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"
from html5lib.constants import entities
encode_entity_map = {}
for k, v in list(entities.items()):
if v != "&" and encode_entity_map.get(v) != k.lower():
# prefer < over < and similarly for &, >, etc.
encode_entity_map[v] = k
def htmlentityreplace_errors(exc):
if isinstance(exc, (UnicodeEncodeError, UnicodeTranslateError)):
res = []
for c in exc.object[exc.start:exc.end]:
e = encode_entity_map.get(c)
if e:
res.append("&")
res.append(e)
if not e.endswith(";"):
res.append(";")
else:
from xml.sax.saxutils import escape
spaceCharacters = u"".join(spaceCharacters)
try:
from codecs import register_error, xmlcharrefreplace_errors
except ImportError:
unicode_encode_errors = "strict"
else:
unicode_encode_errors = "htmlentityreplace"
from html5lib.constants import entities
encode_entity_map = {}
for k, v in entities.items():
if v != "&" and encode_entity_map.get(v) != k.lower():
# prefer < over < and similarly for &, >, etc.
encode_entity_map[v] = k
def htmlentityreplace_errors(exc):
if isinstance(exc, (UnicodeEncodeError, UnicodeTranslateError)):
res = []
for c in exc.object[exc.start:exc.end]:
e = encode_entity_map.get(c)
if e:
res.append("&")
res.append(e)
if not e.endswith(";"):
res.append(";")
else:
from html5lib import utils
from xml.sax.saxutils import escape
spaceCharacters = u"".join(spaceCharacters)
try:
from codecs import register_error, xmlcharrefreplace_errors
except ImportError:
unicode_encode_errors = "strict"
else:
unicode_encode_errors = "htmlentityreplace"
from html5lib.constants import entities
encode_entity_map = {}
for k, v in entities.items():
if v != "&" and encode_entity_map.get(v) != k.lower():
# prefer < over < and similarly for &, >, etc.
encode_entity_map[ord(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])]):
#!/usr/bin/env python2
# vim:fileencoding=utf-8
from __future__ import (unicode_literals, division, absolute_import,
print_function)
__license__ = 'GPL v3'
__copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
from html5lib.constants import entities
html5_entities = {k.replace(';', ''): v for k, v in entities.items()}