scraper 0.14.0

HTML parsing and querying with CSS selectors
Documentation
//! CSS selectors.

use std::convert::TryFrom;
use std::fmt;

use smallvec::SmallVec;

use html5ever::{LocalName, Namespace};
use selectors::parser::SelectorParseErrorKind;
use selectors::{matching, parser, visitor};

use crate::error::SelectorErrorKind;
use crate::ElementRef;

/// Wrapper around CSS selectors.
///
/// Represents a "selector group", i.e. a comma-separated list of selectors.
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct Selector {
    /// The CSS selectors.
    selectors: SmallVec<[parser::Selector<Simple>; 1]>,
}

impl Selector {
    /// Parses a CSS selector group.

    pub fn parse(selectors: &'_ str) -> Result<Self, SelectorErrorKind> {
        let mut parser_input = cssparser::ParserInput::new(selectors);
        let mut parser = cssparser::Parser::new(&mut parser_input);

        parser::SelectorList::parse(&Parser, &mut parser)
            .map(|list| Selector { selectors: list.0 })
            .map_err(SelectorErrorKind::from)
    }

    /// Returns true if the element matches this selector.
    pub fn matches(&self, element: &ElementRef) -> bool {
        self.matches_with_scope(element, None)
    }

    /// Returns true if the element matches this selector.
    /// The optional `scope` argument is used to specify which element has `:scope` pseudo-class.
    /// When it is `None`, `:scope` will match the root element.
    pub fn matches_with_scope(&self, element: &ElementRef, scope: Option<ElementRef>) -> bool {
        let mut context = matching::MatchingContext::new(
            matching::MatchingMode::Normal,
            None,
            None,
            matching::QuirksMode::NoQuirks,
        );
        context.scope_element = scope.map(|x| selectors::Element::opaque(&x));
        self.selectors
            .iter()
            .any(|s| matching::matches_selector(s, 0, None, element, &mut context, &mut |_, _| {}))
    }
}

/// An implementation of `Parser` for `selectors`
struct Parser;
impl<'i> parser::Parser<'i> for Parser {
    type Impl = Simple;
    type Error = SelectorParseErrorKind<'i>;
}

/// A simple implementation of `SelectorImpl` with no pseudo-classes or pseudo-elements.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub struct Simple;

impl parser::SelectorImpl for Simple {
    type AttrValue = String;
    type Identifier = LocalName;
    type ClassName = LocalName;
    type PartName = LocalName;
    type LocalName = LocalName;
    type NamespacePrefix = LocalName;
    type NamespaceUrl = Namespace;
    type BorrowedNamespaceUrl = Namespace;
    type BorrowedLocalName = LocalName;

    type NonTSPseudoClass = NonTSPseudoClass;
    type PseudoElement = PseudoElement;

    // see: https://github.com/servo/servo/pull/19747#issuecomment-357106065
    type ExtraMatchingData = String;
}

/// Non Tree-Structural Pseudo-Class.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum NonTSPseudoClass {}

impl parser::NonTSPseudoClass for NonTSPseudoClass {
    type Impl = Simple;

    fn is_active_or_hover(&self) -> bool {
        false
    }

    fn is_user_action_state(&self) -> bool {
        false
    }

    fn has_zero_specificity(&self) -> bool {
        false
    }
}

impl parser::Visit for NonTSPseudoClass {
    type Impl = Simple;

    fn visit<V>(&self, _visitor: &mut V) -> bool
    where
        V: visitor::SelectorVisitor<Impl = Self::Impl>,
    {
        true
    }
}

impl cssparser::ToCss for NonTSPseudoClass {
    fn to_css<W>(&self, dest: &mut W) -> fmt::Result
    where
        W: fmt::Write,
    {
        dest.write_str("")
    }
}

/// CSS Pseudo-Element
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum PseudoElement {}

impl parser::PseudoElement for PseudoElement {
    type Impl = Simple;
}

impl cssparser::ToCss for PseudoElement {
    fn to_css<W>(&self, dest: &mut W) -> fmt::Result
    where
        W: fmt::Write,
    {
        dest.write_str("")
    }
}

impl<'i> TryFrom<&'i str> for Selector {
    type Error = SelectorErrorKind<'i>;

    fn try_from(s: &'i str) -> Result<Self, Self::Error> {
        Selector::parse(s)
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::convert::TryInto;

    #[test]
    fn selector_conversions() {
        let s = "#testid.testclass";
        let _sel: Selector = s.try_into().unwrap();

        let s = s.to_owned();
        let _sel: Selector = (*s).try_into().unwrap();
    }

    #[test]
    #[should_panic]
    fn invalid_selector_conversions() {
        let s = "<failing selector>";
        let _sel: Selector = s.try_into().unwrap();
    }
}