def lang(self):
"""The language of this element, as a string."""
# http://whatwg.org/C#language
xml_lang = self.etree_element.get(
'{http://www.w3.org/XML/1998/namespace}lang')
if xml_lang is not None:
return ascii_lower(xml_lang)
is_html = (self.in_html_document
or self.namespace_url == 'http://www.w3.org/1999/xhtml')
if is_html:
lang = self.etree_element.get('lang')
if lang is not None:
return ascii_lower(lang)
if self.parent is not None:
return self.parent.lang
# Root elememnt
if is_html:
content_language = None
iterator = self.etree_element.iter(
'{http://www.w3.org/1999/xhtml}meta')
for meta in iterator:
http_equiv = meta.get('http-equiv', '')
if ascii_lower(http_equiv) == 'content-language':
content_language = _parse_content_language(
meta.get('content'))
if content_language is not None:
return ascii_lower(content_language)
# Empty string means unknown
return _parse_content_language(self.transport_content_language) or ''
def lang(self):
"""The language of this element, as a string."""
# http://whatwg.org/C#language
xml_lang = self.etree_element.get(
'{http://www.w3.org/XML/1998/namespace}lang')
if xml_lang is not None:
return ascii_lower(xml_lang)
is_html = (
self.in_html_document or
self.namespace_url == 'http://www.w3.org/1999/xhtml')
if is_html:
lang = self.etree_element.get('lang')
if lang is not None:
return ascii_lower(lang)
if self.parent is not None:
return self.parent.lang
# Root elememnt
if is_html:
content_language = None
for meta in etree_iter(self.etree_element,
'{http://www.w3.org/1999/xhtml}meta'):
http_equiv = meta.get('http-equiv', '')
if ascii_lower(http_equiv) == 'content-language':
content_language = _parse_content_language(
meta.get('content'))
if content_language is not None:
return ascii_lower(content_language)
# Empty string means unknown
return _parse_content_language(self.transport_content_language) or ''
def __init__(self, line, column, value):
Node.__init__(self, line, column)
self.value = value
try:
self.lower_value = ascii_lower(value)
except UnicodeEncodeError:
self.lower_value = value
def __init__(self, line, column, value):
Node.__init__(self, line, column)
self.value = value
try:
self.lower_value = ascii_lower(value)
except UnicodeEncodeError:
self.lower_value = value
def __init__(self, line, column, value, int_value, representation, unit):
Node.__init__(self, line, column)
self.value = value
self.int_value = int_value
self.is_integer = int_value is not None
self.representation = representation
self.unit = unit
self.lower_unit = ascii_lower(unit)
def __init__(self, line, column, value, int_value, representation, unit):
Node.__init__(self, line, column)
self.value = value
self.int_value = int_value
self.is_integer = int_value is not None
self.representation = representation
self.unit = unit
self.lower_unit = ascii_lower(unit)
def match(self, element):
"""Match selectors against the given element.
:param element:
An :class:`ElementWrapper`.
:returns:
A list of the payload objects associated to selectors that match
element, in order of lowest to highest
:attr:`compiler.CompiledSelector` specificity and in order of
addition with :meth:`add_selector` among selectors of equal
specificity.
"""
relevant_selectors = []
if element.id is not None and element.id in self.id_selectors:
self.add_relevant_selectors(element, self.id_selectors[element.id],
relevant_selectors)
for class_name in element.classes:
if class_name in self.class_selectors:
self.add_relevant_selectors(element,
self.class_selectors[class_name],
relevant_selectors)
lower_name = ascii_lower(element.local_name)
if lower_name in self.lower_local_name_selectors:
self.add_relevant_selectors(
element, self.lower_local_name_selectors[lower_name],
relevant_selectors)
if element.namespace_url in self.namespace_selectors:
self.add_relevant_selectors(
element, self.namespace_selectors[element.namespace_url],
relevant_selectors)
if 'lang' in element.etree_element.attrib:
self.add_relevant_selectors(element, self.lang_attr_selectors,
relevant_selectors)
self.add_relevant_selectors(element, self.other_selectors,
relevant_selectors)
relevant_selectors.sort(key=SORT_KEY)
return relevant_selectors
def match(self, element):
"""
Match selectors against the given element.
:param element:
An :class:`ElementWrapper`.
:returns:
A list of the :obj:`payload` objects associated
to selectors that match element,
in order of lowest to highest :attr:`~CompiledSelector.specificity`
and in order of addition with :meth:`add_selector`
among selectors of equal specificity.
"""
relevant_selectors = []
if element.id is not None:
relevant_selectors.append(self.id_selectors.get(element.id, []))
for class_name in element.classes:
relevant_selectors.append(self.class_selectors.get(class_name, []))
relevant_selectors.append(
self.lower_local_name_selectors.get(
ascii_lower(element.local_name), []))
relevant_selectors.append(
self.namespace_selectors.get(element.namespace_url, []))
if 'lang' in element.etree_element.attrib:
relevant_selectors.append(self.lang_attr_selectors)
relevant_selectors.append(self.other_selectors)
results = [
(specificity, order, pseudo, payload)
for selector_list in relevant_selectors
for test, specificity, order, pseudo, payload in selector_list
if test(element)
]
results.sort(key=SORT_KEY)
return results
def match(self, element):
"""
Match selectors against the given element.
:param element:
An :class:`ElementWrapper`.
:returns:
A list of the :obj:`payload` objects associated
to selectors that match element,
in order of lowest to highest :attr:`~CompiledSelector.specificity`
and in order of addition with :meth:`add_selector`
among selectors of equal specificity.
"""
relevant_selectors = []
if element.id is not None:
relevant_selectors.append(self.id_selectors.get(element.id, []))
for class_name in element.classes:
relevant_selectors.append(self.class_selectors.get(class_name, []))
relevant_selectors.append(
self.lower_local_name_selectors.get(
ascii_lower(element.local_name), []))
relevant_selectors.append(
self.namespace_selectors.get(element.namespace_url, []))
relevant_selectors.append(self.other_selectors)
results = [
(specificity, order, pseudo, payload)
for selector_list in relevant_selectors
for test, specificity, order, pseudo, payload in selector_list
if test(element)
]
results.sort(key=SORT_KEY)
return results
def __init__(self, line, column, name, arguments):
Node.__init__(self, line, column)
self.name = name
self.lower_name = ascii_lower(name)
self.arguments = arguments
def __init__(self, line, column, value):
Node.__init__(self, line, column)
self.value = value
self.lower_value = ascii_lower(value)
def parse_component_value_list(css, skip_comments=False):
"""Parse a list of component values.
:type css: :obj:`str`
:param css: A CSS string.
:type skip_comments: :obj:`bool`
:param skip_comments:
Ignore CSS comments.
The return values (and recursively its blocks and functions)
will not contain any :class:`~tinycss2.ast.Comment` object.
:returns: A list of :term:`component values`.
"""
css = (css.replace('\0', '\uFFFD')
# This turns out to be faster than a regexp:
.replace('\r\n', '\n').replace('\r', '\n').replace('\f', '\n'))
length = len(css)
token_start_pos = pos = 0 # Character index in the css source.
line = 1 # First line is line 1.
last_newline = -1
root = tokens = []
end_char = None # Pop the stack when encountering this character.
stack = [] # Stack of nested blocks: (tokens, end_char) tuples.
while pos < length:
newline = css.rfind('\n', token_start_pos, pos)
if newline != -1:
line += 1 + css.count('\n', token_start_pos, newline)
last_newline = newline
# First character in a line is in column 1.
column = pos - last_newline
token_start_pos = pos
c = css[pos]
if c in ' \n\t':
pos += 1
while css.startswith((' ', '\n', '\t'), pos):
pos += 1
value = css[token_start_pos:pos]
tokens.append(WhitespaceToken(line, column, value))
continue
elif (c in 'Uu' and pos + 2 < length and css[pos + 1] == '+' and
css[pos + 2] in '0123456789abcdefABCDEF?'):
start, end, pos = _consume_unicode_range(css, pos + 2)
tokens.append(UnicodeRangeToken(line, column, start, end))
continue
elif css.startswith('-->', pos): # Check before identifiers
tokens.append(LiteralToken(line, column, '-->'))
pos += 3
continue
elif _is_ident_start(css, pos):
value, pos = _consume_ident(css, pos)
if not css.startswith('(', pos): # Not a function
tokens.append(IdentToken(line, column, value))
continue
pos += 1 # Skip the '('
if ascii_lower(value) == 'url':
url_pos = pos
while css.startswith((' ', '\n', '\t'), url_pos):
url_pos += 1
if url_pos >= length or css[url_pos] not in ('"', "'"):
value, pos, error = _consume_url(css, pos)
if value is not None:
repr = 'url({})'.format(serialize_url(value))
if error is not None:
error_key = error[0]
if error_key == 'eof-in-string':
repr = repr[:-2]
else:
assert error_key == 'eof-in-url'
repr = repr[:-1]
tokens.append(URLToken(line, column, value, repr))
if error is not None:
tokens.append(ParseError(line, column, *error))
continue
arguments = []
tokens.append(FunctionBlock(line, column, value, arguments))
stack.append((tokens, end_char))
end_char = ')'
tokens = arguments
continue
match = _NUMBER_RE.match(css, pos)
if match:
pos = match.end()
repr_ = css[token_start_pos:pos]
value = float(repr_)
int_value = int(repr_) if not any(match.groups()) else None
if pos < length and _is_ident_start(css, pos):
unit, pos = _consume_ident(css, pos)
tokens.append(DimensionToken(
line, column, value, int_value, repr_, unit))
elif css.startswith('%', pos):
pos += 1
tokens.append(PercentageToken(
line, column, value, int_value, repr_))
else:
tokens.append(NumberToken(
line, column, value, int_value, repr_))
elif c == '@':
pos += 1
if pos < length and _is_ident_start(css, pos):
value, pos = _consume_ident(css, pos)
tokens.append(AtKeywordToken(line, column, value))
else:
tokens.append(LiteralToken(line, column, '@'))
elif c == '#':
pos += 1
if pos < length and (
css[pos] in '0123456789abcdefghijklmnopqrstuvwxyz'
'-_ABCDEFGHIJKLMNOPQRSTUVWXYZ' or
ord(css[pos]) > 0x7F or # Non-ASCII
# Valid escape:
(css[pos] == '\\' and not css.startswith('\\\n', pos))):
is_identifier = _is_ident_start(css, pos)
value, pos = _consume_ident(css, pos)
tokens.append(HashToken(line, column, value, is_identifier))
else:
tokens.append(LiteralToken(line, column, '#'))
elif c == '{':
content = []
tokens.append(CurlyBracketsBlock(line, column, content))
stack.append((tokens, end_char))
end_char = '}'
tokens = content
pos += 1
elif c == '[':
content = []
tokens.append(SquareBracketsBlock(line, column, content))
stack.append((tokens, end_char))
end_char = ']'
tokens = content
pos += 1
elif c == '(':
content = []
tokens.append(ParenthesesBlock(line, column, content))
stack.append((tokens, end_char))
end_char = ')'
tokens = content
pos += 1
elif c == end_char: # Matching }, ] or )
# The top-level end_char is None (never equal to a character),
# so we never get here if the stack is empty.
tokens, end_char = stack.pop()
pos += 1
elif c in '}])':
tokens.append(ParseError(line, column, c, 'Unmatched ' + c))
pos += 1
elif c in ('"', "'"):
value, pos, error = _consume_quoted_string(css, pos)
if value is not None:
repr = '"{}"'.format(serialize_string_value(value))
if error is not None:
repr = repr[:-1]
tokens.append(StringToken(line, column, value, repr))
if error is not None:
tokens.append(ParseError(line, column, *error))
elif css.startswith('/*', pos): # Comment
pos = css.find('*/', pos + 2)
if pos == -1:
if not skip_comments:
tokens.append(
Comment(line, column, css[token_start_pos + 2:]))
break
if not skip_comments:
tokens.append(
Comment(line, column, css[token_start_pos + 2:pos]))
pos += 2
elif css.startswith('<!--', pos):
tokens.append(LiteralToken(line, column, '<!--'))
pos += 4
elif css.startswith('||', pos):
tokens.append(LiteralToken(line, column, '||'))
pos += 2
elif c in '~|^$*':
pos += 1
if css.startswith('=', pos):
pos += 1
tokens.append(LiteralToken(line, column, c + '='))
else:
tokens.append(LiteralToken(line, column, c))
else:
tokens.append(LiteralToken(line, column, c))
pos += 1
return root
def parse_component_value_list(css, skip_comments=False):
"""Parse a list of component values.
:param css: A :term:`string`.
:param skip_comments:
Ignore CSS comments.
The return values (and recursively its blocks and functions)
will not contain any :class:`~tinycss2.ast.Comment` object.
:returns: A list of :term:`component values`.
"""
css = (css.replace('\0', '\uFFFD')
# This turns out to be faster than a regexp:
.replace('\r\n', '\n').replace('\r', '\n').replace('\f', '\n'))
length = len(css)
token_start_pos = pos = 0 # Character index in the css source.
line = 1 # First line is line 1.
last_newline = -1
root = tokens = []
end_char = None # Pop the stack when encountering this character.
stack = [] # Stack of nested blocks: (tokens, end_char) tuples.
while pos < length:
newline = css.rfind('\n', token_start_pos, pos)
if newline != -1:
line += 1 + css.count('\n', token_start_pos, newline)
last_newline = newline
# First character in a line is in column 1.
column = pos - last_newline
token_start_pos = pos
c = css[pos]
if c in ' \n\t':
pos += 1
while css.startswith((' ', '\n', '\t'), pos):
pos += 1
value = css[token_start_pos:pos]
tokens.append(WhitespaceToken(line, column, value))
continue
elif (c in 'Uu' and pos + 2 < length and css[pos + 1] == '+' and
css[pos + 2] in '0123456789abcdefABCDEF?'):
start, end, pos = _consume_unicode_range(css, pos + 2)
tokens.append(UnicodeRangeToken(line, column, start, end))
continue
elif css.startswith('-->', pos): # Check before identifiers
tokens.append(LiteralToken(line, column, '-->'))
pos += 3
continue
elif _is_ident_start(css, pos):
value, pos = _consume_ident(css, pos)
if not css.startswith('(', pos): # Not a function
tokens.append(IdentToken(line, column, value))
continue
pos += 1 # Skip the '('
if ascii_lower(value) == 'url':
value, pos = _consume_url(css, pos)
tokens.append(
URLToken(line, column, value) if value is not None
else ParseError(line, column, 'bad-url', 'bad URL token'))
continue
arguments = []
tokens.append(FunctionBlock(line, column, value, arguments))
stack.append((tokens, end_char))
end_char = ')'
tokens = arguments
continue
match = _NUMBER_RE.match(css, pos)
if match:
pos = match.end()
repr_ = css[token_start_pos:pos]
value = float(repr_)
int_value = int(repr_) if not any(match.groups()) else None
if pos < length and _is_ident_start(css, pos):
unit, pos = _consume_ident(css, pos)
tokens.append(DimensionToken(
line, column, value, int_value, repr_, unit))
elif css.startswith('%', pos):
pos += 1
tokens.append(PercentageToken(
line, column, value, int_value, repr_))
else:
tokens.append(NumberToken(
line, column, value, int_value, repr_))
elif c == '@':
pos += 1
if pos < length and _is_ident_start(css, pos):
value, pos = _consume_ident(css, pos)
tokens.append(AtKeywordToken(line, column, value))
else:
tokens.append(LiteralToken(line, column, '@'))
elif c == '#':
pos += 1
if pos < length and (
css[pos] in '0123456789abcdefghijklmnopqrstuvwxyz'
'-_ABCDEFGHIJKLMNOPQRSTUVWXYZ' or
ord(css[pos]) > 0x7F or # Non-ASCII
# Valid escape:
(css[pos] == '\\' and not css.startswith('\\\n', pos))):
is_identifier = _is_ident_start(css, pos)
value, pos = _consume_ident(css, pos)
tokens.append(HashToken(line, column, value, is_identifier))
else:
tokens.append(LiteralToken(line, column, '#'))
elif c == '{':
content = []
tokens.append(CurlyBracketsBlock(line, column, content))
stack.append((tokens, end_char))
end_char = '}'
tokens = content
pos += 1
elif c == '[':
content = []
tokens.append(SquareBracketsBlock(line, column, content))
stack.append((tokens, end_char))
end_char = ']'
tokens = content
pos += 1
elif c == '(':
content = []
tokens.append(ParenthesesBlock(line, column, content))
stack.append((tokens, end_char))
end_char = ')'
tokens = content
pos += 1
elif c == end_char: # Matching }, ] or )
# The top-level end_char is None (never equal to a character),
# so we never get here if the stack is empty.
tokens, end_char = stack.pop()
pos += 1
elif c in '}])':
tokens.append(ParseError(line, column, c, 'Unmatched ' + c))
pos += 1
elif c in ('"', "'"):
value, pos = _consume_quoted_string(css, pos)
tokens.append(
StringToken(line, column, value) if value is not None
else ParseError(line, column, 'bad-string',
'bad string token'))
elif css.startswith('/*', pos): # Comment
pos = css.find('*/', pos + 2)
if pos == -1:
if not skip_comments:
tokens.append(
Comment(line, column, css[token_start_pos + 2:]))
break
if not skip_comments:
tokens.append(
Comment(line, column, css[token_start_pos + 2:pos]))
pos += 2
elif css.startswith('<!--', pos):
tokens.append(LiteralToken(line, column, '<!--'))
pos += 4
elif css.startswith('||', pos):
tokens.append(LiteralToken(line, column, '||'))
pos += 2
elif c in '~|^$*':
pos += 1
if css.startswith('=', pos):
pos += 1
tokens.append(LiteralToken(line, column, c + '='))
else:
tokens.append(LiteralToken(line, column, c))
else:
tokens.append(LiteralToken(line, column, c))
pos += 1
return root
def __init__(self, line, column, name, arguments):
Node.__init__(self, line, column)
self.name = name
self.lower_name = ascii_lower(name)
self.arguments = arguments
def __init__(self, line, column, value):
Node.__init__(self, line, column)
self.value = value
self.lower_value = ascii_lower(value)
def _compile_node(selector):
"""Return a boolean expression, as a Python source string.
When evaluated in a context where the `el` variable is an
:class:`cssselect2.tree.Element` object,
tells whether the element is a subject of `selector`.
"""
# To avoid precedence-related bugs, any sub-expression that is passed
# around must be "atomic": add parentheses when the top-level would be
# an operator. Bare literals and function calls are fine.
# 1 and 0 are used for True and False to avoid global lookups.
if isinstance(selector, parser.CombinedSelector):
left_inside = _compile_node(selector.left)
if left_inside == '0':
return '0' # 0 and x == 0
elif left_inside == '1':
# 1 and x == x, but the element matching 1 still needs to exist.
if selector.combinator in (' ', '>'):
left = 'el.parent is not None'
elif selector.combinator in ('~', '+'):
left = 'el.previous is not None'
else:
raise SelectorError('Unknown combinator', selector.combinator)
# Rebind the `el` name inside a generator-expressions (in a new scope)
# so that 'left_inside' applies to different elements.
elif selector.combinator == ' ':
left = 'any((%s) for el in el.ancestors)' % left_inside
elif selector.combinator == '>':
left = ('next(el is not None and (%s) for el in [el.parent])' %
left_inside)
elif selector.combinator == '+':
left = ('next(el is not None and (%s) for el in [el.previous])' %
left_inside)
elif selector.combinator == '~':
left = 'any((%s) for el in el.previous_siblings)' % left_inside
else:
raise SelectorError('Unknown combinator', selector.combinator)
right = _compile_node(selector.right)
if right == '0':
return '0' # 0 and x == 0
elif right == '1':
return left # 1 and x == x
else:
# Evaluate combinators right to left:
return '(%s) and (%s)' % (right, left)
elif isinstance(selector, parser.CompoundSelector):
sub_expressions = [
expr for expr in map(_compile_node, selector.simple_selectors)
if expr != '1'
]
if len(sub_expressions) == 1:
test = sub_expressions[0]
elif '0' in sub_expressions:
test = '0'
elif sub_expressions:
test = ' and '.join('(%s)' % e for e in sub_expressions)
else:
test = '1' # all([]) == True
return test
elif isinstance(selector, parser.NegationSelector):
sub_expressions = [
expr for expr in map(_compile_node, selector.selector_list)
if expr != '1'
]
if not sub_expressions:
return '0'
return f'not ({" or ".join(f"({expr})" for expr in sub_expressions)})'
elif isinstance(
selector,
(parser.MatchesAnySelector, parser.SpecificityAdjustmentSelector)):
sub_expressions = [
expr for expr in map(_compile_node, selector.selector_list)
if expr != '0'
]
if not sub_expressions:
return '0'
return ' or '.join(f'({expr})' for expr in sub_expressions)
elif isinstance(selector, parser.LocalNameSelector):
if selector.lower_local_name == selector.local_name:
return 'el.local_name == %r' % selector.local_name
else:
return ('el.local_name == (%r if el.in_html_document else %r)' %
(selector.lower_local_name, selector.local_name))
elif isinstance(selector, parser.NamespaceSelector):
return 'el.namespace_url == %r' % selector.namespace
elif isinstance(selector, parser.ClassSelector):
return '%r in el.classes' % selector.class_name
elif isinstance(selector, parser.IDSelector):
return 'el.id == %r' % selector.ident
elif isinstance(selector, parser.AttributeSelector):
if selector.namespace is not None:
if selector.namespace:
if selector.name == selector.lower_name:
key = repr('{%s}%s' % (selector.namespace, selector.name))
else:
key = '(%r if el.in_html_document else %r)' % (
'{%s}%s' % (selector.namespace, selector.lower_name),
'{%s}%s' % (selector.namespace, selector.name),
)
else:
if selector.name == selector.lower_name:
key = repr(selector.name)
else:
key = '(%r if el.in_html_document else %r)' % (
selector.lower_name, selector.name)
value = selector.value
if selector.operator is None:
return '%s in el.etree_element.attrib' % key
elif selector.operator == '=':
return 'el.etree_element.get(%s) == %r' % (key, value)
elif selector.operator == '~=':
if len(value.split()) != 1 or value.strip() != value:
return '0'
else:
return (
'%r in split_whitespace(el.etree_element.get(%s, ""))'
% (value, key))
elif selector.operator == '|=':
return ('next(v == %r or (v is not None and v.startswith(%r))'
' for v in [el.etree_element.get(%s)])' %
(value, value + '-', key))
elif selector.operator == '^=':
if value:
return 'el.etree_element.get(%s, "").startswith(%r)' % (
key, value)
else:
return '0'
elif selector.operator == '$=':
if value:
return 'el.etree_element.get(%s, "").endswith(%r)' % (
key, value)
else:
return '0'
elif selector.operator == '*=':
if value:
return '%r in el.etree_element.get(%s, "")' % (value, key)
else:
return '0'
else:
raise SelectorError('Unknown attribute operator',
selector.operator)
else: # In any namespace
raise NotImplementedError # TODO
elif isinstance(selector, parser.PseudoClassSelector):
if selector.name in ('link', 'any-link', 'local-link'):
test = '%s and el.etree_element.get("href") is not None '
if selector.name == 'local-link':
test += 'and not urlparse(el.etree_element.get("href")).scheme'
return test % html_tag_eq('a', 'area', 'link')
elif selector.name == 'enabled':
return ('(%s and el.etree_element.get("disabled") is None'
' and not el.in_disabled_fieldset) or'
'(%s and el.etree_element.get("disabled") is None) or '
'(%s and el.etree_element.get("href") is not None)' % (
html_tag_eq('button', 'input', 'select', 'textarea',
'option'),
html_tag_eq('optgroup', 'menuitem', 'fieldset'),
html_tag_eq('a', 'area', 'link'),
))
elif selector.name == 'disabled':
return ('(%s and (el.etree_element.get("disabled") is not None'
' or el.in_disabled_fieldset)) or'
'(%s and el.etree_element.get("disabled") is not None)' % (
html_tag_eq('button', 'input', 'select', 'textarea',
'option'),
html_tag_eq('optgroup', 'menuitem', 'fieldset'),
))
elif selector.name == 'checked':
return (
'(%s and el.etree_element.get("checked") is not None and'
' ascii_lower(el.etree_element.get("type", "")) '
' in ("checkbox", "radio"))'
'or (%s and el.etree_element.get("selected") is not None)' % (
html_tag_eq('input', 'menuitem'),
html_tag_eq('option'),
))
elif selector.name in ('visited', 'hover', 'active', 'focus',
'focus-within', 'focus-visible', 'target',
'target-within', 'current', 'past', 'future',
'playing', 'paused', 'seeking', 'buffering',
'stalled', 'muted', 'volume-locked',
'user-valid', 'user-invalid'):
# Not applicable in a static context: never match.
return '0'
elif selector.name in ('root', 'scope'):
return 'el.parent is None'
elif selector.name == 'first-child':
return 'el.index == 0'
elif selector.name == 'last-child':
return 'el.index + 1 == len(el.etree_siblings)'
elif selector.name == 'first-of-type':
return ('all(s.tag != el.etree_element.tag'
' for s in el.etree_siblings[:el.index])')
elif selector.name == 'last-of-type':
return ('all(s.tag != el.etree_element.tag'
' for s in el.etree_siblings[el.index + 1:])')
elif selector.name == 'only-child':
return 'len(el.etree_siblings) == 1'
elif selector.name == 'only-of-type':
return ('all(s.tag != el.etree_element.tag or i == el.index'
' for i, s in enumerate(el.etree_siblings))')
elif selector.name == 'empty':
return 'not (el.etree_children or el.etree_element.text)'
else:
raise SelectorError('Unknown pseudo-class', selector.name)
elif isinstance(selector, parser.FunctionalPseudoClassSelector):
if selector.name == 'lang':
langs = []
tokens = [
token for token in selector.arguments
if token.type not in ('whitespace', 'comment')
]
while tokens:
token = tokens.pop(0)
if token.type == 'ident':
langs.append(token.lower_value)
elif token.type == 'string':
langs.append(ascii_lower(token.value))
else:
raise SelectorError('Invalid arguments for :lang()')
if tokens:
token = tokens.pop(0)
if token.type != 'ident' and token.value != ',':
raise SelectorError('Invalid arguments for :lang()')
return ' or '.join(
f'el.lang == {lang!r} or el.lang.startswith({lang + "-"!r})'
for lang in langs)
else:
nth = []
selector_list = []
current_list = nth
for argument in selector.arguments:
if argument.type == 'ident' and argument.value == 'of':
if current_list is nth:
current_list = selector_list
continue
current_list.append(argument)
if selector_list:
test = ' and '.join(
_compile_node(selector.parsed_tree)
for selector in parser.parse(selector_list))
if selector.name == 'nth-child':
count = (
f'sum(1 for el in el.previous_siblings if ({test}))')
elif selector.name == 'nth-last-child':
count = ('sum(1 for el in'
' tuple(el.iter_siblings())[el.index + 1:]'
f' if ({test}))')
elif selector.name == 'nth-of-type':
count = (
'sum(1 for s in ('
' el for el in el.previous_siblings'
f' if ({test}))'
' if s.etree_element.tag == el.etree_element.tag)')
elif selector.name == 'nth-last-of-type':
count = (
'sum(1 for s in ('
' el for el in'
' tuple(el.iter_siblings())[el.index + 1:]'
f' if ({test}))'
' if s.etree_element.tag == el.etree_element.tag)')
else:
raise SelectorError('Unknown pseudo-class', selector.name)
count += f'if ({test}) else float("nan")'
else:
if current_list is selector_list:
raise SelectorError('Invalid arguments for :%s()' %
selector.name)
if selector.name == 'nth-child':
count = 'el.index'
elif selector.name == 'nth-last-child':
count = 'len(el.etree_siblings) - el.index - 1'
elif selector.name == 'nth-of-type':
count = ('sum(1 for s in el.etree_siblings[:el.index]'
' if s.tag == el.etree_element.tag)')
elif selector.name == 'nth-last-of-type':
count = ('sum(1 for s in el.etree_siblings[el.index + 1:]'
' if s.tag == el.etree_element.tag)')
else:
raise SelectorError('Unknown pseudo-class', selector.name)
result = parse_nth(nth)
if result is None:
raise SelectorError('Invalid arguments for :%s()' %
selector.name)
a, b = result
# x is the number of siblings before/after the element
# Matches if a positive or zero integer n exists so that:
# x = a*n + b-1
# x = a*n + B
B = b - 1
if a == 0:
# x = B
return '(%s) == %i' % (count, B)
else:
# n = (x - B) / a
return ('next(r == 0 and n >= 0'
' for n, r in [divmod((%s) - %i, %i)])' %
(count, B, a))
else:
raise TypeError(type(selector), selector)
