def loadbytes(self, data):
# defer cleaning the tree to prevent custom cleaners from cleaning
# WebAnnotator markup
tree = html_document_fromstring(data, encoding=self.encoding)
webannotator.apply_wa_title(tree)
if self.known_entities:
self._prune_tags(tree)
entities = self._get_entities(tree)
self._process_entities(entities)
return self._cleanup_tree(tree)
def loadbytes(self, data):
# defer cleaning the tree to prevent custom cleaners from cleaning
# WebAnnotator markup
tree = html_document_fromstring(data, encoding=self.encoding)
webannotator.apply_wa_title(tree)
if self.known_entities:
self._prune_tags(tree)
entities = self._get_entities(tree)
self._process_entities(entities)
return self._cleanup_tree(tree)
def test_wa_title(self):
tree = html_document_fromstring(b"""
<html>
<head><title>Foo</title></head>
<body>contents</body>
<wa-title><b>hello</b>, world</wa-title>
</html>
""")
webannotator.apply_wa_title(tree)
self.assertHtmlTreeEqual(tree, html_document_fromstring(b"""
<html>
<head><title><b>hello</b>, world</title></head>
<body>contents</body>
</html>
"""))
def assertApplyWaTitle(self, source, result):
tree = html_document_fromstring(source)
webannotator.apply_wa_title(tree)
self.assertHtmlTreeEqual(tree, html_document_fromstring(result))
def assertApplyWaTitle(self, source, result):
tree = html_document_fromstring(source)
webannotator.apply_wa_title(tree)
self.assertHtmlTreeEqual(tree, html_document_fromstring(result))
