Struct Html 

pub struct Html<'s> { /* private fields */ }

解析好的文档。借用源字符串（输入含 \r 或 NUL 时持有规范化后的副本）。

Implementations

impl<'s> Html<'s>

pub fn parse_with(html: &'s str, opts: &ParseOptions<'_>) -> Html<'s>

按选项解析：关闭脚本、在给定上下文元素里解析片段。

pub fn parse(html: &'s str) -> Html<'s>

解析整个 HTML 文档。永不失败：任何输入都会得到一棵树。

Examples found in repository

examples/parse_bench.rs (line 30)

fn main() {
    let mut args = std::env::args().skip(1);
    let manifest = args
        .next()
        .expect("usage: parse_bench <manifest.jsonl> [rounds]");
    let rounds: usize = args.next().and_then(|s| s.parse().ok()).unwrap_or(5);
    let pages: Vec<String> = std::fs::read_to_string(&manifest)
        .unwrap()
        .lines()
        .filter_map(|l| {
            let i = l.find("\"path\"")?;
            let rest = &l[i + 6..];
            let s = rest.find('"')? + 1;
            let e = s + rest[s..].find('"')?;
            std::fs::read(&rest[s..e]).ok()
        })
        .map(|b| String::from_utf8_lossy(&b).into_owned())
        .collect();
    let mut best = f64::MAX;
    let mut nodes = 0;
    for _ in 0..rounds {
        let t = Instant::now();
        nodes = 0;
        for p in &pages {
            nodes += std::hint::black_box(pagesift_html::Html::parse(p)).len();
        }
        let ms = t.elapsed().as_secs_f64() * 1e3;
        best = best.min(ms);
        print!("{ms:.1} ");
    }
    println!("| pages={} nodes={nodes} best_ms={best:.1}", pages.len());
}

pub fn parse_with_scratch(html: &'s str, scratch: &mut ParseScratch) -> Html<'s>

Parse while reusing node, attribute, and decoded-text allocations from the preceding document handled by this worker.

pub fn root(&self) -> NodeRef<'_>

文档根节点（不是 <html> 元素，是它的父节点）。

pub fn len(&self) -> usize

节点总数。节点 id 的范围是 0..len()，可以用来开按节点下标索引的数组。

Examples found in repository

examples/parse_bench.rs (line 30)

fn main() {
    let mut args = std::env::args().skip(1);
    let manifest = args
        .next()
        .expect("usage: parse_bench <manifest.jsonl> [rounds]");
    let rounds: usize = args.next().and_then(|s| s.parse().ok()).unwrap_or(5);
    let pages: Vec<String> = std::fs::read_to_string(&manifest)
        .unwrap()
        .lines()
        .filter_map(|l| {
            let i = l.find("\"path\"")?;
            let rest = &l[i + 6..];
            let s = rest.find('"')? + 1;
            let e = s + rest[s..].find('"')?;
            std::fs::read(&rest[s..e]).ok()
        })
        .map(|b| String::from_utf8_lossy(&b).into_owned())
        .collect();
    let mut best = f64::MAX;
    let mut nodes = 0;
    for _ in 0..rounds {
        let t = Instant::now();
        nodes = 0;
        for p in &pages {
            nodes += std::hint::black_box(pagesift_html::Html::parse(p)).len();
        }
        let ms = t.elapsed().as_secs_f64() * 1e3;
        best = best.min(ms);
        print!("{ms:.1} ");
    }
    println!("| pages={} nodes={nodes} best_ms={best:.1}", pages.len());
}

pub fn is_empty(&self) -> bool

pub fn get(&self, id: NodeId) -> NodeRef<'_>

pub fn body(&self) -> NodeRef<'_>

第一个 <body>，没有时返回根节点。

impl<'a> Html<'a>

pub fn select<'s>(&'a self, sel: &'s Selector) -> Select<'a, 's>

文档里匹配选择器的所有元素，按文档顺序。

Trait Implementations

impl Debug for Html<'_>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.