def runTest(self):
f = open(doc_fn, 'rb')
c = f.read()
f.close()
doc = html.document_fromstring(c)
body = doc.xpath('//body')[0]
bad = []
selector, count = self.selectors[self.index]
options = dict(regex_prefix='re')
xpath = cssselect.css_to_xpath(cssselect.parse(selector, options),
**options)
try:
results = body.xpath(xpath, namespaces=namespaces)
except Exception:
e = sys.exc_info()[1]
e.args = ("%s for xpath %r" % (e, xpath), )
raise
found = {}
for item in results:
if item in found:
assert 0, ("Element shows up multiple times: %r" % item)
found[item] = None
if isinstance(results, basestring):
assert 0, ("Got string result (%r), not element, for xpath %r" %
(results[:20], str(xpath)))
if len(results) != count:
#if self.shortDescription() == 'div.character, div.dialog':
# import pdb; pdb.set_trace()
assert 0, (
"Did not get expected results (%s) instead %s for xpath %r" %
(count, len(results), str(xpath)))
def runTest(self):
f = open(doc_fn, "rb")
c = f.read()
f.close()
doc = html.document_fromstring(c)
body = doc.xpath("//body")[0]
bad = []
selector, count = self.selectors[self.index]
xpath = cssselect.css_to_xpath(cssselect.parse(selector))
try:
results = body.xpath(xpath)
except Exception:
e = sys.exc_info()[1]
e.args = "%s for xpath %r" % (e, xpath)
raise
found = {}
for item in results:
if item in found:
assert 0, "Element shows up multiple times: %r" % item
found[item] = None
if isinstance(results, basestring):
assert 0, "Got string result (%r), not element, for xpath %r" % (results[:20], str(xpath))
if len(results) != count:
# if self.shortDescription() == 'div.character, div.dialog':
# import pdb; pdb.set_trace()
assert 0, "Did not get expected results (%s) instead %s for xpath %r" % (count, len(results), str(xpath))
def match_selectors_against_html_root_element(selectors, html_element):
'''
Find the selectors that match with the DOM from the given HTML.
@param selectors set of CSS selectors (strings)
@param html_element lxml.etree.Element object
@return set of found selectors
'''
found_selectors = set()
css_to_xpath_translator = CssDeadwoodHtmlTranslator()
for selector_str in selectors:
try:
# Instead of just calling css_to_xpath(selector_str),
# we first convert the css selector string to a cssselect.Selector instance
# to pass to selector_to_xpath(), so we can properly ignore pseudo elements.
# Note that cssselect.parse() always returns a list, so we do a for loop.
for selector in cssselect.parse(selector_str):
selector.pseudo_element = None
xpath_expr = css_to_xpath_translator.selector_to_xpath(
selector)
if len(html_element.xpath(xpath_expr)) > 0:
found_selectors.add(selector_str)
except Exception:
global _log
_log.exception('lxml css select failed on selector %r' %
selector_str)
return found_selectors
def selector_to_xpath(selector):
"""Return ``pseudo_type, selector_callable`` from a cssutils ``selector``.
``pseudo_type`` is a string and ``selector_callable`` is a
:class:`lxml.cssselect` XPath callable.
"""
try:
return selector._x_weasyprint_parsed_cssselect
except AttributeError:
parsed_selector = cssselect.parse(selector.selectorText)
# cssutils made sure that `selector` is not a "group of selectors"
# in CSS3 terms (`rule.selectorList` is) so `parsed_selector` cannot be
# of type `cssselect.Or`.
# This leaves only three cases:
# - The selector ends with a pseudo-element. As `cssselect.parse()`
# parses left-to-right, `parsed_selector` is a `cssselect.Pseudo`
# instance that we can unwrap. This is the only place where CSS
# allows pseudo-element selectors.
# - The selector has a pseudo-element not at the end. This is invalid
# and the whole ruleset should be ignored.
# cssselect.CSSSelector() will raise a cssselect.ExpressionError.
# - The selector has no pseudo-element and is supported by
# `cssselect.CSSSelector`.
if isinstance(parsed_selector, cssselect.CombinedSelector):
simple_selector = parsed_selector.subselector
if isinstance(simple_selector, cssselect.Pseudo) \
and simple_selector.ident in PSEUDO_ELEMENTS:
pseudo_type = str(simple_selector.ident)
# Remove the pseudo-element from the selector
parsed_selector.subselector = simple_selector.element
else:
# No pseudo-element or invalid selector.
pseudo_type = None
else:
if isinstance(parsed_selector, cssselect.Pseudo) \
and parsed_selector.ident in PSEUDO_ELEMENTS:
pseudo_type = str(parsed_selector.ident)
# Remove the pseudo-element from the selector
parsed_selector = parsed_selector.element
else:
# No pseudo-element or invalid selector.
pseudo_type = None
selector_callable = cssselect.CSSSelector(parsed_selector)
result = (pseudo_type, selector_callable)
# Cache for next time we use the same stylesheet
selector._x_weasyprint_parsed_cssselect = result
return result