e9c-\u1e9f\u1efa-\u1eff\u1f16-\u1f17\u1f1e-\u1f1f\u1f46-\u1f47\u1f4e-\u1f4f\u1f58\u1f5a\u1f5c\u1f5e\u1f7e-\u1f7f\u1fb5\u1fbd\u1fbf-\u1fc1\u1fc5\u1fcd-\u1fcf\u1fd4-\u1fd5\u1fdc-\u1fdf\u1fed-\u1ff1\u1ff5\u1ffd-\u20cf\u20dd-\u20e0\u20e2-\u2125\u2127-\u2129\u212c-\u212d\u212f-\u217f\u2183-\u3004\u3006\u3008-\u3020\u3030\u3036-\u3040\u3095-\u3098\u309b-\u309c\u309f-\u30a0\u30fb\u30ff-\u3104\u312d-\u4dff\u9fa6-\uabff\ud7a4-\uffff]')
+
+nonXmlNameFirstBMPRegexp = re.compile('[\x00-@\\[-\\^`\\{-\xbf\xd7\xf7\u0132-\u0133\u013f-\u0140\u0149\u017f\u01c4-\u01cc\u01f1-\u01f3\u01f6-\u01f9\u0218-\u024f\u02a9-\u02ba\u02c2-\u0385\u0387\u038b\u038d\u03a2\u03cf\u03d7-\u03d9\u03db\u03dd\u03df\u03e1\u03f4-\u0400\u040d\u0450\u045d\u0482-\u048f\u04c5-\u04c6\u04c9-\u04ca\u04cd-\u04cf\u04ec-\u04ed\u04f6-\u04f7\u04fa-\u0530\u0557-\u0558\u055a-\u0560\u0587-\u05cf\u05eb-\u05ef\u05f3-\u0620\u063b-\u0640\u064b-\u0670\u06b8-\u06b9\u06bf\u06cf\u06d4\u06d6-\u06e4\u06e7-\u0904\u093a-\u093c\u093e-\u0957\u0962-\u0984\u098d-\u098e\u0991-\u0992\u09a9\u09b1\u09b3-\u09b5\u09ba-\u09db\u09de\u09e2-\u09ef\u09f2-\u0a04\u0a0b-\u0a0e\u0a11-\u0a12\u0a29\u0a31\u0a34\u0a37\u0a3a-\u0a58\u0a5d\u0a5f-\u0a71\u0a75-\u0a84\u0a8c\u0a8e\u0a92\u0aa9\u0ab1\u0ab4\u0aba-\u0abc\u0abe-\u0adf\u0ae1-\u0b04\u0b0d-\u0b0e\u0b11-\u0b12\u0b29\u0b31\u0b34-\u0b35\u0b3a-\u0b3c\u0b3e-\u0b5b\u0b5e\u0b62-\u0b84\u0b8b-\u0b8d\u0b91\u0b96-\u0b98\u0b9b\u0b9d\u0ba0-\u0ba2\u0ba5-\u0ba7\u0bab-\u0bad\u0bb6\u0bba-\u0c04\u0c0d\u0c11\u0c29\u0c34\u0c3a-\u0c5f\u0c62-\u0c84\u0c8d\u0c91\u0ca9\u0cb4\u0cba-\u0cdd\u0cdf\u0ce2-\u0d04\u0d0d\u0d11\u0d29\u0d3a-\u0d5f\u0d62-\u0e00\u0e2f\u0e31\u0e34-\u0e3f\u0e46-\u0e80\u0e83\u0e85-\u0e86\u0e89\u0e8b-\u0e8c\u0e8e-\u0e93\u0e98\u0ea0\u0ea4\u0ea6\u0ea8-\u0ea9\u0eac\u0eaf\u0eb1\u0eb4-\u0ebc\u0ebe-\u0ebf\u0ec5-\u0f3f\u0f48\u0f6a-\u109f\u10c6-\u10cf\u10f7-\u10ff\u1101\u1104\u1108\u110a\u110d\u1113-\u113b\u113d\u113f\u1141-\u114b\u114d\u114f\u1151-\u1153\u1156-\u1158\u115a-\u115e\u1162\u1164\u1166\u1168\u116a-\u116c\u116f-\u1171\u1174\u1176-\u119d\u119f-\u11a7\u11a9-\u11aa\u11ac-\u11ad\u11b0-\u11b6\u11b9\u11bb\u11c3-\u11ea\u11ec-\u11ef\u11f1-\u11f8\u11fa-\u1dff\u1e9c-\u1e9f\u1efa-\u1eff\u1f16-\u1f17\u1f1e-\u1f1f\u1f46-\u1f47\u1f4e-\u1f4f\u1f58\u1f5a\u1f5c\u1f5e\u1f7e-\u1f7f\u1fb5\u1fbd\u1fbf-\u1fc1\u1fc5\u1fcd-\u1fcf\u1fd4-\u1fd5\u1fdc-\u1fdf\u1fed-\u1ff1\u1ff5\u1ffd-\u2125\u2127-\u2129\u212c-\u212d\u212f-\u217f\u2183-\u3006\u3008-\u3020\u302a-\u3040\u3095-\u30a0\u30fb-\u3104\u312d-\u4dff\u9fa6-\uabff\ud7a4-\uffff]')
+
+# Simpler things
+nonPubidCharRegexp = re.compile("[^\x20\x0D\x0Aa-zA-Z0-9\-\'()+,./:=?;!*#@$_%]")
+
+
+class InfosetFilter(object):
+    replacementRegexp = re.compile(r"U[\dA-F]{5,5}")
+
+    def __init__(self, replaceChars=None,
+                 dropXmlnsLocalName=False,
+                 dropXmlnsAttrNs=False,
+                 preventDoubleDashComments=False,
+                 preventDashAtCommentEnd=False,
+                 replaceFormFeedCharacters=True,
+                 preventSingleQuotePubid=False):
+
+        self.dropXmlnsLocalName = dropXmlnsLocalName
+        self.dropXmlnsAttrNs = dropXmlnsAttrNs
+
+        self.preventDoubleDashComments = preventDoubleDashComments
+        self.preventDashAtCommentEnd = preventDashAtCommentEnd
+
+        self.replaceFormFeedCharacters = replaceFormFeedCharacters
+
+        self.preventSingleQuotePubid = preventSingleQuotePubid
+
+        self.replaceCache = {}
+
+    def coerceAttribute(self, name, namespace=None):
+        if self.dropXmlnsLocalName and name.startswith("xmlns:"):
+            warnings.warn("Attributes cannot begin with xmlns", DataLossWarning)
+            return None
+        elif (self.dropXmlnsAttrNs and
+              namespace == "http://www.w3.org/2000/xmlns/"):
+            warnings.warn("Attributes cannot be in the xml namespace", DataLossWarning)
+            return None
+        else:
+            return self.toXmlName(name)
+
+    def coerceElement(self, name, namespace=None):
+        return self.toXmlName(name)
+
+    def coerceComment(self, data):
+        if self.preventDoubleDashComments:
+            while "--" in data:
+                warnings.warn("Comments cannot contain adjacent dashes", DataLossWarning)
+                data = data.replace("--", "- -")
+        return data
+
+    def coerceCharacters(self, data):
+        if self.replaceFormFeedCharacters:
+            for i in range(data.count("\x0C")):
+                warnings.warn("Text cannot contain U+000C", DataLossWarning)
+            data = data.replace("\x0C", " ")
+        # Other non-xml characters
+        return data
+
+    def coercePubid(self, data):
+        dataOutput = data
+        for char in nonPubidCharRegexp.findall(data):
+            warnings.warn("Coercing non-XML pubid", DataLossWarning)
+            replacement = self.getReplacementCharacter(char)
+            dataOutput = dataOutput.replace(char, replacement)
+        if self.preventSingleQuotePubid and dataOutput.find("'") >= 0:
+            warnings.warn("Pubid cannot contain single quote", DataLossWarning)
+            dataOutput = dataOutput.replace("'", self.getReplacementCharacter("'"))
+        return dataOutput
+
+    def toXmlName(self, name):
+        nameFirst = name[0]
+        nameRest = name[1:]
+        m = nonXmlNameFirstBMPRegexp.match(nameFirst)
+        if m:
+            warnings.warn("Coercing non-XML name", DataLossWarning)
+            nameFirstOutput = self.getReplacementCharacter(nameFirst)
+        else:
+            nameFirstOutput = nameFirst
+
+        nameRestOutput = nameRest
+        replaceChars = set(nonXmlNameBMPRegexp.findall(nameRest))
+        for char in replaceChars:
+            warnings.warn("Coercing non-XML name", DataLossWarning)
+            replacement = self.getReplacementCharacter(char)
+            nameRestOutput = nameRestOutput.replace(char, replacement)
+        return nameFirstOutput + nameRestOutput
+
+    def getReplacementCharacter(self, char):
+        if char in self.replaceCache:
+            replacement = self.replaceCache[char]
+        else:
+            replacement = self.escapeChar(char)
+        return replacement
+
+    def fromXmlName(self, name):
+        for item in set(self.replacementRegexp.findall(name)):
+            name = name.replace(item, self.unescapeChar(item))
+        return name
+
+    def escapeChar(self, char):
+        replacement = "U%05X" % ord(char)
+        self.replaceCache[char] = replacement
+        return replacement
+
+    def unescapeChar(self, charcode):
+        return chr(int(charcode[1:], 16))

lib/html5lib/inputstream.py

@@ -0,0 +1,886 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+from six.moves import http_client
+
+import codecs
+import re
+
+from .constants import EOF, spaceCharacters, asciiLetters, asciiUppercase
+from .constants import encodings, ReparseException
+from . import utils
+
+from io import StringIO
+
+try:
+    from io import BytesIO
+except ImportError:
+    BytesIO = StringIO
+
+try:
+    from io import BufferedIOBase
+except ImportError:
+    class BufferedIOBase(object):
+        pass
+
+# Non-unicode versions of constants for use in the pre-parser
+spaceCharactersBytes = frozenset([item.encode("ascii") for item in spaceCharacters])
+asciiLettersBytes = frozenset([item.encode("ascii") for item in asciiLetters])
+asciiUppercaseBytes = frozenset([item.encode("ascii") for item in asciiUppercase])
+spacesAngleBrackets = spaceCharactersBytes | frozenset([b">", b"<"])
+
+invalid_unicode_re = re.compile("[\u0001-\u0008\u000B\u000E-\u001F\u007F-\u009F\uD800-\uDFFF\uFDD0-\uFDEF\uFFFE\uFFFF\U0001FFFE\U0001FFFF\U0002FFFE\U0002FFFF\U0003FFFE\U0003FFFF\U0004FFFE\U0004FFFF\U0005FFFE\U0005FFFF\U0006FFFE\U0006FFFF\U0007FFFE\U0007FFFF\U0008FFFE\U0008FFFF\U0009FFFE\U0009FFFF\U000AFFFE\U000AFFFF\U000BFFFE\U000BFFFF\U000CFFFE\U000CFFFF\U000DFFFE\U000DFFFF\U000EFFFE\U000EFFFF\U000FFFFE\U000FFFFF\U0010FFFE\U0010FFFF]")
+
+non_bmp_invalid_codepoints = set([0x1FFFE, 0x1FFFF, 0x2FFFE, 0x2FFFF, 0x3FFFE,
+                                  0x3FFFF, 0x4FFFE, 0x4FFFF, 0x5FFFE, 0x5FFFF,
+                                  0x6FFFE, 0x6FFFF, 0x7FFFE, 0x7FFFF, 0x8FFFE,
+                                  0x8FFFF, 0x9FFFE, 0x9FFFF, 0xAFFFE, 0xAFFFF,
+                                  0xBFFFE, 0xBFFFF, 0xCFFFE, 0xCFFFF, 0xDFFFE,
+                                  0xDFFFF, 0xEFFFE, 0xEFFFF, 0xFFFFE, 0xFFFFF,
+                                  0x10FFFE, 0x10FFFF])
+
+ascii_punctuation_re = re.compile("[\u0009-\u000D\u0020-\u002F\u003A-\u0040\u005B-\u0060\u007B-\u007E]")
+
+# Cache for charsUntil()
+charsUntilRegEx = {}
+
+
+class BufferedStream(object):
+    """Buffering for streams that do not have buffering of their own
+
+    The buffer is implemented as a list of chunks on the assumption that
+    joining many strings will be slow since it is O(n**2)
+    """
+
+    def __init__(self, stream):
+        self.stream = stream
+        self.buffer = []
+        self.position = [-1, 0]  # chunk number, offset
+
+    def tell(self):
+        pos = 0
+        for chunk in self.buffer[:self.position[0]]:
+            pos += len(chunk)
+        pos += self.position[1]
+        return pos
+
+    def seek(self, pos):
+        assert pos <= self._bufferedBytes()
+        offset = pos
+        i = 0
+        while len(self.buffer[i]) < offset:
+            offset -= len(self.buffer[i])
+            i += 1
+        self.position = [i, offset]
+
+    def read(self, bytes):
+        if not self.buffer:
+            return self._readStream(bytes)
+        elif (self.position[0] == len(self.buffer) and
+              self.position[1] == len(self.buffer[-1])):
+            return self._readStream(bytes)
+        else:
+            return self._readFromBuffer(bytes)
+
+    def _bufferedBytes(self):
+        return sum([len(item) for item in self.buffer])
+
+    def _readStream(self, bytes):
+        data = self.stream.read(bytes)
+        self.buffer.append(data)
+        self.position[0] += 1
+        self.position[1] = len(data)
+        return data
+
+    def _readFromBuffer(self, bytes):
+        remainingBytes = bytes
+        rv = []
+        bufferIndex = self.position[0]
+        bufferOffset = self.position[1]
+        while bufferIndex < len(self.buffer) and remainingBytes != 0:
+            assert remainingBytes > 0
+            bufferedData = self.buffer[bufferIndex]
+
+            if remainingBytes <= len(bufferedData) - bufferOffset:
+                bytesToRead = remainingBytes
+                self.position = [bufferIndex, bufferOffset + bytesToRead]
+            else:
+                bytesToRead = len(bufferedData) - bufferOffset
+                self.position = [bufferIndex, len(bufferedData)]
+                bufferIndex += 1
+            rv.append(bufferedData[bufferOffset:bufferOffset + bytesToRead])
+            remainingBytes -= bytesToRead
+
+            bufferOffset = 0
+
+        if remainingBytes:
+            rv.append(self._readStream(remainingBytes))
+
+        return b"".join(rv)
+
+
+def HTMLInputStream(source, encoding=None, parseMeta=True, chardet=True):
+    if isinstance(source, http_client.HTTPResponse):
+        # Work around Python bug #20007: read(0) closes the connection.
+        # http://bugs.python.org/issue20007
+        isUnicode = False
+    elif hasattr(source, "read"):
+        isUnicode = isinstance(source.read(0), text_type)
+    else:
+        isUnicode = isinstance(source, text_type)
+
+    if isUnicode:
+        if encoding is not None:
+            raise TypeError("Cannot explicitly set an encoding with a unicode string")
+
+        return HTMLUnicodeInputStream(source)
+    else:
+        return HTMLBinaryInputStream(source, encoding, parseMeta, chardet)
+
+
+class HTMLUnicodeInputStream(object):
+    """Provides a unicode stream of characters to the HTMLTokenizer.
+
+    This class takes care of character encoding and removing or replacing
+    incorrect byte-sequences and also provides column and line tracking.
+
+    """
+
+    _defaultChunkSize = 10240
+
+    def __init__(self, source):
+        """Initialises the HTMLInputStream.
+
+        HTMLInputStream(source, [encoding]) -> Normalized stream from source
+        for use by html5lib.
+
+        source can be either a file-object, local filename or a string.
+
+        The optional encoding parameter must be a string that indicates
+        the encoding.  If specified, that encoding will be used,
+        regardless of any BOM or later declaration (such as in a meta
+        element)
+
+        parseMeta - Look for a <meta> element containing encoding information
+
+        """
+
+        # Craziness
+        if len("\U0010FFFF") == 1:
+            self.reportCharacterErrors = self.characterErrorsUCS4
+            self.replaceCharactersRegexp = re.compile("[\uD800-\uDFFF]")
+        else:
+            self.reportCharacterErrors = self.characterErrorsUCS2
+            self.replaceCharactersRegexp = re.compile("([\uD800-\uDBFF](?![\uDC00-\uDFFF])|(?<![\uD800-\uDBFF])[\uDC00-\uDFFF])")
+
+        # List of where new lines occur
+        self.newLines = [0]
+
+        self.charEncoding = ("utf-8", "certain")
+        self.dataStream = self.openStream(source)
+
+        self.reset()
+
+    def reset(self):
+        self.chunk = ""
+        self.chunkSize = 0
+        self.chunkOffset = 0
+        self.errors = []
+
+        # number of (complete) lines in previous chunks
+        self.prevNumLines = 0
+        # number of columns in the last line of the previous chunk
+        self.prevNumCols = 0
+
+        # Deal with CR LF and surrogates split over chunk boundaries
+        self._bufferedCharacter = None
+
+    def openStream(self, source):
+        """Produces a file object from source.
+
+        source can be either a file object, local filename or a string.
+
+        """
+        # Already a file object
+        if hasattr(source, 'read'):
+            stream = source
+        else:
+            stream = StringIO(source)
+
+        return stream
+
+    def _position(self, offset):
+        chunk = self.chunk
+        nLines = chunk.count('\n', 0, offset)
+        positionLine = self.prevNumLines + nLines
+        lastLinePos = chunk.rfind('\n', 0, offset)
+        if lastLinePos == -1:
+            positionColumn = self.prevNumCols + offset
+        else:
+            positionColumn = offset - (lastLinePos + 1)
+        return (positionLine, positionColumn)
+
+    def position(self):
+        """Returns (line, col) of the current position in the stream."""
+        line, col = self._position(self.chunkOffset)
+        return (line + 1, col)
+
+    def char(self):
+        """ Read one character from the stream or queue if available. Return
+            EOF when EOF is reached.
+        """
+        # Read a new chunk from the input stream if necessary
+        if self.chunkOffset >= self.chunkSize:
+            if not self.readChunk():
+                return EOF
+
+        chunkOffset = self.chunkOffset
+        char = self.chunk[chunkOffset]
+        self.chunkOffset = chunkOffset + 1
+
+        return char
+
+    def readChunk(self, chunkSize=None):
+        if chunkSize is None:
+            chunkSize = self._defaultChunkSize
+
+        self.prevNumLines, self.prevNumCols = self._position(self.chunkSize)
+
+        self.chunk = ""
+        self.chunkSize = 0
+        self.chunkOffset = 0
+
+        data = self.dataStream.read(chunkSize)
+
+        # Deal with CR LF and surrogates broken across chunks
+        if self._bufferedCharacter:
+            data = self._bufferedCharacter + data
+            self._bufferedCharacter = None
+        elif not data:
+            # We have no more data, bye-bye stream
+            return False
+
+        if len(data) > 1:
+            lastv = ord(data[-1])
+            if lastv == 0x0D or 0xD800 <= lastv <= 0xDBFF:
+                self._bufferedCharacter = data[-1]
+                data = data[:-1]
+
+        self.reportCharacterErrors(data)
+
+        # Replace invalid characters
+        # Note U+0000 is dealt with in the tokenizer
+        data = self.replaceCharactersRegexp.sub("\ufffd", data)
+
+        data = data.replace("\r\n", "\n")
+        data = data.replace("\r", "\n")
+
+        self.chunk = data
+        self.chunkSize = len(data)
+
+        return True
+
+    def characterErrorsUCS4(self, data):
+        for i in range(len(invalid_unicode_re.findall(data))):
+            self.errors.append("invalid-codepoint")
+
+    def characterErrorsUCS2(self, data):
+        # Someone picked the wrong compile option
+        # You lose
+        skip = False
+        for match in invalid_unicode_re.finditer(data):
+            if skip:
+                continue
+            codepoint = ord(match.group())
+            pos = match.start()
+            # Pretty sure there should be endianness issues here
+            if utils.isSurrogatePair(data[pos:pos + 2]):
+                # We have a surrogate pair!
+                char_val = utils.surrogatePairToCodepoint(data[pos:pos + 2])
+                if char_val in non_bmp_invalid_codepoints:
+                    self.errors.append("invalid-codepoint")
+                skip = True
+            elif (codepoint >= 0xD800 and codepoint <= 0xDFFF and
+                  pos == len(data) - 1):
+                self.errors.append("invalid-codepoint")
+            else:
+                skip = False
+                self.errors.append("invalid-codepoint")
+
+    def charsUntil(self, characters, opposite=False):
+        """ Returns a string of characters from the stream up to but not
+        including any character in 'characters' or EOF. 'characters' must be
+        a container that supports the 'in' method and iteration over its
+        characters.
+        """
+
+        # Use a cache of regexps to find the required characters
+        try:
+            chars = charsUntilRegEx[(characters, opposite)]
+        except KeyError:
+            if __debug__:
+                for c in characters:
+                    assert(ord(c) < 128)
+            regex = "".join(["\\x%02x" % ord(c) for c in characters])
+            if not opposite:
+                regex = "^%s" % regex
+            chars = charsUntilRegEx[(characters, opposite)] = re.compile("[%s]+" % regex)
+
+        rv = []
+
+        while True:
+            # Find the longest matching prefix
+            m = chars.match(self.chunk, self.chunkOffset)
+            if m is None:
+                # If nothing matched, and it wasn't because we ran out of chunk,
+                # then stop
+                if self.chunkOffset != self.chunkSize:
+                    break
+            else:
+                end = m.end()
+                # If not the whole chunk matched, return everything
+                # up to the part that didn't match
+                if end != self.chunkSize:
+                    rv.append(self.chunk[self.chunkOffset:end])
+                    self.chunkOffset = end
+                    break
+            # If the whole remainder of the chunk matched,
+            # use it all and read the next chunk
+            rv.append(self.chunk[self.chunkOffset:])
+            if not self.readChunk():
+                # Reached EOF
+                break
+
+        r = "".join(rv)
+        return r
+
+    def unget(self, char):
+        # Only one character is allowed to be ungotten at once - it must
+        # be consumed again before any further call to unget
+        if char is not None:
+            if self.chunkOffset == 0:
+                # unget is called quite rarely, so it's a good idea to do
+                # more work here if it saves a bit of work in the frequently
+                # called char and charsUntil.
+                # So, just prepend the ungotten character onto the current
+                # chunk:
+                self.chunk = char + self.chunk
+                self.chunkSize += 1
+            else:
+                self.chunkOffset -= 1
+                assert self.chunk[self.chunkOffset] == char
+
+
+class HTMLBinaryInputStream(HTMLUnicodeInputStream):
+    """Provides a unicode stream of characters to the HTMLTokenizer.
+
+    This class takes care of character encoding and removing or replacing
+    incorrect byte-sequences and also provides column and line tracking.
+
+    """
+
+    def __init__(self, source, encoding=None, parseMeta=True, chardet=True):
+        """Initialises the HTMLInputStream.
+
+        HTMLInputStream(source, [encoding]) -> Normalized stream from source
+        for use by html5lib.
+
+        source can be either a file-object, local filename or a string.
+
+        The optional encoding parameter must be a string that indicates
+        the encoding.  If specified, that encoding will be used,
+        regardless of any BOM or later declaration (such as in a meta
+        element)
+
+        parseMeta - Look for a <meta> element containing encoding information
+
+        """
+        # Raw Stream - for unicode objects this will encode to utf-8 and set
+        #              self.charEncoding as appropriate
+        self.rawStream = self.openStream(source)
+
+        HTMLUnicodeInputStream.__init__(self, self.rawStream)
+
+        self.charEncoding = (codecName(encoding), "certain")
+
+        # Encoding Information
+        # Number of bytes to use when looking for a meta element with
+        # encoding information
+        self.numBytesMeta = 512
+        # Number of bytes to use when using detecting encoding using chardet
+        self.numBytesChardet = 100
+        # Encoding to use if no other information can be found
+        self.defaultEncoding = "windows-1252"
+
+        # Detect encoding iff no explicit "transport level" encoding is supplied
+        if (self.charEncoding[0] is None):
+            self.charEncoding = self.detectEncoding(parseMeta, chardet)
+
+        # Call superclass
+        self.reset()
+
+    def reset(self):
+        self.dataStream = codecs.getreader(self.charEncoding[0])(self.rawStream,
+                                                                 'replace')
+        HTMLUnicodeInputStream.reset(self)
+
+    def openStream(self, source):
+        """Produces a file object from source.
+
+        source can be either a file object, local filename or a string.
+
+        """
+        # Already a file object
+        if hasattr(source, 'read'):
+            stream = source
+        else:
+            stream = BytesIO(source)
+
+        try:
+            stream.seek(stream.tell())
+        except:
+            stream = BufferedStream(stream)
+
+        return stream
+
+    def detectEncoding(self, parseMeta=True, chardet=True):
+        # First look for a BOM
+        # This will also read past the BOM if present
+        encoding = self.detectBOM()
+        confidence = "certain"
+        # If there is no BOM need to look for meta elements with encoding
+        # information
+        if encoding is None and parseMeta:
+            encoding = self.detectEncodingMeta()
+            confidence = "tentative"
+        # Guess with chardet, if avaliable
+        if encoding is None and chardet:
+            confidence = "tentative"
+            try:
+                try:
+                    from charade.universaldetector import UniversalDetector
+                except ImportError:
+                    from chardet.universaldetector import UniversalDetector
+                buffers = []
+                detector = UniversalDetector()
+                while not detector.done:
+                    buffer = self.rawStream.read(self.numBytesChardet)
+                    assert isinstance(buffer, bytes)
+                    if not buffer:
+                        break
+                    buffers.append(buffer)
+                    detector.feed(buffer)
+                detector.close()
+                encoding = detector.result['encoding']
+                self.rawStream.seek(0)
+            except ImportError:
+                pass
+        # If all else fails use the default encoding
+        if encoding is None:
+            confidence = "tentative"
+            encoding = self.defaultEncoding
+
+        # Substitute for equivalent encodings:
+        encodingSub = {"iso-8859-1": "windows-1252"}
+
+        if encoding.lower() in encodingSub:
+            encoding = encodingSub[encoding.lower()]
+
+        return encoding, confidence
+
+    def changeEncoding(self, newEncoding):
+        assert self.charEncoding[1] != "certain"
+        newEncoding = codecName(newEncoding)
+        if newEncoding in ("utf-16", "utf-16-be", "utf-16-le"):
+            newEncoding = "utf-8"
+        if newEncoding is None:
+            return
+        elif newEncoding == self.charEncoding[0]:
+            self.charEncoding = (self.charEncoding[0], "certain")
+        else:
+            self.rawStream.seek(0)
+            self.reset()
+            self.charEncoding = (newEncoding, "certain")
+            raise ReparseException("Encoding changed from %s to %s" % (self.charEncoding[0], newEncoding))
+
+    def detectBOM(self):
+        """Attempts to detect at BOM at the start of the stream. If
+        an encoding can be determined from the BOM return the name of the
+        encoding otherwise return None"""
+        bomDict = {
+            codecs.BOM_UTF8: 'utf-8',
+            codecs.BOM_UTF16_LE: 'utf-16-le', codecs.BOM_UTF16_BE: 'utf-16-be',
+            codecs.BOM_UTF32_LE: 'utf-32-le', codecs.BOM_UTF32_BE: 'utf-32-be'
+        }
+
+        # Go to beginning of file and read in 4 bytes
+        string = self.rawStream.read(4)
+        assert isinstance(string, bytes)
+
+        # Try detecting the BOM using bytes from the string
+        encoding = bomDict.get(string[:3])         # UTF-8
+        seek = 3
+        if not encoding:
+            # Need to detect UTF-32 before UTF-16
+            encoding = bomDict.get(string)         # UTF-32
+            seek = 4
+            if not encoding:
+                encoding = bomDict.get(string[:2])  # UTF-16
+                seek = 2
+
+        # Set the read position past the BOM if one was found, otherwise
+        # set it to the start of the stream
+        self.rawStream.seek(encoding and seek or 0)
+
+        return encoding
+
+    def detectEncodingMeta(self):
+        """Report the encoding declared by the meta element
+        """
+        buffer = self.rawStream.read(self.numBytesMeta)
+        assert isinstance(buffer, bytes)
+        parser = EncodingParser(buffer)
+        self.rawStream.seek(0)
+        encoding = parser.getEncoding()
+
+        if encoding in ("utf-16", "utf-16-be", "utf-16-le"):
+            encoding = "utf-8"
+
+        return encoding
+
+
+class EncodingBytes(bytes):
+    """String-like object with an associated position and various extra methods
+    If the position is ever greater than the string length then an exception is
+    raised"""
+    def __new__(self, value):
+        assert isinstance(value, bytes)
+        return bytes.__new__(self, value.lower())
+
+    def __init__(self, value):
+        self._position = -1
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        p = self._position = self._position + 1
+        if p >= len(self):
+            raise StopIteration
+        elif p < 0:
+            raise TypeError
+        return self[p:p + 1]
+
+    def next(self):
+        # Py2 compat
+        return self.__next__()
+
+    def previous(self):
+        p = self._position
+        if p >= len(self):
+            raise StopIteration
+        elif p < 0:
+            raise TypeError
+        self._position = p = p - 1
+        return self[p:p + 1]
+
+    def setPosition(self, position):
+        if self._position >= len(self):
+            raise StopIteration
+        self._position = position
+
+    def getPosition(self):
+        if self._position >= len(self):
+            raise StopIteration
+        if self._position >= 0:
+            return self._position
+        else:
+            return None
+
+    position = property(getPosition, setPosition)
+
+    def getCurrentByte(self):
+        return self[self.position:self.position + 1]
+
+    currentByte = property(getCurrentByte)
+
+    def skip(self, chars=spaceCharactersBytes):
+        """Skip past a list of characters"""
+        p = self.position               # use property for the error-checking
+        while p < len(self):
+            c = self[p:p + 1]
+            if c not in chars:
+                self._position = p
+                return c
+            p += 1
+        self._position = p
+        return None
+
+    def skipUntil(self, chars):
+        p = self.position
+        while p < len(self):
+            c = self[p:p + 1]
+            if c in chars:
+                self._position = p
+                return c
+            p += 1
+        self._position = p
+        return None
+
+    def matchBytes(self, bytes):
+        """Look for a sequence of bytes at the start of a string. If the bytes
+        are found return True and advance the position to the byte after the
+        match. Otherwise return False and leave the position alone"""
+        p = self.position
+        data = self[p:p + len(bytes)]
+        rv = data.startswith(bytes)
+        if rv:
+            self.position += len(bytes)
+        return rv
+
+    def jumpTo(self, bytes):
+        """Look for the next sequence of bytes matching a given sequence. If
+        a match is found advance the position to the last byte of the match"""
+        newPosition = self[self.position:].find(bytes)
+        if newPosition > -1:
+            # XXX: This is ugly, but I can't see a nicer way to fix this.
+            if self._position == -1:
+                self._position = 0
+            self._position += (newPosition + len(bytes) - 1)
+            return True
+        else:
+            raise StopIteration
+
+
+class EncodingParser(object):
+    """Mini parser for detecting character encoding from meta elements"""
+
+    def __init__(self, data):
+        """string - the data to work on for encoding detection"""
+        self.data = EncodingBytes(data)
+        self.encoding = None
+
+    def getEncoding(self):
+        methodDispatch = (
+            (b"<!--", self.handleComment),
+            (b"<meta", self.handleMeta),
+            (b"</", self.handlePossibleEndTag),
+            (b"<!", self.handleOther),
+            (b"<?", self.handleOther),
+            (b"<", self.handlePossibleStartTag))
+        for byte in self.data:
+            keepParsing = True
+            for key, method in methodDispatch:
+                if self.data.matchBytes(key):
+                    try:
+                        keepParsing = method()
+                        break
+                    except StopIteration:
+                        keepParsing = False
+                        break
+            if not keepParsing:
+                break
+
+        return self.encoding
+
+    def handleComment(self):
+        """Skip over comments"""
+        return self.data.jumpTo(b"-->")
+
+    def handleMeta(self):
+        if self.data.currentByte not in spaceCharactersBytes:
+            # if we have <meta not followed by a space so just keep going
+            return True
+        # We have a valid meta element we want to search for attributes
+        hasPragma = False
+        pendingEncoding = None
+        while True:
+            # Try to find the next attribute after the current position
+            attr = self.getAttribute()
+            if attr is None:
+                return True
+            else:
+                if attr[0] == b"http-equiv":
+                    hasPragma = attr[1] == b"content-type"
+                    if hasPragma and pendingEncoding is not None:
+                        self.encoding = pendingEncoding
+                        return False
+                elif attr[0] == b"charset":
+                    tentativeEncoding = attr[1]
+                    codec = codecName(tentativeEncoding)
+                    if codec is not None:
+                        self.encoding = codec
+                        return False
+                elif attr[0] == b"content":
+                    contentParser = ContentAttrParser(EncodingBytes(attr[1]))
+                    tentativeEncoding = contentParser.parse()
+                    if tentativeEncoding is not None:
+                        codec = codecName(tentativeEncoding)
+                        if codec is not None:
+                            if hasPragma:
+                                self.encoding = codec
+                                return False
+                            else:
+                                pendingEncoding = codec
+
+    def handlePossibleStartTag(self):
+        return self.handlePossibleTag(False)
+
+    def handlePossibleEndTag(self):
+        next(self.data)
+        return self.handlePossibleTag(True)
+
+    def handlePossibleTag(self, endTag):
+        data = self.data
+        if data.currentByte not in asciiLettersBytes:
+            # If the next byte is not an ascii letter either ignore this
+            # fragment (possible start tag case) or treat it according to
+            # handleOther
+            if endTag:
+                data.previous()
+                self.handleOther()
+            return True
+
+        c = data.skipUntil(spacesAngleBrackets)
+        if c == b"<":
+            # return to the first step in the overall "two step" algorithm
+            # reprocessing the < byte
+            data.previous()
+        else:
+            # Read all attributes
+            attr = self.getAttribute()
+            while attr is not None:
+                attr = self.getAttribute()
+        return True
+
+    def handleOther(self):
+        return self.data.jumpTo(b">")
+
+    def getAttribute(self):
+        """Return a name,value pair for the next attribute in the stream,
+        if one is found, or None"""
+        data = self.data
+        # Step 1 (skip chars)
+        c = data.skip(spaceCharactersBytes | frozenset([b"/"]))
+        assert c is None or len(c) == 1
+        # Step 2
+        if c in (b">", None):
+            return None
+        # Step 3
+        attrName = []
+        attrValue = []
+        # Step 4 attribute name
+        while True:
+            if c == b"=" and attrName:
+                break
+            elif c in spaceCharactersBytes:
+                # Step 6!
+                c = data.skip()
+                break
+            elif c in (b"/", b">"):
+                return b"".join(attrName), b""
+            elif c in asciiUppercaseBytes:
+                attrName.append(c.lower())
+            elif c is None:
+                return None
+            else:
+                attrName.append(c)
+            # Step 5
+            c = next(data)
+        # Step 7
+        if c != b"=":
+            data.previous()
+            return b"".join(attrName), b""
+        # Step 8
+        next(data)
+        # Step 9
+        c = data.skip()
+        # Step 10
+        if c in (b"'", b'"'):
+            # 10.1
+            quoteChar = c
+            while True:
+                # 10.2
+                c = next(data)
+                # 10.3
+                if c == quoteChar:
+                    next(data)
+                    return b"".join(attrName), b"".join(attrValue)
+                # 10.4
+                elif c in asciiUppercaseBytes:
+                    attrValue.append(c.lower())
+                # 10.5
+                else:
+                    attrValue.append(c)
+        elif c == b">":
+            return b"".join(attrName), b""
+        elif c in asciiUppercaseBytes:
+            attrValue.append(c.lower())
+        elif c is None:
+            return None
+        else:
+            attrValue.append(c)
+        # Step 11
+        while True:
+            c = next(data)
+            if c in spacesAngleBrackets:
+                return b"".join(attrName), b"".join(attrValue)
+            elif c in asciiUppercaseBytes:
+                attrValue.append(c.lower())
+            elif c is None:
+                return None
+            else:
+                attrValue.append(c)
+
+
+class ContentAttrParser(object):
+    def __init__(self, data):
+        assert isinstance(data, bytes)
+        self.data = data
+
+    def parse(self):
+        try:
+            # Check if the attr name is charset
+            # otherwise return
+            self.data.jumpTo(b"charset")
+            self.data.position += 1
+            self.data.skip()
+            if not self.data.currentByte == b"=":
+                # If there is no = sign keep looking for attrs
+                return None
+            self.data.position += 1
+            self.data.skip()
+            # Look for an encoding between matching quote marks
+            if self.data.currentByte in (b'"', b"'"):
+                quoteMark = self.data.currentByte
+                self.data.position += 1
+                oldPosition = self.data.position
+                if self.data.jumpTo(quoteMark):
+                    return self.data[oldPosition:self.data.position]
+                else:
+                    return None
+            else:
+                # Unquoted value
+                oldPosition = self.data.position
+                try:
+                    self.data.skipUntil(spaceCharactersBytes)
+                    return self.data[oldPosition:self.data.position]
+                except StopIteration:
+                    # Return the whole remaining value
+                    return self.data[oldPosition:]
+        except StopIteration:
+            return None
+
+
+def codecName(encoding):
+    """Return the python codec name corresponding to an encoding or None if the
+    string doesn't correspond to a valid encoding."""
+    if isinstance(encoding, bytes):
+        try:
+            encoding = encoding.decode("ascii")
+        except UnicodeDecodeError:
+            return None
+    if encoding:
+        canonicalName = ascii_punctuation_re.sub("", encoding).lower()
+        return encodings.get(canonicalName, None)
+    else:
+        return None

lib/html5lib/sanitizer.py

@@ -0,0 +1,271 @@
+from __future__ import absolute_import, division, unicode_literals
+
+import re
+from xml.sax.saxutils import escape, unescape
+
+from .tokenizer import HTMLTokenizer
+from .constants import tokenTypes
+
+
+class HTMLSanitizerMixin(object):
+    """ sanitization of XHTML+MathML+SVG and of inline style attributes."""
+
+    acceptable_elements = ['a', 'abbr', 'acronym', 'address', 'area',
+                           'article', 'aside', 'audio', 'b', 'big', 'blockquote', 'br', 'button',
+                           'canvas', 'caption', 'center', 'cite', 'code', 'col', 'colgroup',
+                           'command', 'datagrid', 'datalist', 'dd', 'del', 'details', 'dfn',
+                           'dialog', 'dir', 'div', 'dl', 'dt', 'em', 'event-source', 'fieldset',
+                           'figcaption', 'figure', 'footer', 'font', 'form', 'header', 'h1',
+                           'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'i', 'img', 'input', 'ins',
+                           'keygen', 'kbd', 'label', 'legend', 'li', 'm', 'map', 'menu', 'meter',
+                           'multicol', 'nav', 'nextid', 'ol', 'output', 'optgroup', 'option',
+                           'p', 'pre', 'progress', 'q', 's', 'samp', 'section', 'select',
+                           'small', 'sound', 'source', 'spacer', 'span', 'strike', 'strong',
+                           'sub', 'sup', 'table', 'tbody', 'td', 'textarea', 'time', 'tfoot',
+                           'th', 'thead', 'tr', 'tt', 'u', 'ul', 'var', 'video']
+
+    mathml_elements = ['maction', 'math', 'merror', 'mfrac', 'mi',
+                       'mmultiscripts', 'mn', 'mo', 'mover', 'mpadded', 'mphantom',
+                       'mprescripts', 'mroot', 'mrow', 'mspace', 'msqrt', 'mstyle', 'msub',
+                       'msubsup', 'msup', 'mtable', 'mtd', 'mtext', 'mtr', 'munder',
+                       'munderover', 'none']
+
+    svg_elements = ['a', 'animate', 'animateColor', 'animateMotion',
+                    'animateTransform', 'clipPath', 'circle', 'defs', 'desc', 'ellipse',
+                    'font-face', 'font-face-name', 'font-face-src', 'g', 'glyph', 'hkern',
+                    'linearGradient', 'line', 'marker', 'metadata', 'missing-glyph',
+                    'mpath', 'path', 'polygon', 'polyline', 'radialGradient', 'rect',
+                    'set', 'stop', 'svg', 'switch', 'text', 'title', 'tspan', 'use']
+
+    acceptable_attributes = ['abbr', 'accept', 'accept-charset', 'accesskey',
+                             'action', 'align', 'alt', 'autocomplete', 'autofocus', 'axis',
+                             'background', 'balance', 'bgcolor', 'bgproperties', 'border',
+                             'bordercolor', 'bordercolordark', 'bordercolorlight', 'bottompadding',
+                             'cellpadding', 'cellspacing', 'ch', 'challenge', 'char', 'charoff',
+                             'choff', 'charset', 'checked', 'cite', 'class', 'clear', 'color',
+                             'cols', 'colspan', 'compact', 'contenteditable', 'controls', 'coords',
+                             'data', 'datafld', 'datapagesize', 'datasrc', 'datetime', 'default',
+                             'delay', 'dir', 'disabled', 'draggable', 'dynsrc', 'enctype', 'end',
+                             'face', 'for', 'form', 'frame', 'galleryimg', 'gutter', 'headers',
+                             'height', 'hidefocus', 'hidden', 'high', 'href', 'hreflang', 'hspace',
+                             'icon', 'id', 'inputmode', 'ismap', 'keytype', 'label', 'leftspacing',
+                             'lang', 'list', 'longdesc', 'loop', 'loopcount', 'loopend',
+                             'loopstart', 'low', 'lowsrc', 'max', 'maxlength', 'media', 'method',
+                             'min', 'multiple', 'name', 'nohref', 'noshade', 'nowrap', 'open',
+                             'optimum', 'pattern', 'ping', 'point-size', 'poster', 'pqg', 'preload',
+                             'prompt', 'radiogroup', 'readonly', 'rel', 'repeat-max', 'repeat-min',
+                             'replace', 'required', 'rev', 'rightspacing', 'rows', 'rowspan',
+                             'rules', 'scope', 'selected', 'shape', 'size', 'span', 'src', 'start',
+                             'step', 'style', 'summary', 'suppress', 'tabindex', 'target',
+                             'template', 'title', 'toppadding', 'type', 'unselectable', 'usemap',
+                             'urn', 'valign', 'value', 'variable', 'volume', 'vspace', 'vrml',
+                             'width', 'wrap', 'xml:lang']
+
+    mathml_attributes = ['actiontype', 'align', 'columnalign', 'columnalign',
+                         'columnalign', 'columnlines', 'columnspacing', 'columnspan', 'depth',
+                         'display', 'displaystyle', 'equalcolumns', 'equalrows', 'fence',
+                         'fontstyle', 'fontweight', 'frame', 'height', 'linethickness', 'lspace',
+                         'mathbackground', 'mathcolor', 'mathvariant', 'mathvariant', 'maxsize',
+                         'minsize', 'other', 'rowalign', 'rowalign', 'rowalign', 'rowlines',
+                         'rowspacing', 'rowspan', 'rspace', 'scriptlevel', 'selection',
+                         'separator', 'stretchy', 'width', 'width', 'xlink:href', 'xlink:show',
+                         'xlink:type', 'xmlns', 'xmlns:xlink']
+
+    svg_attributes = ['accent-height', 'accumulate', 'additive', 'alphabetic',
+                      'arabic-form', 'ascent', 'attributeName', 'attributeType',
+                      'baseProfile', 'bbox', 'begin', 'by', 'calcMode', 'cap-height',
+                      'class', 'clip-path', 'color', 'color-rendering', 'content', 'cx',
+                      'cy', 'd', 'dx', 'dy', 'descent', 'display', 'dur', 'end', 'fill',
+                      'fill-opacity', 'fill-rule', 'font-family', 'font-size',
+                      'font-stretch', 'font-style', 'font-variant', 'font-weight', 'from',
+                      'fx', 'fy', 'g1', 'g2', 'glyph-name', 'gradientUnits', 'hanging',
+                      'height', 'horiz-adv-x', 'horiz-origin-x', 'id', 'ideographic', 'k',
+                      'keyPoints', 'keySplines', 'keyTimes', 'lang', 'marker-end',
+                      'marker-mid', 'marker-start', 'markerHeight', 'markerUnits',
+                      'markerWidth', 'mathematical', 'max', 'min', 'name', 'offset',
+                      'opacity', 'orient', 'origin', 'overline-position',
+                      'overline-thickness', 'panose-1', 'path', 'pathLength', 'points',
+                      'preserveAspectRatio', 'r', 'refX', 'refY', 'repeatCount',
+                      'repeatDur', 'requiredExtensions', 'requiredFeatures', 'restart',
+                      'rotate', 'rx', 'ry', 'slope', 'stemh', 'stemv', 'stop-color',
+                      'stop-opacity', 'strikethrough-position', 'strikethrough-thickness',
+                      'stroke', 'stroke-dasharray', 'stroke-dashoffset', 'stroke-linecap',
+                      'stroke-linejoin', 'stroke-miterlimit', 'stroke-opacity',
+                      'stroke-width', 'systemLanguage', 'target', 'text-anchor', 'to',
+                      'transform', 'type', 'u1', 'u2', 'underline-position',
+                      'underline-thickness', 'unicode', 'unicode-range', 'units-per-em',
+                      'values', 'version', 'viewBox', 'visibility', 'width', 'widths', 'x',
+                      'x-height', 'x1', 'x2', 'xlink:actuate', 'xlink:arcrole',
+                      'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type',
+                      'xml:base', 'xml:lang', 'xml:space', 'xmlns', 'xmlns:xlink', 'y',
+                      'y1', 'y2', 'zoomAndPan']
+
+    attr_val_is_uri = ['href', 'src', 'cite', 'action', 'longdesc', 'poster',
+                       'xlink:href', 'xml:base']
+
+    svg_attr_val_allows_ref = ['clip-path', 'color-profile', 'cursor', 'fill',
+                               'filter', 'marker', 'marker-start', 'marker-mid', 'marker-end',
+                               'mask', 'stroke']
+
+    svg_allow_local_href = ['altGlyph', 'animate', 'animateColor',
+                            'animateMotion', 'animateTransform', 'cursor', 'feImage', 'filter',
+                            'linearGradient', 'pattern', 'radialGradient', 'textpath', 'tref',
+                            'set', 'use']
+
+    acceptable_css_properties = ['azimuth', 'background-color',
+                                 'border-bottom-color', 'border-collapse', 'border-color',
+                                 'border-left-color', 'border-right-color', 'border-top-color', 'clear',
+                                 'color', 'cursor', 'direction', 'display', 'elevation', 'float', 'font',
+                                 'font-family', 'font-size', 'font-style', 'font-variant', 'font-weight',
+                                 'height', 'letter-spacing', 'line-height', 'overflow', 'pause',
+                                 'pause-after', 'pause-before', 'pitch', 'pitch-range', 'richness',
+                                 'speak', 'speak-header', 'speak-numeral', 'speak-punctuation',
+                                 'speech-rate', 'stress', 'text-align', 'text-decoration', 'text-indent',
+                                 'unicode-bidi', 'vertical-align', 'voice-family', 'volume',
+                                 'white-space', 'width']
+
+    acceptable_css_keywords = ['auto', 'aqua', 'black', 'block', 'blue',
+                               'bold', 'both', 'bottom', 'brown', 'center', 'collapse', 'dashed',
+                               'dotted', 'fuchsia', 'gray', 'green', '!important', 'italic', 'left',
+                               'lime', 'maroon', 'medium', 'none', 'navy', 'normal', 'nowrap', 'olive',
+                               'pointer', 'purple', 'red', 'right', 'solid', 'silver', 'teal', 'top',
+                               'transparent', 'underline', 'white', 'yellow']
+
+    acceptable_svg_properties = ['fill', 'fill-opacity', 'fill-rule',
+                                 'stroke', 'stroke-width', 'stroke-linecap', 'stroke-linejoin',
+                                 'stroke-opacity']
+
+    acceptable_protocols = ['ed2k', 'ftp', 'http', 'https', 'irc',
+                            'mailto', 'news', 'gopher', 'nntp', 'telnet', 'webcal',
+                            'xmpp', 'callto', 'feed', 'urn', 'aim', 'rsync', 'tag',
+                            'ssh', 'sftp', 'rtsp', 'afs']
+
+    # subclasses may define their own versions of these constants
+    allowed_elements = acceptable_elements + mathml_elements + svg_elements
+    allowed_attributes = acceptable_attributes + mathml_attributes + svg_attributes
+    allowed_css_properties = acceptable_css_properties
+    allowed_css_keywords = acceptable_css_keywords
+    allowed_svg_properties = acceptable_svg_properties
+    allowed_protocols = acceptable_protocols
+
+    # Sanitize the +html+, escaping all elements not in ALLOWED_ELEMENTS, and
+    # stripping out all # attributes not in ALLOWED_ATTRIBUTES. Style
+    # attributes are parsed, and a restricted set, # specified by
+    # ALLOWED_CSS_PROPERTIES and ALLOWED_CSS_KEYWORDS, are allowed through.
+    # attributes in ATTR_VAL_IS_URI are scanned, and only URI schemes specified
+    # in ALLOWED_PROTOCOLS are allowed.
+    #
+    #   sanitize_html('<script> do_nasty_stuff() </script>')
+    #    => &lt;script> do_nasty_stuff() &lt;/script>
+    #   sanitize_html('<a href="javascript: sucker();">Click here for $100</a>')
+    #    => <a>Click here for $100</a>
+    def sanitize_token(self, token):
+
+        # accommodate filters which use token_type differently
+        token_type = token["type"]
+        if token_type in list(tokenTypes.keys()):
+            token_type = tokenTypes[token_type]
+
+        if token_type in (tokenTypes["StartTag"], tokenTypes["EndTag"],
+                          tokenTypes["EmptyTag"]):
+            if token["name"] in self.allowed_elements:
+                return self.allowed_token(token, token_type)
+            else:
+                return self.disallowed_token(token, token_type)
+        elif token_type == tokenTypes["Comment"]:
+            pass
+        else:
+            return token
+
+    def allowed_token(self, token, token_type):
+        if "data" in token:
+            attrs = dict([(name, val) for name, val in
+                          token["data"][::-1]
+                          if name in self.allowed_attributes])
+            for attr in self.attr_val_is_uri:
+                if attr not in attrs:
+                    continue
+                val_unescaped = re.sub("[`\000-\040\177-\240\s]+", '',
+                                       unescape(attrs[attr])).lower()
+                # remove replacement characters from unescaped characters
+                val_unescaped = val_unescaped.replace("\ufffd", "")
+                if (re.match("^[a-z0-9][-+.a-z0-9]*:", val_unescaped) and
+                    (val_unescaped.split(':')[0] not in
+                     self.allowed_protocols)):
+                    del attrs[attr]
+            for attr in self.svg_attr_val_allows_ref:
+                if attr in attrs:
+                    attrs[attr] = re.sub(r'url\s*\(\s*[^#\s][^)]+?\)',
+                                         ' ',
+                                         unescape(attrs[attr]))
+            if (token["name"] in self.svg_allow_local_href and
+                'xlink:href' in attrs and re.search('^\s*[^#\s].*',
+                                                    attrs['xlink:href'])):
+                del attrs['xlink:href']
+            if 'style' in attrs:
+                attrs['style'] = self.sanitize_css(attrs['style'])
+            token["data"] = [[name, val] for name, val in list(attrs.items())]
+        return token
+
+    def disallowed_token(self, token, token_type):
+        if token_type == tokenTypes["EndTag"]:
+            token["data"] = "</%s>" % token["name"]
+        elif token["data"]:
+            attrs = ''.join([' %s="%s"' % (k, escape(v)) for k, v in token["data"]])
+            token["data"] = "<%s%s>" % (token["name"], attrs)
+        else:
+            token["data"] = "<%s>" % token["name"]
+        if token.get("selfClosing"):
+            token["data"] = token["data"][:-1] + "/>"
+
+        if token["type"] in list(tokenTypes.keys()):
+            token["type"] = "Characters"
+        else:
+            token["type"] = tokenTypes["Characters"]
+
+        del token["name"]
+        return token
+
+    def sanitize_css(self, style):
+        # disallow urls
+        style = re.compile('url\s*\(\s*[^\s)]+?\s*\)\s*').sub(' ', style)
+
+        # gauntlet
+        if not re.match("""^([:,;#%.\sa-zA-Z0-9!]|\w-\w|'[\s\w]+'|"[\s\w]+"|\([\d,\s]+\))*$""", style):
+            return ''
+        if not re.match("^\s*([-\w]+\s*:[^:;]*(;\s*|$))*$", style):
+            return ''
+
+        clean = []
+        for prop, value in re.findall("([-\w]+)\s*:\s*([^:;]*)", style):
+            if not value:
+                continue
+            if prop.lower() in self.allowed_css_properties:
+                clean.append(prop + ': ' + value + ';')
+            elif prop.split('-')[0].lower() in ['background', 'border', 'margin',
+                                                'padding']:
+                for keyword in value.split():
+                    if not keyword in self.acceptable_css_keywords and \
+                            not re.match("^(#[0-9a-f]+|rgb\(\d+%?,\d*%?,?\d*%?\)?|\d{0,2}\.?\d{0,2}(cm|em|ex|in|mm|pc|pt|px|%|,|\))?)$", keyword):
+                        break
+                else:
+                    clean.append(prop + ': ' + value + ';')
+            elif prop.lower() in self.allowed_svg_properties:
+                clean.append(prop + ': ' + value + ';')
+
+        return ' '.join(clean)
+
+
+class HTMLSanitizer(HTMLTokenizer, HTMLSanitizerMixin):
+    def __init__(self, stream, encoding=None, parseMeta=True, useChardet=True,
+                 lowercaseElementName=False, lowercaseAttrName=False, parser=None):
+        # Change case matching defaults as we only output lowercase html anyway
+        # This solution doesn't seem ideal...
+        HTMLTokenizer.__init__(self, stream, encoding, parseMeta, useChardet,
+                               lowercaseElementName, lowercaseAttrName, parser=parser)
+
+    def __iter__(self):
+        for token in HTMLTokenizer.__iter__(self):
+            token = self.sanitize_token(token)
+            if token:
+                yield token

lib/html5lib/serializer/__init__.py

@@ -0,0 +1,16 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from .. import treewalkers
+
+from .htmlserializer import HTMLSerializer
+
+
+def serialize(input, tree="etree", format="html", encoding=None,
+              **serializer_opts):
+    # XXX: Should we cache this?
+    walker = treewalkers.getTreeWalker(tree)
+    if format == "html":
+        s = HTMLSerializer(**serializer_opts)
+    else:
+        raise ValueError("type must be html")
+    return s.render(walker(input), encoding)

lib/html5lib/serializer/htmlserializer.py

@@ -0,0 +1,320 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+
+import gettext
+_ = gettext.gettext
+
+try:
+    from functools import reduce
+except ImportError:
+    pass
+
+from ..constants import voidElements, booleanAttributes, spaceCharacters
+from ..constants import rcdataElements, entities, xmlEntities
+from .. import utils
+from xml.sax.saxutils import escape
+
+spaceCharacters = "".join(spaceCharacters)
+
+try:
+    from codecs import register_error, xmlcharrefreplace_errors
+except ImportError:
+    unicode_encode_errors = "strict"
+else:
+    unicode_encode_errors = "htmlentityreplace"
+
+    encode_entity_map = {}
+    is_ucs4 = len("\U0010FFFF") == 1
+    for k, v in list(entities.items()):
+        # skip multi-character entities
+        if ((is_ucs4 and len(v) > 1) or
+                (not is_ucs4 and len(v) > 2)):
+            continue
+        if v != "&":
+            if len(v) == 2:
+                v = utils.surrogatePairToCodepoint(v)
+            else:
+                v = ord(v)
+            if not v in encode_entity_map or k.islower():
+                # prefer < over < and similarly for &, >, etc.
+                encode_entity_map[v] = k
+
+    def htmlentityreplace_errors(exc):
+        if isinstance(exc, (UnicodeEncodeError, UnicodeTranslateError)):
+            res = []
+            codepoints = []
+            skip = False
+            for i, c in enumerate(exc.object[exc.start:exc.end]):
+                if skip:
+                    skip = False
+                    continue
+                index = i + exc.start
+                if utils.isSurrogatePair(exc.object[index:min([exc.end, index + 2])]):
+                    codepoint = utils.surrogatePairToCodepoint(exc.object[index:index + 2])
+                    skip = True
+                else:
+                    codepoint = ord(c)
+                codepoints.append(codepoint)
+            for cp in codepoints:
+                e = encode_entity_map.get(cp)
+                if e:
+                    res.append("&")
+                    res.append(e)
+                    if not e.endswith(";"):
+                        res.append(";")
+                else:
+                    res.append("&#x%s;" % (hex(cp)[2:]))
+            return ("".join(res), exc.end)
+        else:
+            return xmlcharrefreplace_errors(exc)
+
+    register_error(unicode_encode_errors, htmlentityreplace_errors)
+
+    del register_error
+
+
+class HTMLSerializer(object):
+
+    # attribute quoting options
+    quote_attr_values = False
+    quote_char = '"'
+    use_best_quote_char = True
+
+    # tag syntax options
+    omit_optional_tags = True
+    minimize_boolean_attributes = True
+    use_trailing_solidus = False
+    space_before_trailing_solidus = True
+
+    # escaping options
+    escape_lt_in_attrs = False
+    escape_rcdata = False
+    resolve_entities = True
+
+    # miscellaneous options
+    alphabetical_attributes = False
+    inject_meta_charset = True
+    strip_whitespace = False
+    sanitize = False
+
+    options = ("quote_attr_values", "quote_char", "use_best_quote_char",
+               "omit_optional_tags", "minimize_boolean_attributes",
+               "use_trailing_solidus", "space_before_trailing_solidus",
+               "escape_lt_in_attrs", "escape_rcdata", "resolve_entities",
+               "alphabetical_attributes", "inject_meta_charset",
+               "strip_whitespace", "sanitize")
+
+    def __init__(self, **kwargs):
+        """Initialize HTMLSerializer.
+
+        Keyword options (default given first unless specified) include:
+
+        inject_meta_charset=True|False
+          Whether it insert a meta element to define the character set of the
+          document.
+        quote_attr_values=True|False
+          Whether to quote attribute values that don't require quoting
+          per HTML5 parsing rules.
+        quote_char=u'"'|u"'"
+          Use given quote character for attribute quoting. Default is to
+          use double quote unless attribute value contains a double quote,
+          in which case single quotes are used instead.
+        escape_lt_in_attrs=False|True
+          Whether to escape < in attribute values.
+        escape_rcdata=False|True
+          Whether to escape characters that need to be escaped within normal
+          elements within rcdata elements such as style.
+        resolve_entities=True|False
+          Whether to resolve named character entities that appear in the
+          source tree. The XML predefined entities &lt; &gt; &amp; &quot; &apos;
+          are unaffected by this setting.
+        strip_whitespace=False|True
+          Whether to remove semantically meaningless whitespace. (This
+          compresses all whitespace to a single space except within pre.)
+        minimize_boolean_attributes=True|False
+          Shortens boolean attributes to give just the attribute value,
+          for example <input disabled="disabled"> becomes <input disabled>.
+        use_trailing_solidus=False|True
+          Includes a close-tag slash at the end of the start tag of void
+          elements (empty elements whose end tag is forbidden). E.g. <hr/>.
+        space_before_trailing_solidus=True|False
+          Places a space immediately before the closing slash in a tag
+          using a trailing solidus. E.g. <hr />. Requires use_trailing_solidus.
+        sanitize=False|True
+          Strip all unsafe or unknown constructs from output.
+          See `html5lib user documentation`_
+        omit_optional_tags=True|False
+          Omit start/end tags that are optional.
+        alphabetical_attributes=False|True
+          Reorder attributes to be in alphabetical order.
+
+        .. _html5lib user documentation: http://code.google.com/p/html5lib/wiki/UserDocumentation
+        """
+        if 'quote_char' in kwargs:
+            self.use_best_quote_char = False
+        for attr in self.options:
+            setattr(self, attr, kwargs.get(attr, getattr(self, attr)))
+        self.errors = []
+        self.strict = False
+
+    def encode(self, string):
+        assert(isinstance(string, text_type))
+        if self.encoding:
+            return string.encode(self.encoding, unicode_encode_errors)
+        else:
+            return string
+
+    def encodeStrict(self, string):
+        assert(isinstance(string, text_type))
+        if self.encoding:
+            return string.encode(self.encoding, "strict")
+        else:
+            return string
+
+    def serialize(self, treewalker, encoding=None):
+        self.encoding = encoding
+        in_cdata = False
+        self.errors = []
+
+        if encoding and self.inject_meta_charset:
+            from ..filters.inject_meta_charset import Filter
+            treewalker = Filter(treewalker, encoding)
+        # WhitespaceFilter should be used before OptionalTagFilter
+        # for maximum efficiently of this latter filter
+        if self.strip_whitespace:
+            from ..filters.whitespace import Filter
+            treewalker = Filter(treewalker)
+        if self.sanitize:
+            from ..filters.sanitizer import Filter
+            treewalker = Filter(treewalker)
+        if self.omit_optional_tags:
+            from ..filters.optionaltags import Filter
+            treewalker = Filter(treewalker)
+        # Alphabetical attributes must be last, as other filters
+        # could add attributes and alter the order
+        if self.alphabetical_attributes:
+            from ..filters.alphabeticalattributes import Filter
+            treewalker = Filter(treewalker)
+
+        for token in treewalker:
+            type = token["type"]
+            if type == "Doctype":
+                doctype = "<!DOCTYPE %s" % token["name"]
+
+                if token["publicId"]:
+                    doctype += ' PUBLIC "%s"' % token["publicId"]
+                elif token["systemId"]:
+                    doctype += " SYSTEM"
+                if token["systemId"]:
+                    if token["systemId"].find('"') >= 0:
+                        if token["systemId"].find("'") >= 0:
+                            self.serializeError(_("System identifer contains both single and double quote characters"))
+                        quote_char = "'"
+                    else:
+                        quote_char = '"'
+                    doctype += " %s%s%s" % (quote_char, token["systemId"], quote_char)
+
+                doctype += ">"
+                yield self.encodeStrict(doctype)
+
+            elif type in ("Characters", "SpaceCharacters"):
+                if type == "SpaceCharacters" or in_cdata:
+                    if in_cdata and token["data"].find("</") >= 0:
+                        self.serializeError(_("Unexpected </ in CDATA"))
+                    yield self.encode(token["data"])
+                else:
+                    yield self.encode(escape(token["data"]))
+
+            elif type in ("StartTag", "EmptyTag"):
+                name = token["name"]
+                yield self.encodeStrict("<%s" % name)
+                if name in rcdataElements and not self.escape_rcdata:
+                    in_cdata = True
+                elif in_cdata:
+                    self.serializeError(_("Unexpected child element of a CDATA element"))
+                for (attr_namespace, attr_name), attr_value in token["data"].items():
+                    # TODO: Add namespace support here
+                    k = attr_name
+                    v = attr_value
+                    yield self.encodeStrict(' ')
+
+                    yield self.encodeStrict(k)
+                    if not self.minimize_boolean_attributes or \
+                        (k not in booleanAttributes.get(name, tuple())
+                         and k not in booleanAttributes.get("", tuple())):
+                        yield self.encodeStrict("=")
+                        if self.quote_attr_values or not v:
+                            quote_attr = True
+                        else:
+                            quote_attr = reduce(lambda x, y: x or (y in v),
+                                                spaceCharacters + ">\"'=", False)
+                        v = v.replace("&", "&amp;")
+                        if self.escape_lt_in_attrs:
+                            v = v.replace("<", "&lt;")
+                        if quote_attr:
+                            quote_char = self.quote_char
+                            if self.use_best_quote_char:
+                                if "'" in v and '"' not in v:
+                                    quote_char = '"'
+                                elif '"' in v and "'" not in v:
+                                    quote_char = "'"
+                            if quote_char == "'":
+                                v = v.replace("'", "&#39;")
+                            else:
+                                v = v.replace('"', "&quot;")
+                            yield self.encodeStrict(quote_char)
+                            yield self.encode(v)
+                            yield self.encodeStrict(quote_char)
+                        else:
+                            yield self.encode(v)
+                if name in voidElements and self.use_trailing_solidus:
+                    if self.space_before_trailing_solidus:
+                        yield self.encodeStrict(" /")
+                    else:
+                        yield self.encodeStrict("/")
+                yield self.encode(">")
+
+            elif type == "EndTag":
+                name = token["name"]
+                if name in rcdataElements:
+                    in_cdata = False
+                elif in_cdata:
+                    self.serializeError(_("Unexpected child element of a CDATA element"))
+                yield self.encodeStrict("</%s>" % name)
+
+            elif type == "Comment":
+                data = token["data"]
+                if data.find("--") >= 0:
+                    self.serializeError(_("Comment contains --"))
+                yield self.encodeStrict("<!--%s-->" % token["data"])
+
+            elif type == "Entity":
+                name = token["name"]
+                key = name + ";"
+                if not key in entities:
+                    self.serializeError(_("Entity %s not recognized" % name))
+                if self.resolve_entities and key not in xmlEntities:
+                    data = entities[key]
+                else:
+                    data = "&%s;" % name
+                yield self.encodeStrict(data)
+
+            else:
+                self.serializeError(token["data"])
+
+    def render(self, treewalker, encoding=None):
+        if encoding:
+            return b"".join(list(self.serialize(treewalker, encoding)))
+        else:
+            return "".join(list(self.serialize(treewalker)))
+
+    def serializeError(self, data="XXX ERROR MESSAGE NEEDED"):
+        # XXX The idea is to make data mandatory.
+        self.errors.append(data)
+        if self.strict:
+            raise SerializeError
+
+
+def SerializeError(Exception):
+    """Error in serialized tree"""
+    pass

lib/html5lib/treeadapters/sax.py

@@ -0,0 +1,44 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from xml.sax.xmlreader import AttributesNSImpl
+
+from ..constants import adjustForeignAttributes, unadjustForeignAttributes
+
+prefix_mapping = {}
+for prefix, localName, namespace in adjustForeignAttributes.values():
+    if prefix is not None:
+        prefix_mapping[prefix] = namespace
+
+
+def to_sax(walker, handler):
+    """Call SAX-like content handler based on treewalker walker"""
+    handler.startDocument()
+    for prefix, namespace in prefix_mapping.items():
+        handler.startPrefixMapping(prefix, namespace)
+
+    for token in walker:
+        type = token["type"]
+        if type == "Doctype":
+            continue
+        elif type in ("StartTag", "EmptyTag"):
+            attrs = AttributesNSImpl(token["data"],
+                                     unadjustForeignAttributes)
+            handler.startElementNS((token["namespace"], token["name"]),
+                                   token["name"],
+                                   attrs)
+            if type == "EmptyTag":
+                handler.endElementNS((token["namespace"], token["name"]),
+                                     token["name"])
+        elif type == "EndTag":
+            handler.endElementNS((token["namespace"], token["name"]),
+                                 token["name"])
+        elif type in ("Characters", "SpaceCharacters"):
+            handler.characters(token["data"])
+        elif type == "Comment":
+            pass
+        else:
+            assert False, "Unknown token type"
+
+    for prefix, namespace in prefix_mapping.items():
+        handler.endPrefixMapping(prefix)
+    handler.endDocument()

lib/html5lib/treebuilders/__init__.py

@@ -0,0 +1,76 @@
+"""A collection of modules for building different kinds of tree from
+HTML documents.
+
+To create a treebuilder for a new type of tree, you need to do
+implement several things:
+
+1) A set of classes for various types of elements: Document, Doctype,
+Comment, Element. These must implement the interface of
+_base.treebuilders.Node (although comment nodes have a different
+signature for their constructor, see treebuilders.etree.Comment)
+Textual content may also be implemented as another node type, or not, as
+your tree implementation requires.
+
+2) A treebuilder object (called TreeBuilder by convention) that
+inherits from treebuilders._base.TreeBuilder. This has 4 required attributes:
+documentClass - the class to use for the bottommost node of a document
+elementClass - the class to use for HTML Elements
+commentClass - the class to use for comments
+doctypeClass - the class to use for doctypes
+It also has one required method:
+getDocument - Returns the root node of the complete document tree
+
+3) If you wish to run the unit tests, you must also create a
+testSerializer method on your treebuilder which accepts a node and
+returns a string containing Node and its children serialized according
+to the format used in the unittests
+"""
+
+from __future__ import absolute_import, division, unicode_literals
+
+from ..utils import default_etree
+
+treeBuilderCache = {}
+
+
+def getTreeBuilder(treeType, implementation=None, **kwargs):
+    """Get a TreeBuilder class for various types of tree with built-in support
+
+    treeType - the name of the tree type required (case-insensitive). Supported
+               values are:
+
+               "dom" - A generic builder for DOM implementations, defaulting to
+                       a xml.dom.minidom based implementation.
+               "etree" - A generic builder for tree implementations exposing an
+                         ElementTree-like interface, defaulting to
+                         xml.etree.cElementTree if available and
+                         xml.etree.ElementTree if not.
+               "lxml" - A etree-based builder for lxml.etree, handling
+                        limitations of lxml's implementation.
+
+    implementation - (Currently applies to the "etree" and "dom" tree types). A
+                      module implementing the tree type e.g.
+                      xml.etree.ElementTree or xml.etree.cElementTree."""
+
+    treeType = treeType.lower()
+    if treeType not in treeBuilderCache:
+        if treeType == "dom":
+            from . import dom
+            # Come up with a sane default (pref. from the stdlib)
+            if implementation is None:
+                from xml.dom import minidom
+                implementation = minidom
+            # NEVER cache here, caching is done in the dom submodule
+            return dom.getDomModule(implementation, **kwargs).TreeBuilder
+        elif treeType == "lxml":
+            from . import etree_lxml
+            treeBuilderCache[treeType] = etree_lxml.TreeBuilder
+        elif treeType == "etree":
+            from . import etree
+            if implementation is None:
+                implementation = default_etree
+            # NEVER cache here, caching is done in the etree submodule
+            return etree.getETreeModule(implementation, **kwargs).TreeBuilder
+        else:
+            raise ValueError("""Unrecognised treebuilder "%s" """ % treeType)
+    return treeBuilderCache.get(treeType)

lib/html5lib/treebuilders/_base.py

@@ -0,0 +1,377 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+
+from ..constants import scopingElements, tableInsertModeElements, namespaces
+
+# The scope markers are inserted when entering object elements,
+# marquees, table cells, and table captions, and are used to prevent formatting
+# from "leaking" into tables, object elements, and marquees.
+Marker = None
+
+listElementsMap = {
+    None: (frozenset(scopingElements), False),
+    "button": (frozenset(scopingElements | set([(namespaces["html"], "button")])), False),
+    "list": (frozenset(scopingElements | set([(namespaces["html"], "ol"),
+                                              (namespaces["html"], "ul")])), False),
+    "table": (frozenset([(namespaces["html"], "html"),
+                         (namespaces["html"], "table")]), False),
+    "select": (frozenset([(namespaces["html"], "optgroup"),
+                          (namespaces["html"], "option")]), True)
+}
+
+
+class Node(object):
+    def __init__(self, name):
+        """Node representing an item in the tree.
+        name - The tag name associated with the node
+        parent - The parent of the current node (or None for the document node)
+        value - The value of the current node (applies to text nodes and
+        comments
+        attributes - a dict holding name, value pairs for attributes of the node
+        childNodes - a list of child nodes of the current node. This must
+        include all elements but not necessarily other node types
+        _flags - A list of miscellaneous flags that can be set on the node
+        """
+        self.name = name
+        self.parent = None
+        self.value = None
+        self.attributes = {}
+        self.childNodes = []
+        self._flags = []
+
+    def __str__(self):
+        attributesStr = " ".join(["%s=\"%s\"" % (name, value)
+                                  for name, value in
+                                  self.attributes.items()])
+        if attributesStr:
+            return "<%s %s>" % (self.name, attributesStr)
+        else:
+            return "<%s>" % (self.name)
+
+    def __repr__(self):
+        return "<%s>" % (self.name)
+
+    def appendChild(self, node):
+        """Insert node as a child of the current node
+        """
+        raise NotImplementedError
+
+    def insertText(self, data, insertBefore=None):
+        """Insert data as text in the current node, positioned before the
+        start of node insertBefore or to the end of the node's text.
+        """
+        raise NotImplementedError
+
+    def insertBefore(self, node, refNode):
+        """Insert node as a child of the current node, before refNode in the
+        list of child nodes. Raises ValueError if refNode is not a child of
+        the current node"""
+        raise NotImplementedError
+
+    def removeChild(self, node):
+        """Remove node from the children of the current node
+        """
+        raise NotImplementedError
+
+    def reparentChildren(self, newParent):
+        """Move all the children of the current node to newParent.
+        This is needed so that trees that don't store text as nodes move the
+        text in the correct way
+        """
+        # XXX - should this method be made more general?
+        for child in self.childNodes:
+            newParent.appendChild(child)
+        self.childNodes = []
+
+    def cloneNode(self):
+        """Return a shallow copy of the current node i.e. a node with the same
+        name and attributes but with no parent or child nodes
+        """
+        raise NotImplementedError
+
+    def hasContent(self):
+        """Return true if the node has children or text, false otherwise
+        """
+        raise NotImplementedError
+
+
+class ActiveFormattingElements(list):
+    def append(self, node):
+        equalCount = 0
+        if node != Marker:
+            for element in self[::-1]:
+                if element == Marker:
+                    break
+                if self.nodesEqual(element, node):
+                    equalCount += 1
+                if equalCount == 3:
+                    self.remove(element)
+                    break
+        list.append(self, node)
+
+    def nodesEqual(self, node1, node2):
+        if not node1.nameTuple == node2.nameTuple:
+            return False
+
+        if not node1.attributes == node2.attributes:
+            return False
+
+        return True
+
+
+class TreeBuilder(object):
+    """Base treebuilder implementation
+    documentClass - the class to use for the bottommost node of a document
+    elementClass - the class to use for HTML Elements
+    commentClass - the class to use for comments
+    doctypeClass - the class to use for doctypes
+    """
+
+    # Document class
+    documentClass = None
+
+    # The class to use for creating a node
+    elementClass = None
+
+    # The class to use for creating comments
+    commentClass = None
+
+    # The class to use for creating doctypes
+    doctypeClass = None
+
+    # Fragment class
+    fragmentClass = None
+
+    def __init__(self, namespaceHTMLElements):
+        if namespaceHTMLElements:
+            self.defaultNamespace = "http://www.w3.org/1999/xhtml"
+        else:
+            self.defaultNamespace = None
+        self.reset()
+
+    def reset(self):
+        self.openElements = []
+        self.activeFormattingElements = ActiveFormattingElements()
+
+        # XXX - rename these to headElement, formElement
+        self.headPointer = None
+        self.formPointer = None
+
+        self.insertFromTable = False
+
+        self.document = self.documentClass()
+
+    def elementInScope(self, target, variant=None):
+
+        # If we pass a node in we match that. if we pass a string
+        # match any node with that name
+        exactNode = hasattr(target, "nameTuple")
+
+        listElements, invert = listElementsMap[variant]
+
+        for node in reversed(self.openElements):
+            if (node.name == target and not exactNode or
+                    node == target and exactNode):
+                return True
+            elif (invert ^ (node.nameTuple in listElements)):
+                return False
+
+        assert False  # We should never reach this point
+
+    def reconstructActiveFormattingElements(self):
+        # Within this algorithm the order of steps described in the
+        # specification is not quite the same as the order of steps in the
+        # code. It should still do the same though.
+
+        # Step 1: stop the algorithm when there's nothing to do.
+        if not self.activeFormattingElements:
+            return
+
+        # Step 2 and step 3: we start with the last element. So i is -1.
+        i = len(self.activeFormattingElements) - 1
+        entry = self.activeFormattingElements[i]
+        if entry == Marker or entry in self.openElements:
+            return
+
+        # Step 6
+        while entry != Marker and entry not in self.openElements:
+            if i == 0:
+                # This will be reset to 0 below
+                i = -1
+                break
+            i -= 1
+            # Step 5: let entry be one earlier in the list.
+            entry = self.activeFormattingElements[i]
+
+        while True:
+            # Step 7
+            i += 1
+
+            # Step 8
+            entry = self.activeFormattingElements[i]
+            clone = entry.cloneNode()  # Mainly to get a new copy of the attributes
+
+            # Step 9
+            element = self.insertElement({"type": "StartTag",
+                                          "name": clone.name,
+                                          "namespace": clone.namespace,
+                                          "data": clone.attributes})
+
+            # Step 10
+            self.activeFormattingElements[i] = element
+
+            # Step 11
+            if element == self.activeFormattingElements[-1]:
+                break
+
+    def clearActiveFormattingElements(self):
+        entry = self.activeFormattingElements.pop()
+        while self.activeFormattingElements and entry != Marker:
+            entry = self.activeFormattingElements.pop()
+
+    def elementInActiveFormattingElements(self, name):
+        """Check if an element exists between the end of the active
+        formatting elements and the last marker. If it does, return it, else
+        return false"""
+
+        for item in self.activeFormattingElements[::-1]:
+            # Check for Marker first because if it's a Marker it doesn't have a
+            # name attribute.
+            if item == Marker:
+                break
+            elif item.name == name:
+                return item
+        return False
+
+    def insertRoot(self, token):
+        element = self.createElement(token)
+        self.openElements.append(element)
+        self.document.appendChild(element)
+
+    def insertDoctype(self, token):
+        name = token["name"]
+        publicId = token["publicId"]
+        systemId = token["systemId"]
+
+        doctype = self.doctypeClass(name, publicId, systemId)
+        self.document.appendChild(doctype)
+
+    def insertComment(self, token, parent=None):
+        if parent is None:
+            parent = self.openElements[-1]
+        parent.appendChild(self.commentClass(token["data"]))
+
+    def createElement(self, token):
+        """Create an element but don't insert it anywhere"""
+        name = token["name"]
+        namespace = token.get("namespace", self.defaultNamespace)
+        element = self.elementClass(name, namespace)
+        element.attributes = token["data"]
+        return element
+
+    def _getInsertFromTable(self):
+        return self._insertFromTable
+
+    def _setInsertFromTable(self, value):
+        """Switch the function used to insert an element from the
+        normal one to the misnested table one and back again"""
+        self._insertFromTable = value
+        if value:
+            self.insertElement = self.insertElementTable
+        else:
+            self.insertElement = self.insertElementNormal
+
+    insertFromTable = property(_getInsertFromTable, _setInsertFromTable)
+
+    def insertElementNormal(self, token):
+        name = token["name"]
+        assert isinstance(name, text_type), "Element %s not unicode" % name
+        namespace = token.get("namespace", self.defaultNamespace)
+        element = self.elementClass(name, namespace)
+        element.attributes = token["data"]
+        self.openElements[-1].appendChild(element)
+        self.openElements.append(element)
+        return element
+
+    def insertElementTable(self, token):
+        """Create an element and insert it into the tree"""
+        element = self.createElement(token)
+        if self.openElements[-1].name not in tableInsertModeElements:
+            return self.insertElementNormal(token)
+        else:
+            # We should be in the InTable mode. This means we want to do
+            # special magic element rearranging
+            parent, insertBefore = self.getTableMisnestedNodePosition()
+            if insertBefore is None:
+                parent.appendChild(element)
+            else:
+                parent.insertBefore(element, insertBefore)
+            self.openElements.append(element)
+        return element
+
+    def insertText(self, data, parent=None):
+        """Insert text data."""
+        if parent is None:
+            parent = self.openElements[-1]
+
+        if (not self.insertFromTable or (self.insertFromTable and
+                                         self.openElements[-1].name
+                                         not in tableInsertModeElements)):
+            parent.insertText(data)
+        else:
+            # We should be in the InTable mode. This means we want to do
+            # special magic element rearranging
+            parent, insertBefore = self.getTableMisnestedNodePosition()
+            parent.insertText(data, insertBefore)
+
+    def getTableMisnestedNodePosition(self):
+        """Get the foster parent element, and sibling to insert before
+        (or None) when inserting a misnested table node"""
+        # The foster parent element is the one which comes before the most
+        # recently opened table element
+        # XXX - this is really inelegant
+        lastTable = None
+        fosterParent = None
+        insertBefore = None
+        for elm in self.openElements[::-1]:
+            if elm.name == "table":
+                lastTable = elm
+                break
+        if lastTable:
+            # XXX - we should really check that this parent is actually a
+            # node here
+            if lastTable.parent:
+                fosterParent = lastTable.parent
+                insertBefore = lastTable
+            else:
+                fosterParent = self.openElements[
+                    self.openElements.index(lastTable) - 1]
+        else:
+            fosterParent = self.openElements[0]
+        return fosterParent, insertBefore
+
+    def generateImpliedEndTags(self, exclude=None):
+        name = self.openElements[-1].name
+        # XXX td, th and tr are not actually needed
+        if (name in frozenset(("dd", "dt", "li", "option", "optgroup", "p", "rp", "rt"))
+                and name != exclude):
+            self.openElements.pop()
+            # XXX This is not entirely what the specification says. We should
+            # investigate it more closely.
+            self.generateImpliedEndTags(exclude)
+
+    def getDocument(self):
+        "Return the final tree"
+        return self.document
+
+    def getFragment(self):
+        "Return the final fragment"
+        # assert self.innerHTML
+        fragment = self.fragmentClass()
+        self.openElements[0].reparentChildren(fragment)
+        return fragment
+
+    def testSerializer(self, node):
+        """Serialize the subtree of node in the format required by unit tests
+        node - the node from which to start serializing"""
+        raise NotImplementedError

lib/html5lib/treebuilders/dom.py

@@ -0,0 +1,227 @@
+from __future__ import absolute_import, division, unicode_literals
+
+
+from xml.dom import minidom, Node
+import weakref
+
+from . import _base
+from .. import constants
+from ..constants import namespaces
+from ..utils import moduleFactoryFactory
+
+
+def getDomBuilder(DomImplementation):
+    Dom = DomImplementation
+
+    class AttrList(object):
+        def __init__(self, element):
+            self.element = element
+
+        def __iter__(self):
+            return list(self.element.attributes.items()).__iter__()
+
+        def __setitem__(self, name, value):
+            self.element.setAttribute(name, value)
+
+        def __len__(self):
+            return len(list(self.element.attributes.items()))
+
+        def items(self):
+            return [(item[0], item[1]) for item in
+                    list(self.element.attributes.items())]
+
+        def keys(self):
+            return list(self.element.attributes.keys())
+
+        def __getitem__(self, name):
+            return self.element.getAttribute(name)
+
+        def __contains__(self, name):
+            if isinstance(name, tuple):
+                raise NotImplementedError
+            else:
+                return self.element.hasAttribute(name)
+
+    class NodeBuilder(_base.Node):
+        def __init__(self, element):
+            _base.Node.__init__(self, element.nodeName)
+            self.element = element
+
+        namespace = property(lambda self: hasattr(self.element, "namespaceURI")
+                             and self.element.namespaceURI or None)
+
+        def appendChild(self, node):
+            node.parent = self
+            self.element.appendChild(node.element)
+
+        def insertText(self, data, insertBefore=None):
+            text = self.element.ownerDocument.createTextNode(data)
+            if insertBefore:
+                self.element.insertBefore(text, insertBefore.element)
+            else:
+                self.element.appendChild(text)
+
+        def insertBefore(self, node, refNode):
+            self.element.insertBefore(node.element, refNode.element)
+            node.parent = self
+
+        def removeChild(self, node):
+            if node.element.parentNode == self.element:
+                self.element.removeChild(node.element)
+            node.parent = None
+
+        def reparentChildren(self, newParent):
+            while self.element.hasChildNodes():
+                child = self.element.firstChild
+                self.element.removeChild(child)
+                newParent.element.appendChild(child)
+            self.childNodes = []
+
+        def getAttributes(self):
+            return AttrList(self.element)
+
+        def setAttributes(self, attributes):
+            if attributes:
+                for name, value in list(attributes.items()):
+                    if isinstance(name, tuple):
+                        if name[0] is not None:
+                            qualifiedName = (name[0] + ":" + name[1])
+                        else:
+                            qualifiedName = name[1]
+                        self.element.setAttributeNS(name[2], qualifiedName,
+                                                    value)
+                    else:
+                        self.element.setAttribute(
+                            name, value)
+        attributes = property(getAttributes, setAttributes)
+
+        def cloneNode(self):
+            return NodeBuilder(self.element.cloneNode(False))
+
+        def hasContent(self):
+            return self.element.hasChildNodes()
+
+        def getNameTuple(self):
+            if self.namespace is None:
+                return namespaces["html"], self.name
+            else:
+                return self.namespace, self.name
+
+        nameTuple = property(getNameTuple)
+
+    class TreeBuilder(_base.TreeBuilder):
+        def documentClass(self):
+            self.dom = Dom.getDOMImplementation().createDocument(None, None, None)
+            return weakref.proxy(self)
+
+        def insertDoctype(self, token):
+            name = token["name"]
+            publicId = token["publicId"]
+            systemId = token["systemId"]
+
+            domimpl = Dom.getDOMImplementation()
+            doctype = domimpl.createDocumentType(name, publicId, systemId)
+            self.document.appendChild(NodeBuilder(doctype))
+            if Dom == minidom:
+                doctype.ownerDocument = self.dom
+
+        def elementClass(self, name, namespace=None):
+            if namespace is None and self.defaultNamespace is None:
+                node = self.dom.createElement(name)
+            else:
+                node = self.dom.createElementNS(namespace, name)
+
+            return NodeBuilder(node)
+
+        def commentClass(self, data):
+            return NodeBuilder(self.dom.createComment(data))
+
+        def fragmentClass(self):
+            return NodeBuilder(self.dom.createDocumentFragment())
+
+        def appendChild(self, node):
+            self.dom.appendChild(node.element)
+
+        def testSerializer(self, element):
+            return testSerializer(element)
+
+        def getDocument(self):
+            return self.dom
+
+        def getFragment(self):
+            return _base.TreeBuilder.getFragment(self).element
+
+        def insertText(self, data, parent=None):
+            data = data
+            if parent != self:
+                _base.TreeBuilder.insertText(self, data, parent)
+            else:
+                # HACK: allow text nodes as children of the document node
+                if hasattr(self.dom, '_child_node_types'):
+                    if not Node.TEXT_NODE in self.dom._child_node_types:
+                        self.dom._child_node_types = list(self.dom._child_node_types)
+                        self.dom._child_node_types.append(Node.TEXT_NODE)
+                self.dom.appendChild(self.dom.createTextNode(data))
+
+        implementation = DomImplementation
+        name = None
+
+    def testSerializer(element):
+        element.normalize()
+        rv = []
+
+        def serializeElement(element, indent=0):
+            if element.nodeType == Node.DOCUMENT_TYPE_NODE:
+                if element.name:
+                    if element.publicId or element.systemId:
+                        publicId = element.publicId or ""
+                        systemId = element.systemId or ""
+                        rv.append("""|%s<!DOCTYPE %s "%s" "%s">""" %
+                                  (' ' * indent, element.name, publicId, systemId))
+                    else:
+                        rv.append("|%s<!DOCTYPE %s>" % (' ' * indent, element.name))
+                else:
+                    rv.append("|%s<!DOCTYPE >" % (' ' * indent,))
+            elif element.nodeType == Node.DOCUMENT_NODE:
+                rv.append("#document")
+            elif element.nodeType == Node.DOCUMENT_FRAGMENT_NODE:
+                rv.append("#document-fragment")
+            elif element.nodeType == Node.COMMENT_NODE:
+                rv.append("|%s<!-- %s -->" % (' ' * indent, element.nodeValue))
+            elif element.nodeType == Node.TEXT_NODE:
+                rv.append("|%s\"%s\"" % (' ' * indent, element.nodeValue))
+            else:
+                if (hasattr(element, "namespaceURI") and
+                        element.namespaceURI is not None):
+                    name = "%s %s" % (constants.prefixes[element.namespaceURI],
+                                      element.nodeName)
+                else:
+                    name = element.nodeName
+                rv.append("|%s<%s>" % (' ' * indent, name))
+                if element.hasAttributes():
+                    attributes = []
+                    for i in range(len(element.attributes)):
+                        attr = element.attributes.item(i)
+                        name = attr.nodeName
+                        value = attr.value
+                        ns = attr.namespaceURI
+                        if ns:
+                            name = "%s %s" % (constants.prefixes[ns], attr.localName)
+                        else:
+                            name = attr.nodeName
+                        attributes.append((name, value))
+
+                    for name, value in sorted(attributes):
+                        rv.append('|%s%s="%s"' % (' ' * (indent + 2), name, value))
+            indent += 2
+            for child in element.childNodes:
+                serializeElement(child, indent)
+        serializeElement(element, 0)
+
+        return "\n".join(rv)
+
+    return locals()
+
+
+# The actual means to get a module!
+getDomModule = moduleFactoryFactory(getDomBuilder)

lib/html5lib/treebuilders/etree.py

@@ -0,0 +1,337 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+
+import re
+
+from . import _base
+from .. import ihatexml
+from .. import constants
+from ..constants import namespaces
+from ..utils import moduleFactoryFactory
+
+tag_regexp = re.compile("{([^}]*)}(.*)")
+
+
+def getETreeBuilder(ElementTreeImplementation, fullTree=False):
+    ElementTree = ElementTreeImplementation
+    ElementTreeCommentType = ElementTree.Comment("asd").tag
+
+    class Element(_base.Node):
+        def __init__(self, name, namespace=None):
+            self._name = name
+            self._namespace = namespace
+            self._element = ElementTree.Element(self._getETreeTag(name,
+                                                                  namespace))
+            if namespace is None:
+                self.nameTuple = namespaces["html"], self._name
+            else:
+                self.nameTuple = self._namespace, self._name
+            self.parent = None
+            self._childNodes = []
+            self._flags = []
+
+        def _getETreeTag(self, name, namespace):
+            if namespace is None:
+                etree_tag = name
+            else:
+                etree_tag = "{%s}%s" % (namespace, name)
+            return etree_tag
+
+        def _setName(self, name):
+            self._name = name
+            self._element.tag = self._getETreeTag(self._name, self._namespace)
+
+        def _getName(self):
+            return self._name
+
+        name = property(_getName, _setName)
+
+        def _setNamespace(self, namespace):
+            self._namespace = namespace
+            self._element.tag = self._getETreeTag(self._name, self._namespace)
+
+        def _getNamespace(self):
+            return self._namespace
+
+        namespace = property(_getNamespace, _setNamespace)
+
+        def _getAttributes(self):
+            return self._element.attrib
+
+        def _setAttributes(self, attributes):
+            # Delete existing attributes first
+            # XXX - there may be a better way to do this...
+            for key in list(self._element.attrib.keys()):
+                del self._element.attrib[key]
+            for key, value in attributes.items():
+                if isinstance(key, tuple):
+                    name = "{%s}%s" % (key[2], key[1])
+                else:
+                    name = key
+                self._element.set(name, value)
+
+        attributes = property(_getAttributes, _setAttributes)
+
+        def _getChildNodes(self):
+            return self._childNodes
+
+        def _setChildNodes(self, value):
+            del self._element[:]
+            self._childNodes = []
+            for element in value:
+                self.insertChild(element)
+
+        childNodes = property(_getChildNodes, _setChildNodes)
+
+        def hasContent(self):
+            """Return true if the node has children or text"""
+            return bool(self._element.text or len(self._element))
+
+        def appendChild(self, node):
+            self._childNodes.append(node)
+            self._element.append(node._element)
+            node.parent = self
+
+        def insertBefore(self, node, refNode):
+            index = list(self._element).index(refNode._element)
+            self._element.insert(index, node._element)
+            node.parent = self
+
+        def removeChild(self, node):
+            self._element.remove(node._element)
+            node.parent = None
+
+        def insertText(self, data, insertBefore=None):
+            if not(len(self._element)):
+                if not self._element.text:
+                    self._element.text = ""
+                self._element.text += data
+            elif insertBefore is None:
+                # Insert the text as the tail of the last child element
+                if not self._element[-1].tail:
+                    self._element[-1].tail = ""
+                self._element[-1].tail += data
+            else:
+                # Insert the text before the specified node
+                children = list(self._element)
+                index = children.index(insertBefore._element)
+                if index > 0:
+                    if not self._element[index - 1].tail:
+                        self._element[index - 1].tail = ""
+                    self._element[index - 1].tail += data
+                else:
+                    if not self._element.text:
+                        self._element.text = ""
+                    self._element.text += data
+
+        def cloneNode(self):
+            element = type(self)(self.name, self.namespace)
+            for name, value in self.attributes.items():
+                element.attributes[name] = value
+            return element
+
+        def reparentChildren(self, newParent):
+            if newParent.childNodes:
+                newParent.childNodes[-1]._element.tail += self._element.text
+            else:
+                if not newParent._element.text:
+                    newParent._element.text = ""
+                if self._element.text is not None:
+                    newParent._element.text += self._element.text
+            self._element.text = ""
+            _base.Node.reparentChildren(self, newParent)
+
+    class Comment(Element):
+        def __init__(self, data):
+            # Use the superclass constructor to set all properties on the
+            # wrapper element
+            self._element = ElementTree.Comment(data)
+            self.parent = None
+            self._childNodes = []
+            self._flags = []
+
+        def _getData(self):
+            return self._element.text
+
+        def _setData(self, value):
+            self._element.text = value
+
+        data = property(_getData, _setData)
+
+    class DocumentType(Element):
+        def __init__(self, name, publicId, systemId):
+            Element.__init__(self, "<!DOCTYPE>")
+            self._element.text = name
+            self.publicId = publicId
+            self.systemId = systemId
+
+        def _getPublicId(self):
+            return self._element.get("publicId", "")
+
+        def _setPublicId(self, value):
+            if value is not None:
+                self._element.set("publicId", value)
+
+        publicId = property(_getPublicId, _setPublicId)
+
+        def _getSystemId(self):
+            return self._element.get("systemId", "")
+
+        def _setSystemId(self, value):
+            if value is not None:
+                self._element.set("systemId", value)
+
+        systemId = property(_getSystemId, _setSystemId)
+
+    class Document(Element):
+        def __init__(self):
+            Element.__init__(self, "DOCUMENT_ROOT")
+
+    class DocumentFragment(Element):
+        def __init__(self):
+            Element.__init__(self, "DOCUMENT_FRAGMENT")
+
+    def testSerializer(element):
+        rv = []
+
+        def serializeElement(element, indent=0):
+            if not(hasattr(element, "tag")):
+                element = element.getroot()
+            if element.tag == "<!DOCTYPE>":
+                if element.get("publicId") or element.get("systemId"):
+                    publicId = element.get("publicId") or ""
+                    systemId = element.get("systemId") or ""
+                    rv.append("""<!DOCTYPE %s "%s" "%s">""" %
+                              (element.text, publicId, systemId))
+                else:
+                    rv.append("<!DOCTYPE %s>" % (element.text,))
+            elif element.tag == "DOCUMENT_ROOT":
+                rv.append("#document")
+                if element.text is not None:
+                    rv.append("|%s\"%s\"" % (' ' * (indent + 2), element.text))
+                if element.tail is not None:
+                    raise TypeError("Document node cannot have tail")
+                if hasattr(element, "attrib") and len(element.attrib):
+                    raise TypeError("Document node cannot have attributes")
+            elif element.tag == ElementTreeCommentType:
+                rv.append("|%s<!-- %s -->" % (' ' * indent, element.text))
+            else:
+                assert isinstance(element.tag, text_type), \
+                    "Expected unicode, got %s, %s" % (type(element.tag), element.tag)
+                nsmatch = tag_regexp.match(element.tag)
+
+                if nsmatch is None:
+                    name = element.tag
+                else:
+                    ns, name = nsmatch.groups()
+                    prefix = constants.prefixes[ns]
+                    name = "%s %s" % (prefix, name)
+                rv.append("|%s<%s>" % (' ' * indent, name))
+
+                if hasattr(element, "attrib"):
+                    attributes = []
+                    for name, value in element.attrib.items():
+                        nsmatch = tag_regexp.match(name)
+                        if nsmatch is not None:
+                            ns, name = nsmatch.groups()
+                            prefix = constants.prefixes[ns]
+                            attr_string = "%s %s" % (prefix, name)
+                        else:
+                            attr_string = name
+                        attributes.append((attr_string, value))
+
+                    for name, value in sorted(attributes):
+                        rv.append('|%s%s="%s"' % (' ' * (indent + 2), name, value))
+                if element.text:
+                    rv.append("|%s\"%s\"" % (' ' * (indent + 2), element.text))
+            indent += 2
+            for child in element:
+                serializeElement(child, indent)
+            if element.tail:
+                rv.append("|%s\"%s\"" % (' ' * (indent - 2), element.tail))
+        serializeElement(element, 0)
+
+        return "\n".join(rv)
+
+    def tostring(element):
+        """Serialize an element and its child nodes to a string"""
+        rv = []
+        filter = ihatexml.InfosetFilter()
+
+        def serializeElement(element):
+            if isinstance(element, ElementTree.ElementTree):
+                element = element.getroot()
+
+            if element.tag == "<!DOCTYPE>":
+                if element.get("publicId") or element.get("systemId"):
+                    publicId = element.get("publicId") or ""
+                    systemId = element.get("systemId") or ""
+                    rv.append("""<!DOCTYPE %s PUBLIC "%s" "%s">""" %
+                              (element.text, publicId, systemId))
+                else:
+                    rv.append("<!DOCTYPE %s>" % (element.text,))
+            elif element.tag == "DOCUMENT_ROOT":
+                if element.text is not None:
+                    rv.append(element.text)
+                if element.tail is not None:
+                    raise TypeError("Document node cannot have tail")
+                if hasattr(element, "attrib") and len(element.attrib):
+                    raise TypeError("Document node cannot have attributes")
+
+                for child in element:
+                    serializeElement(child)
+
+            elif element.tag == ElementTreeCommentType:
+                rv.append("<!--%s-->" % (element.text,))
+            else:
+                # This is assumed to be an ordinary element
+                if not element.attrib:
+                    rv.append("<%s>" % (filter.fromXmlName(element.tag),))
+                else:
+                    attr = " ".join(["%s=\"%s\"" % (
+                        filter.fromXmlName(name), value)
+                        for name, value in element.attrib.items()])
+                    rv.append("<%s %s>" % (element.tag, attr))
+                if element.text:
+                    rv.append(element.text)
+
+                for child in element:
+                    serializeElement(child)
+
+                rv.append("</%s>" % (element.tag,))
+
+            if element.tail:
+                rv.append(element.tail)
+
+        serializeElement(element)
+
+        return "".join(rv)
+
+    class TreeBuilder(_base.TreeBuilder):
+        documentClass = Document
+        doctypeClass = DocumentType
+        elementClass = Element
+        commentClass = Comment
+        fragmentClass = DocumentFragment
+        implementation = ElementTreeImplementation
+
+        def testSerializer(self, element):
+            return testSerializer(element)
+
+        def getDocument(self):
+            if fullTree:
+                return self.document._element
+            else:
+                if self.defaultNamespace is not None:
+                    return self.document._element.find(
+                        "{%s}html" % self.defaultNamespace)
+                else:
+                    return self.document._element.find("html")
+
+        def getFragment(self):
+            return _base.TreeBuilder.getFragment(self)._element
+
+    return locals()
+
+
+getETreeModule = moduleFactoryFactory(getETreeBuilder)

lib/html5lib/treebuilders/etree_lxml.py

@@ -0,0 +1,369 @@
+"""Module for supporting the lxml.etree library. The idea here is to use as much
+of the native library as possible, without using fragile hacks like custom element
+names that break between releases. The downside of this is that we cannot represent
+all possible trees; specifically the following are known to cause problems:
+
+Text or comments as siblings of the root element
+Docypes with no name
+
+When any of these things occur, we emit a DataLossWarning
+"""
+
+from __future__ import absolute_import, division, unicode_literals
+
+import warnings
+import re
+import sys
+
+from . import _base
+from ..constants import DataLossWarning
+from .. import constants
+from . import etree as etree_builders
+from .. import ihatexml
+
+import lxml.etree as etree
+
+
+fullTree = True
+tag_regexp = re.compile("{([^}]*)}(.*)")
+
+comment_type = etree.Comment("asd").tag
+
+
+class DocumentType(object):
+    def __init__(self, name, publicId, systemId):
+        self.name = name
+        self.publicId = publicId
+        self.systemId = systemId
+
+
+class Document(object):
+    def __init__(self):
+        self._elementTree = None
+        self._childNodes = []
+
+    def appendChild(self, element):
+        self._elementTree.getroot().addnext(element._element)
+
+    def _getChildNodes(self):
+        return self._childNodes
+
+    childNodes = property(_getChildNodes)
+
+
+def testSerializer(element):
+    rv = []
+    finalText = None
+    infosetFilter = ihatexml.InfosetFilter()
+
+    def serializeElement(element, indent=0):
+        if not hasattr(element, "tag"):
+            if hasattr(element, "getroot"):
+                # Full tree case
+                rv.append("#document")
+                if element.docinfo.internalDTD:
+                    if not (element.docinfo.public_id or
+                            element.docinfo.system_url):
+                        dtd_str = "<!DOCTYPE %s>" % element.docinfo.root_name
+                    else:
+                        dtd_str = """<!DOCTYPE %s "%s" "%s">""" % (
+                            element.docinfo.root_name,
+                            element.docinfo.public_id,
+                            element.docinfo.system_url)
+                    rv.append("|%s%s" % (' ' * (indent + 2), dtd_str))
+                next_element = element.getroot()
+                while next_element.getprevious() is not None:
+                    next_element = next_element.getprevious()
+                while next_element is not None:
+                    serializeElement(next_element, indent + 2)
+                    next_element = next_element.getnext()
+            elif isinstance(element, str) or isinstance(element, bytes):
+                # Text in a fragment
+                assert isinstance(element, str) or sys.version_info.major == 2
+                rv.append("|%s\"%s\"" % (' ' * indent, element))
+            else:
+                # Fragment case
+                rv.append("#document-fragment")
+                for next_element in element:
+                    serializeElement(next_element, indent + 2)
+        elif element.tag == comment_type:
+            rv.append("|%s<!-- %s -->" % (' ' * indent, element.text))
+            if hasattr(element, "tail") and element.tail:
+                rv.append("|%s\"%s\"" % (' ' * indent, element.tail))
+        else:
+            assert isinstance(element, etree._Element)
+            nsmatch = etree_builders.tag_regexp.match(element.tag)
+            if nsmatch is not None:
+                ns = nsmatch.group(1)
+                tag = nsmatch.group(2)
+                prefix = constants.prefixes[ns]
+                rv.append("|%s<%s %s>" % (' ' * indent, prefix,
+                                          infosetFilter.fromXmlName(tag)))
+            else:
+                rv.append("|%s<%s>" % (' ' * indent,
+                                       infosetFilter.fromXmlName(element.tag)))
+
+            if hasattr(element, "attrib"):
+                attributes = []
+                for name, value in element.attrib.items():
+                    nsmatch = tag_regexp.match(name)
+                    if nsmatch is not None:
+                        ns, name = nsmatch.groups()
+                        name = infosetFilter.fromXmlName(name)
+                        prefix = constants.prefixes[ns]
+                        attr_string = "%s %s" % (prefix, name)
+                    else:
+                        attr_string = infosetFilter.fromXmlName(name)
+                    attributes.append((attr_string, value))
+
+                for name, value in sorted(attributes):
+                    rv.append('|%s%s="%s"' % (' ' * (indent + 2), name, value))
+
+            if element.text:
+                rv.append("|%s\"%s\"" % (' ' * (indent + 2), element.text))
+            indent += 2
+            for child in element:
+                serializeElement(child, indent)
+            if hasattr(element, "tail") and element.tail:
+                rv.append("|%s\"%s\"" % (' ' * (indent - 2), element.tail))
+    serializeElement(element, 0)
+
+    if finalText is not None:
+        rv.append("|%s\"%s\"" % (' ' * 2, finalText))
+
+    return "\n".join(rv)
+
+
+def tostring(element):
+    """Serialize an element and its child nodes to a string"""
+    rv = []
+    finalText = None
+
+    def serializeElement(element):
+        if not hasattr(element, "tag"):
+            if element.docinfo.internalDTD:
+                if element.docinfo.doctype:
+                    dtd_str = element.docinfo.doctype
+                else:
+                    dtd_str = "<!DOCTYPE %s>" % element.docinfo.root_name
+                rv.append(dtd_str)
+            serializeElement(element.getroot())
+
+        elif element.tag == comment_type:
+            rv.append("<!--%s-->" % (element.text,))
+
+        else:
+            # This is assumed to be an ordinary element
+            if not element.attrib:
+                rv.append("<%s>" % (element.tag,))
+            else:
+                attr = " ".join(["%s=\"%s\"" % (name, value)
+                                 for name, value in element.attrib.items()])
+                rv.append("<%s %s>" % (element.tag, attr))
+            if element.text:
+                rv.append(element.text)
+
+            for child in element:
+                serializeElement(child)
+
+            rv.append("</%s>" % (element.tag,))
+
+        if hasattr(element, "tail") and element.tail:
+            rv.append(element.tail)
+
+    serializeElement(element)
+
+    if finalText is not None:
+        rv.append("%s\"" % (' ' * 2, finalText))
+
+    return "".join(rv)
+
+
+class TreeBuilder(_base.TreeBuilder):
+    documentClass = Document
+    doctypeClass = DocumentType
+    elementClass = None
+    commentClass = None
+    fragmentClass = Document
+    implementation = etree
+
+    def __init__(self, namespaceHTMLElements, fullTree=False):
+        builder = etree_builders.getETreeModule(etree, fullTree=fullTree)
+        infosetFilter = self.infosetFilter = ihatexml.InfosetFilter()
+        self.namespaceHTMLElements = namespaceHTMLElements
+
+        class Attributes(dict):
+            def __init__(self, element, value={}):
+                self._element = element
+                dict.__init__(self, value)
+                for key, value in self.items():
+                    if isinstance(key, tuple):
+                        name = "{%s}%s" % (key[2], infosetFilter.coerceAttribute(key[1]))
+                    else:
+                        name = infosetFilter.coerceAttribute(key)
+                    self._element._element.attrib[name] = value
+
+            def __setitem__(self, key, value):
+                dict.__setitem__(self, key, value)
+                if isinstance(key, tuple):
+                    name = "{%s}%s" % (key[2], infosetFilter.coerceAttribute(key[1]))
+                else:
+                    name = infosetFilter.coerceAttribute(key)
+                self._element._element.attrib[name] = value
+
+        class Element(builder.Element):
+            def __init__(self, name, namespace):
+                name = infosetFilter.coerceElement(name)
+                builder.Element.__init__(self, name, namespace=namespace)
+                self._attributes = Attributes(self)
+
+            def _setName(self, name):
+                self._name = infosetFilter.coerceElement(name)
+                self._element.tag = self._getETreeTag(
+                    self._name, self._namespace)
+
+            def _getName(self):
+                return infosetFilter.fromXmlName(self._name)
+
+            name = property(_getName, _setName)
+
+            def _getAttributes(self):
+                return self._attributes
+
+            def _setAttributes(self, attributes):
+                self._attributes = Attributes(self, attributes)
+
+            attributes = property(_getAttributes, _setAttributes)
+
+            def insertText(self, data, insertBefore=None):
+                data = infosetFilter.coerceCharacters(data)
+                builder.Element.insertText(self, data, insertBefore)
+
+            def appendChild(self, child):
+                builder.Element.appendChild(self, child)
+
+        class Comment(builder.Comment):
+            def __init__(self, data):
+                data = infosetFilter.coerceComment(data)
+                builder.Comment.__init__(self, data)
+
+            def _setData(self, data):
+                data = infosetFilter.coerceComment(data)
+                self._element.text = data
+
+            def _getData(self):
+                return self._element.text
+
+            data = property(_getData, _setData)
+
+        self.elementClass = Element
+        self.commentClass = builder.Comment
+        # self.fragmentClass = builder.DocumentFragment
+        _base.TreeBuilder.__init__(self, namespaceHTMLElements)
+
+    def reset(self):
+        _base.TreeBuilder.reset(self)
+        self.insertComment = self.insertCommentInitial
+        self.initial_comments = []
+        self.doctype = None
+
+    def testSerializer(self, element):
+        return testSerializer(element)
+
+    def getDocument(self):
+        if fullTree:
+            return self.document._elementTree
+        else:
+            return self.document._elementTree.getroot()
+
+    def getFragment(self):
+        fragment = []
+        element = self.openElements[0]._element
+        if element.text:
+            fragment.append(element.text)
+        fragment.extend(list(element))
+        if element.tail:
+            fragment.append(element.tail)
+        return fragment
+
+    def insertDoctype(self, token):
+        name = token["name"]
+        publicId = token["publicId"]
+        systemId = token["systemId"]
+
+        if not name:
+            warnings.warn("lxml cannot represent empty doctype", DataLossWarning)
+            self.doctype = None
+        else:
+            coercedName = self.infosetFilter.coerceElement(name)
+            if coercedName != name:
+                warnings.warn("lxml cannot represent non-xml doctype", DataLossWarning)
+
+            doctype = self.doctypeClass(coercedName, publicId, systemId)
+            self.doctype = doctype
+
+    def insertCommentInitial(self, data, parent=None):
+        self.initial_comments.append(data)
+
+    def insertCommentMain(self, data, parent=None):
+        if (parent == self.document and
+                self.document._elementTree.getroot()[-1].tag == comment_type):
+                warnings.warn("lxml cannot represent adjacent comments beyond the root elements", DataLossWarning)
+        super(TreeBuilder, self).insertComment(data, parent)
+
+    def insertRoot(self, token):
+        """Create the document root"""
+        # Because of the way libxml2 works, it doesn't seem to be possible to
+        # alter information like the doctype after the tree has been parsed.
+        # Therefore we need to use the built-in parser to create our iniial
+        # tree, after which we can add elements like normal
+        docStr = ""
+        if self.doctype:
+            assert self.doctype.name
+            docStr += "<!DOCTYPE %s" % self.doctype.name
+            if (self.doctype.publicId is not None or
+                    self.doctype.systemId is not None):
+                docStr += (' PUBLIC "%s" ' %
+                           (self.infosetFilter.coercePubid(self.doctype.publicId or "")))
+                if self.doctype.systemId:
+                    sysid = self.doctype.systemId
+                    if sysid.find("'") >= 0 and sysid.find('"') >= 0:
+                        warnings.warn("DOCTYPE system cannot contain single and double quotes", DataLossWarning)
+                        sysid = sysid.replace("'", 'U00027')
+                    if sysid.find("'") >= 0:
+                        docStr += '"%s"' % sysid
+                    else:
+                        docStr += "'%s'" % sysid
+                else:
+                    docStr += "''"
+            docStr += ">"
+            if self.doctype.name != token["name"]:
+                warnings.warn("lxml cannot represent doctype with a different name to the root element", DataLossWarning)
+        docStr += "<THIS_SHOULD_NEVER_APPEAR_PUBLICLY/>"
+        root = etree.fromstring(docStr)
+
+        # Append the initial comments:
+        for comment_token in self.initial_comments:
+            root.addprevious(etree.Comment(comment_token["data"]))
+
+        # Create the root document and add the ElementTree to it
+        self.document = self.documentClass()
+        self.document._elementTree = root.getroottree()
+
+        # Give the root element the right name
+        name = token["name"]
+        namespace = token.get("namespace", self.defaultNamespace)
+        if namespace is None:
+            etree_tag = name
+        else:
+            etree_tag = "{%s}%s" % (namespace, name)
+        root.tag = etree_tag
+
+        # Add the root element to the internal child/open data structures
+        root_element = self.elementClass(name, namespace)
+        root_element._element = root
+        self.document._childNodes.append(root_element)
+        self.openElements.append(root_element)
+
+        # Reset to the default insert comment function
+        self.insertComment = self.insertCommentMain

lib/html5lib/treewalkers/__init__.py

@@ -0,0 +1,57 @@
+"""A collection of modules for iterating through different kinds of
+tree, generating tokens identical to those produced by the tokenizer
+module.
+
+To create a tree walker for a new type of tree, you need to do
+implement a tree walker object (called TreeWalker by convention) that
+implements a 'serialize' method taking a tree as sole argument and
+returning an iterator generating tokens.
+"""
+
+from __future__ import absolute_import, division, unicode_literals
+
+import sys
+
+from ..utils import default_etree
+
+treeWalkerCache = {}
+
+
+def getTreeWalker(treeType, implementation=None, **kwargs):
+    """Get a TreeWalker class for various types of tree with built-in support
+
+    treeType - the name of the tree type required (case-insensitive). Supported
+               values are:
+
+                "dom" - The xml.dom.minidom DOM implementation
+                "pulldom" - The xml.dom.pulldom event stream
+                "etree" - A generic walker for tree implementations exposing an
+                          elementtree-like interface (known to work with
+                          ElementTree, cElementTree and lxml.etree).
+                "lxml" - Optimized walker for lxml.etree
+                "genshi" - a Genshi stream
+
+    implementation - (Currently applies to the "etree" tree type only). A module
+                      implementing the tree type e.g. xml.etree.ElementTree or
+                      cElementTree."""
+
+    treeType = treeType.lower()
+    if treeType not in treeWalkerCache:
+        if treeType in ("dom", "pulldom"):
+            name = "%s.%s" % (__name__, treeType)
+            __import__(name)
+            mod = sys.modules[name]
+            treeWalkerCache[treeType] = mod.TreeWalker
+        elif treeType == "genshi":
+            from . import genshistream
+            treeWalkerCache[treeType] = genshistream.TreeWalker
+        elif treeType == "lxml":
+            from . import lxmletree
+            treeWalkerCache[treeType] = lxmletree.TreeWalker
+        elif treeType == "etree":
+            from . import etree
+            if implementation is None:
+                implementation = default_etree
+            # XXX: NEVER cache here, caching is done in the etree submodule
+            return etree.getETreeModule(implementation, **kwargs).TreeWalker
+    return treeWalkerCache.get(treeType)

lib/html5lib/treewalkers/_base.py

@@ -0,0 +1,200 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type, string_types
+
+import gettext
+_ = gettext.gettext
+
+from xml.dom import Node
+
+DOCUMENT = Node.DOCUMENT_NODE
+DOCTYPE = Node.DOCUMENT_TYPE_NODE
+TEXT = Node.TEXT_NODE
+ELEMENT = Node.ELEMENT_NODE
+COMMENT = Node.COMMENT_NODE
+ENTITY = Node.ENTITY_NODE
+UNKNOWN = "<#UNKNOWN#>"
+
+from ..constants import voidElements, spaceCharacters
+spaceCharacters = "".join(spaceCharacters)
+
+
+def to_text(s, blank_if_none=True):
+    """Wrapper around six.text_type to convert None to empty string"""
+    if s is None:
+        if blank_if_none:
+            return ""
+        else:
+            return None
+    elif isinstance(s, text_type):
+        return s
+    else:
+        return text_type(s)
+
+
+def is_text_or_none(string):
+    """Wrapper around isinstance(string_types) or is None"""
+    return string is None or isinstance(string, string_types)
+
+
+class TreeWalker(object):
+    def __init__(self, tree):
+        self.tree = tree
+
+    def __iter__(self):
+        raise NotImplementedError
+
+    def error(self, msg):
+        return {"type": "SerializeError", "data": msg}
+
+    def emptyTag(self, namespace, name, attrs, hasChildren=False):
+        assert namespace is None or isinstance(namespace, string_types), type(namespace)
+        assert isinstance(name, string_types), type(name)
+        assert all((namespace is None or isinstance(namespace, string_types)) and
+                   isinstance(name, string_types) and
+                   isinstance(value, string_types)
+                   for (namespace, name), value in attrs.items())
+
+        yield {"type": "EmptyTag", "name": to_text(name, False),
+               "namespace": to_text(namespace),
+               "data": attrs}
+        if hasChildren:
+            yield self.error(_("Void element has children"))
+
+    def startTag(self, namespace, name, attrs):
+        assert namespace is None or isinstance(namespace, string_types), type(namespace)
+        assert isinstance(name, string_types), type(name)
+        assert all((namespace is None or isinstance(namespace, string_types)) and
+                   isinstance(name, string_types) and
+                   isinstance(value, string_types)
+                   for (namespace, name), value in attrs.items())
+
+        return {"type": "StartTag",
+                "name": text_type(name),
+                "namespace": to_text(namespace),
+                "data": dict(((to_text(namespace, False), to_text(name)),
+                              to_text(value, False))
+                             for (namespace, name), value in attrs.items())}
+
+    def endTag(self, namespace, name):
+        assert namespace is None or isinstance(namespace, string_types), type(namespace)
+        assert isinstance(name, string_types), type(namespace)
+
+        return {"type": "EndTag",
+                "name": to_text(name, False),
+                "namespace": to_text(namespace),
+                "data": {}}
+
+    def text(self, data):
+        assert isinstance(data, string_types), type(data)
+
+        data = to_text(data)
+        middle = data.lstrip(spaceCharacters)
+        left = data[:len(data) - len(middle)]
+        if left:
+            yield {"type": "SpaceCharacters", "data": left}
+        data = middle
+        middle = data.rstrip(spaceCharacters)
+        right = data[len(middle):]
+        if middle:
+            yield {"type": "Characters", "data": middle}
+        if right:
+            yield {"type": "SpaceCharacters", "data": right}
+
+    def comment(self, data):
+        assert isinstance(data, string_types), type(data)
+
+        return {"type": "Comment", "data": text_type(data)}
+
+    def doctype(self, name, publicId=None, systemId=None, correct=True):
+        assert is_text_or_none(name), type(name)
+        assert is_text_or_none(publicId), type(publicId)
+        assert is_text_or_none(systemId), type(systemId)
+
+        return {"type": "Doctype",
+                "name": to_text(name),
+                "publicId": to_text(publicId),
+                "systemId": to_text(systemId),
+                "correct": to_text(correct)}
+
+    def entity(self, name):
+        assert isinstance(name, string_types), type(name)
+
+        return {"type": "Entity", "name": text_type(name)}
+
+    def unknown(self, nodeType):
+        return self.error(_("Unknown node type: ") + nodeType)
+
+
+class NonRecursiveTreeWalker(TreeWalker):
+    def getNodeDetails(self, node):
+        raise NotImplementedError
+
+    def getFirstChild(self, node):
+        raise NotImplementedError
+
+    def getNextSibling(self, node):
+        raise NotImplementedError
+
+    def getParentNode(self, node):
+        raise NotImplementedError
+
+    def __iter__(self):
+        currentNode = self.tree
+        while currentNode is not None:
+            details = self.getNodeDetails(currentNode)
+            type, details = details[0], details[1:]
+            hasChildren = False
+
+            if type == DOCTYPE:
+                yield self.doctype(*details)
+
+            elif type == TEXT:
+                for token in self.text(*details):
+                    yield token
+
+            elif type == ELEMENT:
+                namespace, name, attributes, hasChildren = details
+                if name in voidElements:
+                    for token in self.emptyTag(namespace, name, attributes,
+                                               hasChildren):
+                        yield token
+                    hasChildren = False
+                else:
+                    yield self.startTag(namespace, name, attributes)
+
+            elif type == COMMENT:
+                yield self.comment(details[0])
+
+            elif type == ENTITY:
+                yield self.entity(details[0])
+
+            elif type == DOCUMENT:
+                hasChildren = True
+
+            else:
+                yield self.unknown(details[0])
+
+            if hasChildren:
+                firstChild = self.getFirstChild(currentNode)
+            else:
+                firstChild = None
+
+            if firstChild is not None:
+                currentNode = firstChild
+            else:
+                while currentNode is not None:
+                    details = self.getNodeDetails(currentNode)
+                    type, details = details[0], details[1:]
+                    if type == ELEMENT:
+                        namespace, name, attributes, hasChildren = details
+                        if name not in voidElements:
+                            yield self.endTag(namespace, name)
+                    if self.tree is currentNode:
+                        currentNode = None
+                        break
+                    nextSibling = self.getNextSibling(currentNode)
+                    if nextSibling is not None:
+                        currentNode = nextSibling
+                        break
+                    else:
+                        currentNode = self.getParentNode(currentNode)

lib/html5lib/treewalkers/dom.py

@@ -0,0 +1,46 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from xml.dom import Node
+
+import gettext
+_ = gettext.gettext
+
+from . import _base
+
+
+class TreeWalker(_base.NonRecursiveTreeWalker):
+    def getNodeDetails(self, node):
+        if node.nodeType == Node.DOCUMENT_TYPE_NODE:
+            return _base.DOCTYPE, node.name, node.publicId, node.systemId
+
+        elif node.nodeType in (Node.TEXT_NODE, Node.CDATA_SECTION_NODE):
+            return _base.TEXT, node.nodeValue
+
+        elif node.nodeType == Node.ELEMENT_NODE:
+            attrs = {}
+            for attr in list(node.attributes.keys()):
+                attr = node.getAttributeNode(attr)
+                if attr.namespaceURI:
+                    attrs[(attr.namespaceURI, attr.localName)] = attr.value
+                else:
+                    attrs[(None, attr.name)] = attr.value
+            return (_base.ELEMENT, node.namespaceURI, node.nodeName,
+                    attrs, node.hasChildNodes())
+
+        elif node.nodeType == Node.COMMENT_NODE:
+            return _base.COMMENT, node.nodeValue
+
+        elif node.nodeType in (Node.DOCUMENT_NODE, Node.DOCUMENT_FRAGMENT_NODE):
+            return (_base.DOCUMENT,)
+
+        else:
+            return _base.UNKNOWN, node.nodeType
+
+    def getFirstChild(self, node):
+        return node.firstChild
+
+    def getNextSibling(self, node):
+        return node.nextSibling
+
+    def getParentNode(self, node):
+        return node.parentNode

lib/html5lib/treewalkers/etree.py

@@ -0,0 +1,138 @@
+from __future__ import absolute_import, division, unicode_literals
+
+try:
+    from collections import OrderedDict
+except ImportError:
+    try:
+        from ordereddict import OrderedDict
+    except ImportError:
+        OrderedDict = dict
+import gettext
+_ = gettext.gettext
+
+import re
+
+from six import text_type
+
+from . import _base
+from ..utils import moduleFactoryFactory
+
+tag_regexp = re.compile("{([^}]*)}(.*)")
+
+
+def getETreeBuilder(ElementTreeImplementation):
+    ElementTree = ElementTreeImplementation
+    ElementTreeCommentType = ElementTree.Comment("asd").tag
+
+    class TreeWalker(_base.NonRecursiveTreeWalker):
+        """Given the particular ElementTree representation, this implementation,
+        to avoid using recursion, returns "nodes" as tuples with the following
+        content:
+
+        1. The current element
+
+        2. The index of the element relative to its parent
+
+        3. A stack of ancestor elements
+
+        4. A flag "text", "tail" or None to indicate if the current node is a
+           text node; either the text or tail of the current element (1)
+        """
+        def getNodeDetails(self, node):
+            if isinstance(node, tuple):  # It might be the root Element
+                elt, key, parents, flag = node
+                if flag in ("text", "tail"):
+                    return _base.TEXT, getattr(elt, flag)
+                else:
+                    node = elt
+
+            if not(hasattr(node, "tag")):
+                node = node.getroot()
+
+            if node.tag in ("DOCUMENT_ROOT", "DOCUMENT_FRAGMENT"):
+                return (_base.DOCUMENT,)
+
+            elif node.tag == "<!DOCTYPE>":
+                return (_base.DOCTYPE, node.text,
+                        node.get("publicId"), node.get("systemId"))
+
+            elif node.tag == ElementTreeCommentType:
+                return _base.COMMENT, node.text
+
+            else:
+                assert type(node.tag) == text_type, type(node.tag)
+                # This is assumed to be an ordinary element
+                match = tag_regexp.match(node.tag)
+                if match:
+                    namespace, tag = match.groups()
+                else:
+                    namespace = None
+                    tag = node.tag
+                attrs = OrderedDict()
+                for name, value in list(node.attrib.items()):
+                    match = tag_regexp.match(name)
+                    if match:
+                        attrs[(match.group(1), match.group(2))] = value
+                    else:
+                        attrs[(None, name)] = value
+                return (_base.ELEMENT, namespace, tag,
+                        attrs, len(node) or node.text)
+
+        def getFirstChild(self, node):
+            if isinstance(node, tuple):
+                element, key, parents, flag = node
+            else:
+                element, key, parents, flag = node, None, [], None
+
+            if flag in ("text", "tail"):
+                return None
+            else:
+                if element.text:
+                    return element, key, parents, "text"
+                elif len(element):
+                    parents.append(element)
+                    return element[0], 0, parents, None
+                else:
+                    return None
+
+        def getNextSibling(self, node):
+            if isinstance(node, tuple):
+                element, key, parents, flag = node
+            else:
+                return None
+
+            if flag == "text":
+                if len(element):
+                    parents.append(element)
+                    return element[0], 0, parents, None
+                else:
+                    return None
+            else:
+                if element.tail and flag != "tail":
+                    return element, key, parents, "tail"
+                elif key < len(parents[-1]) - 1:
+                    return parents[-1][key + 1], key + 1, parents, None
+                else:
+                    return None
+
+        def getParentNode(self, node):
+            if isinstance(node, tuple):
+                element, key, parents, flag = node
+            else:
+                return None
+
+            if flag == "text":
+                if not parents:
+                    return element
+                else:
+                    return element, key, parents, None
+            else:
+                parent = parents.pop()
+                if not parents:
+                    return parent
+                else:
+                    return parent, list(parents[-1]).index(parent), parents, None
+
+    return locals()
+
+getETreeModule = moduleFactoryFactory(getETreeBuilder)

lib/html5lib/treewalkers/genshistream.py

@@ -0,0 +1,69 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from genshi.core import QName
+from genshi.core import START, END, XML_NAMESPACE, DOCTYPE, TEXT
+from genshi.core import START_NS, END_NS, START_CDATA, END_CDATA, PI, COMMENT
+
+from . import _base
+
+from ..constants import voidElements, namespaces
+
+
+class TreeWalker(_base.TreeWalker):
+    def __iter__(self):
+        # Buffer the events so we can pass in the following one
+        previous = None
+        for event in self.tree:
+            if previous is not None:
+                for token in self.tokens(previous, event):
+                    yield token
+            previous = event
+
+        # Don't forget the final event!
+        if previous is not None:
+            for token in self.tokens(previous, None):
+                yield token
+
+    def tokens(self, event, next):
+        kind, data, pos = event
+        if kind == START:
+            tag, attribs = data
+            name = tag.localname
+            namespace = tag.namespace
+            converted_attribs = {}
+            for k, v in attribs:
+                if isinstance(k, QName):
+                    converted_attribs[(k.namespace, k.localname)] = v
+                else:
+                    converted_attribs[(None, k)] = v
+
+            if namespace == namespaces["html"] and name in voidElements:
+                for token in self.emptyTag(namespace, name, converted_attribs,
+                                           not next or next[0] != END
+                                           or next[1] != tag):
+                    yield token
+            else:
+                yield self.startTag(namespace, name, converted_attribs)
+
+        elif kind == END:
+            name = data.localname
+            namespace = data.namespace
+            if name not in voidElements:
+                yield self.endTag(namespace, name)
+
+        elif kind == COMMENT:
+            yield self.comment(data)
+
+        elif kind == TEXT:
+            for token in self.text(data):
+                yield token
+
+        elif kind == DOCTYPE:
+            yield self.doctype(*data)
+
+        elif kind in (XML_NAMESPACE, DOCTYPE, START_NS, END_NS,
+                      START_CDATA, END_CDATA, PI):
+            pass
+
+        else:
+            yield self.unknown(kind)

lib/html5lib/treewalkers/lxmletree.py

@@ -0,0 +1,204 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+
+from lxml import etree
+from ..treebuilders.etree import tag_regexp
+
+from gettext import gettext
+_ = gettext
+
+from . import _base
+
+from .. import ihatexml
+
+
+def ensure_str(s):
+    if s is None:
+        return None
+    elif isinstance(s, text_type):
+        return s
+    else:
+        return s.decode("utf-8", "strict")
+
+
+class Root(object):
+    def __init__(self, et):
+        self.elementtree = et
+        self.children = []
+        if et.docinfo.internalDTD:
+            self.children.append(Doctype(self,
+                                         ensure_str(et.docinfo.root_name),
+                                         ensure_str(et.docinfo.public_id),
+                                         ensure_str(et.docinfo.system_url)))
+        root = et.getroot()
+        node = root
+
+        while node.getprevious() is not None:
+            node = node.getprevious()
+        while node is not None:
+            self.children.append(node)
+            node = node.getnext()
+
+        self.text = None
+        self.tail = None
+
+    def __getitem__(self, key):
+        return self.children[key]
+
+    def getnext(self):
+        return None
+
+    def __len__(self):
+        return 1
+
+
+class Doctype(object):
+    def __init__(self, root_node, name, public_id, system_id):
+        self.root_node = root_node
+        self.name = name
+        self.public_id = public_id
+        self.system_id = system_id
+
+        self.text = None
+        self.tail = None
+
+    def getnext(self):
+        return self.root_node.children[1]
+
+
+class FragmentRoot(Root):
+    def __init__(self, children):
+        self.children = [FragmentWrapper(self, child) for child in children]
+        self.text = self.tail = None
+
+    def getnext(self):
+        return None
+
+
+class FragmentWrapper(object):
+    def __init__(self, fragment_root, obj):
+        self.root_node = fragment_root
+        self.obj = obj
+        if hasattr(self.obj, 'text'):
+            self.text = ensure_str(self.obj.text)
+        else:
+            self.text = None
+        if hasattr(self.obj, 'tail'):
+            self.tail = ensure_str(self.obj.tail)
+        else:
+            self.tail = None
+
+    def __getattr__(self, name):
+        return getattr(self.obj, name)
+
+    def getnext(self):
+        siblings = self.root_node.children
+        idx = siblings.index(self)
+        if idx < len(siblings) - 1:
+            return siblings[idx + 1]
+        else:
+            return None
+
+    def __getitem__(self, key):
+        return self.obj[key]
+
+    def __bool__(self):
+        return bool(self.obj)
+
+    def getparent(self):
+        return None
+
+    def __str__(self):
+        return str(self.obj)
+
+    def __unicode__(self):
+        return str(self.obj)
+
+    def __len__(self):
+        return len(self.obj)
+
+
+class TreeWalker(_base.NonRecursiveTreeWalker):
+    def __init__(self, tree):
+        if hasattr(tree, "getroot"):
+            tree = Root(tree)
+        elif isinstance(tree, list):
+            tree = FragmentRoot(tree)
+        _base.NonRecursiveTreeWalker.__init__(self, tree)
+        self.filter = ihatexml.InfosetFilter()
+
+    def getNodeDetails(self, node):
+        if isinstance(node, tuple):  # Text node
+            node, key = node
+            assert key in ("text", "tail"), _("Text nodes are text or tail, found %s") % key
+            return _base.TEXT, ensure_str(getattr(node, key))
+
+        elif isinstance(node, Root):
+            return (_base.DOCUMENT,)
+
+        elif isinstance(node, Doctype):
+            return _base.DOCTYPE, node.name, node.public_id, node.system_id
+
+        elif isinstance(node, FragmentWrapper) and not hasattr(node, "tag"):
+            return _base.TEXT, node.obj
+
+        elif node.tag == etree.Comment:
+            return _base.COMMENT, ensure_str(node.text)
+
+        elif node.tag == etree.Entity:
+            return _base.ENTITY, ensure_str(node.text)[1:-1]  # strip &;
+
+        else:
+            # This is assumed to be an ordinary element
+            match = tag_regexp.match(ensure_str(node.tag))
+            if match:
+                namespace, tag = match.groups()
+            else:
+                namespace = None
+                tag = ensure_str(node.tag)
+            attrs = {}
+            for name, value in list(node.attrib.items()):
+                name = ensure_str(name)
+                value = ensure_str(value)
+                match = tag_regexp.match(name)
+                if match:
+                    attrs[(match.group(1), match.group(2))] = value
+                else:
+                    attrs[(None, name)] = value
+            return (_base.ELEMENT, namespace, self.filter.fromXmlName(tag),
+                    attrs, len(node) > 0 or node.text)
+
+    def getFirstChild(self, node):
+        assert not isinstance(node, tuple), _("Text nodes have no children")
+
+        assert len(node) or node.text, "Node has no children"
+        if node.text:
+            return (node, "text")
+        else:
+            return node[0]
+
+    def getNextSibling(self, node):
+        if isinstance(node, tuple):  # Text node
+            node, key = node
+            assert key in ("text", "tail"), _("Text nodes are text or tail, found %s") % key
+            if key == "text":
+                # XXX: we cannot use a "bool(node) and node[0] or None" construct here
+                # because node[0] might evaluate to False if it has no child element
+                if len(node):
+                    return node[0]
+                else:
+                    return None
+            else:  # tail
+                return node.getnext()
+
+        return (node, "tail") if node.tail else node.getnext()
+
+    def getParentNode(self, node):
+        if isinstance(node, tuple):  # Text node
+            node, key = node
+            assert key in ("text", "tail"), _("Text nodes are text or tail, found %s") % key
+            if key == "text":
+                return node
+            # else: fallback to "normal" processing
+
+        return node.getparent()

lib/html5lib/treewalkers/pulldom.py

@@ -0,0 +1,63 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from xml.dom.pulldom import START_ELEMENT, END_ELEMENT, \
+    COMMENT, IGNORABLE_WHITESPACE, CHARACTERS
+
+from . import _base
+
+from ..constants import voidElements
+
+
+class TreeWalker(_base.TreeWalker):
+    def __iter__(self):
+        ignore_until = None
+        previous = None
+        for event in self.tree:
+            if previous is not None and \
+                    (ignore_until is None or previous[1] is ignore_until):
+                if previous[1] is ignore_until:
+                    ignore_until = None
+                for token in self.tokens(previous, event):
+                    yield token
+                    if token["type"] == "EmptyTag":
+                        ignore_until = previous[1]
+            previous = event
+        if ignore_until is None or previous[1] is ignore_until:
+            for token in self.tokens(previous, None):
+                yield token
+        elif ignore_until is not None:
+            raise ValueError("Illformed DOM event stream: void element without END_ELEMENT")
+
+    def tokens(self, event, next):
+        type, node = event
+        if type == START_ELEMENT:
+            name = node.nodeName
+            namespace = node.namespaceURI
+            attrs = {}
+            for attr in list(node.attributes.keys()):
+                attr = node.getAttributeNode(attr)
+                attrs[(attr.namespaceURI, attr.localName)] = attr.value
+            if name in voidElements:
+                for token in self.emptyTag(namespace,
+                                           name,
+                                           attrs,
+                                           not next or next[1] is not node):
+                    yield token
+            else:
+                yield self.startTag(namespace, name, attrs)
+
+        elif type == END_ELEMENT:
+            name = node.nodeName
+            namespace = node.namespaceURI
+            if name not in voidElements:
+                yield self.endTag(namespace, name)
+
+        elif type == COMMENT:
+            yield self.comment(node.nodeValue)
+
+        elif type in (IGNORABLE_WHITESPACE, CHARACTERS):
+            for token in self.text(node.nodeValue):
+                yield token
+
+        else:
+            yield self.unknown(type)

lib/html5lib/trie/__init__.py

@@ -0,0 +1,12 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from .py import Trie as PyTrie
+
+Trie = PyTrie
+
+try:
+    from .datrie import Trie as DATrie
+except ImportError:
+    pass
+else:
+    Trie = DATrie

lib/html5lib/trie/_base.py

@@ -0,0 +1,37 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from collections import Mapping
+
+
+class Trie(Mapping):
+    """Abstract base class for tries"""
+
+    def keys(self, prefix=None):
+        keys = super().keys()
+
+        if prefix is None:
+            return set(keys)
+
+        # Python 2.6: no set comprehensions
+        return set([x for x in keys if x.startswith(prefix)])
+
+    def has_keys_with_prefix(self, prefix):
+        for key in self.keys():
+            if key.startswith(prefix):
+                return True
+
+        return False
+
+    def longest_prefix(self, prefix):
+        if prefix in self:
+            return prefix
+
+        for i in range(1, len(prefix) + 1):
+            if prefix[:-i] in self:
+                return prefix[:-i]
+
+        raise KeyError(prefix)
+
+    def longest_prefix_item(self, prefix):
+        lprefix = self.longest_prefix(prefix)
+        return (lprefix, self[lprefix])

lib/html5lib/trie/datrie.py

@@ -0,0 +1,44 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from datrie import Trie as DATrie
+from six import text_type
+
+from ._base import Trie as ABCTrie
+
+
+class Trie(ABCTrie):
+    def __init__(self, data):
+        chars = set()
+        for key in data.keys():
+            if not isinstance(key, text_type):
+                raise TypeError("All keys must be strings")
+            for char in key:
+                chars.add(char)
+
+        self._data = DATrie("".join(chars))
+        for key, value in data.items():
+            self._data[key] = value
+
+    def __contains__(self, key):
+        return key in self._data
+
+    def __len__(self):
+        return len(self._data)
+
+    def __iter__(self):
+        raise NotImplementedError()
+
+    def __getitem__(self, key):
+        return self._data[key]
+
+    def keys(self, prefix=None):
+        return self._data.keys(prefix)
+
+    def has_keys_with_prefix(self, prefix):
+        return self._data.has_keys_with_prefix(prefix)
+
+    def longest_prefix(self, prefix):
+        return self._data.longest_prefix(prefix)
+
+    def longest_prefix_item(self, prefix):
+        return self._data.longest_prefix_item(prefix)

lib/html5lib/trie/py.py

@@ -0,0 +1,67 @@
+from __future__ import absolute_import, division, unicode_literals
+from six import text_type
+
+from bisect import bisect_left
+
+from ._base import Trie as ABCTrie
+
+
+class Trie(ABCTrie):
+    def __init__(self, data):
+        if not all(isinstance(x, text_type) for x in data.keys()):
+            raise TypeError("All keys must be strings")
+
+        self._data = data
+        self._keys = sorted(data.keys())
+        self._cachestr = ""
+        self._cachepoints = (0, len(data))
+
+    def __contains__(self, key):
+        return key in self._data
+
+    def __len__(self):
+        return len(self._data)
+
+    def __iter__(self):
+        return iter(self._data)
+
+    def __getitem__(self, key):
+        return self._data[key]
+
+    def keys(self, prefix=None):
+        if prefix is None or prefix == "" or not self._keys:
+            return set(self._keys)
+
+        if prefix.startswith(self._cachestr):
+            lo, hi = self._cachepoints
+            start = i = bisect_left(self._keys, prefix, lo, hi)
+        else:
+            start = i = bisect_left(self._keys, prefix)
+
+        keys = set()
+        if start == len(self._keys):
+            return keys
+
+        while self._keys[i].startswith(prefix):
+            keys.add(self._keys[i])
+            i += 1
+
+        self._cachestr = prefix
+        self._cachepoints = (start, i)
+
+        return keys
+
+    def has_keys_with_prefix(self, prefix):
+        if prefix in self._data:
+            return True
+
+        if prefix.startswith(self._cachestr):
+            lo, hi = self._cachepoints
+            i = bisect_left(self._keys, prefix, lo, hi)
+        else:
+            i = bisect_left(self._keys, prefix)
+
+        if i == len(self._keys):
+            return False
+
+        return self._keys[i].startswith(prefix)

lib/html5lib/utils.py

@@ -0,0 +1,82 @@
+from __future__ import absolute_import, division, unicode_literals
+
+from types import ModuleType
+
+try:
+    import xml.etree.cElementTree as default_etree
+except ImportError:
+    import xml.etree.ElementTree as default_etree
+
+
+__all__ = ["default_etree", "MethodDispatcher", "isSurrogatePair",
+           "surrogatePairToCodepoint", "moduleFactoryFactory"]
+
+
+class MethodDispatcher(dict):
+    """Dict with 2 special properties:
+
+    On initiation, keys that are lists, sets or tuples are converted to
+    multiple keys so accessing any one of the items in the original
+    list-like object returns the matching value
+
+    md = MethodDispatcher({("foo", "bar"):"baz"})
+    md["foo"] == "baz"
+
+    A default value which can be set through the default attribute.
+    """
+
+    def __init__(self, items=()):
+        # Using _dictEntries instead of directly assigning to self is about
+        # twice as fast. Please do careful performance testing before changing
+        # anything here.
+        _dictEntries = []
+        for name, value in items:
+            if type(name) in (list, tuple, frozenset, set):
+                for item in name:
+                    _dictEntries.append((item, value))
+            else:
+                _dictEntries.append((name, value))
+        dict.__init__(self, _dictEntries)
+        self.default = None
+
+    def __getitem__(self, key):
+        return dict.get(self, key, self.default)
+
+
+# Some utility functions to dal with weirdness around UCS2 vs UCS4
+# python builds
+
+def isSurrogatePair(data):
+    return (len(data) == 2 and
+            ord(data[0]) >= 0xD800 and ord(data[0]) <= 0xDBFF and
+            ord(data[1]) >= 0xDC00 and ord(data[1]) <= 0xDFFF)
+
+
+def surrogatePairToCodepoint(data):
+    char_val = (0x10000 + (ord(data[0]) - 0xD800) * 0x400 +
+                (ord(data[1]) - 0xDC00))
+    return char_val
+
+# Module Factory Factory (no, this isn't Java, I know)
+# Here to stop this being duplicated all over the place.
+
+
+def moduleFactoryFactory(factory):
+    moduleCache = {}
+
+    def moduleFactory(baseModule, *args, **kwargs):
+        if isinstance(ModuleType.__name__, type("")):
+            name = "_%s_factory" % baseModule.__name__
+        else:
+            name = b"_%s_factory" % baseModule.__name__
+
+        if name in moduleCache:
+            return moduleCache[name]
+        else:
+            mod = ModuleType(name)
+            objs = factory(baseModule, *args, **kwargs)
+            mod.__dict__.update(objs)
+            moduleCache[name] = mod
+            return mod
+
+    return moduleFactory