kumo 0.3.13

use bytes::Bytes;
use kumo::extract::{LinkExtractor, Response};

fn make_response(url: &str, html: &str) -> Response {
    Response::from_parts(url, 200, html)
}

#[test]
fn extracts_all_links_by_default() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/a">A</a><a href="/b">B</a>"#,
    );
    let links = LinkExtractor::new().extract(&res);
    assert_eq!(
        links,
        vec!["https://example.com/a", "https://example.com/b"]
    );
}

#[test]
fn resolves_relative_urls() {
    let res = make_response("https://example.com/page/1", r#"<a href="../2">next</a>"#);
    let links = LinkExtractor::new().extract(&res);
    assert_eq!(links, vec!["https://example.com/2"]);
}

#[test]
fn allow_filter_keeps_matching_only() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/product/1">p</a><a href="/about">a</a>"#,
    );
    let links = LinkExtractor::new().allow(r"/product/").extract(&res);
    assert_eq!(links, vec!["https://example.com/product/1"]);
}

#[test]
fn deny_filter_removes_matching() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/page">p</a><a href="/page.pdf">pdf</a>"#,
    );
    let links = LinkExtractor::new().deny(r"\.pdf$").extract(&res);
    assert_eq!(links, vec!["https://example.com/page"]);
}

#[test]
fn deduplicates_links() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/a">1</a><a href="/a">2</a><a href="/b">3</a>"#,
    );
    let links = LinkExtractor::new().extract(&res);
    assert_eq!(links.len(), 2);
    assert_eq!(links[0], "https://example.com/a");
    assert_eq!(links[1], "https://example.com/b");
}

#[test]
fn restrict_css_scopes_search() {
    let res = make_response(
        "https://example.com/",
        r#"<nav><a href="/nav">nav</a></nav><footer><a href="/foot">foot</a></footer>"#,
    );
    let links = LinkExtractor::new().restrict_css("nav").extract(&res);
    assert_eq!(links, vec!["https://example.com/nav"]);
}

#[test]
fn returns_empty_for_binary_response() {
    let res = Response::from_bytes("https://example.com", 200, Bytes::from_static(b"\xff\xfe"));
    let links = LinkExtractor::new().extract(&res);
    assert!(links.is_empty());
}

#[test]
fn allow_and_deny_combine() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/product/1">p1</a>
           <a href="/product/2.pdf">pdf</a>
           <a href="/about">about</a>"#,
    );
    let links = LinkExtractor::new()
        .allow(r"/product/")
        .deny(r"\.pdf$")
        .extract(&res);
    assert_eq!(links, vec!["https://example.com/product/1"]);
}

#[test]
fn canonicalize_strips_fragments() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/page#s1">s1</a><a href="/page#s2">s2</a><a href="/page">p</a>"#,
    );
    let links = LinkExtractor::new().canonicalize(true).extract(&res);
    assert_eq!(links, vec!["https://example.com/page"]);
}

#[test]
fn no_canonicalize_keeps_fragments_distinct() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/page#s1">s1</a><a href="/page#s2">s2</a>"#,
    );
    let links = LinkExtractor::new().canonicalize(false).extract(&res);
    assert_eq!(links.len(), 2);
}

#[test]
fn allow_domains_keeps_matching_domain() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="https://example.com/a">a</a>
           <a href="https://other.com/b">b</a>
           <a href="https://sub.example.com/c">c</a>"#,
    );
    let links = LinkExtractor::new()
        .allow_domains(&["example.com"])
        .extract(&res);
    assert_eq!(
        links,
        vec!["https://example.com/a", "https://sub.example.com/c"]
    );
}

#[test]
fn deny_domains_removes_matching_domain() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="https://example.com/a">a</a>
           <a href="https://ads.com/b">b</a>"#,
    );
    let links = LinkExtractor::new()
        .deny_domains(&["ads.com"])
        .extract(&res);
    assert_eq!(links, vec!["https://example.com/a"]);
}

#[test]
fn allow_domains_and_allow_regex_are_or_ed() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="https://example.com/page">page</a>
           <a href="https://cdn.other.com/img.png">img</a>
           <a href="https://third.com/x">x</a>"#,
    );
    let links = LinkExtractor::new()
        .allow_domains(&["example.com"])
        .allow(r"cdn\.other\.com")
        .extract(&res);
    assert_eq!(
        links,
        vec!["https://example.com/page", "https://cdn.other.com/img.png"]
    );
}

#[test]
fn extracts_from_area_tags_by_default() {
    let res = make_response(
        "https://example.com/",
        r#"<map><area href="/map-link"></map><a href="/a-link">a</a>"#,
    );
    let links = LinkExtractor::new().extract(&res);
    assert!(links.contains(&"https://example.com/map-link".to_string()));
    assert!(links.contains(&"https://example.com/a-link".to_string()));
}

#[test]
fn tags_restricts_to_specified_tags_only() {
    let res = make_response(
        "https://example.com/",
        r#"<a href="/a-link">a</a><area href="/area-link">"#,
    );
    let links = LinkExtractor::new().tags(&["a"]).extract(&res);
    assert_eq!(links, vec!["https://example.com/a-link"]);
    assert!(!links.contains(&"https://example.com/area-link".to_string()));
}

#[test]
fn attrs_extracts_from_custom_attribute() {
    let res = make_response(
        "https://example.com/",
        r#"<a data-href="/custom">x</a><a href="/normal">y</a>"#,
    );
    let links = LinkExtractor::new().attrs(&["data-href"]).extract(&res);
    assert_eq!(links, vec!["https://example.com/custom"]);
}