gerbeur/api/test/remote/html.test.ts

import "../setup.ts";
import { assertEquals } from "@std/assert";
import {
  extractBestIcon,
  extractFirstContentImage,
  extractJsonLd,
  extractLargeImage,
  extractMaskIconColor,
  extractMetaName,
  extractOgTag,
  extractPageTitle,
  extractThemeColor,
  normalizeCssColor,
} from "../../remote/html.ts";

const BASE = "https://example.com/articles/1";

Deno.test("extractOgTag / extractMetaName: either attribute order, entities decoded", () => {
  const page = `
    <meta property="og:title" content="Tom &amp; Jerry &#8212; &#x2764;">
    <meta content='Desc' property='og:description'>
    <meta name="twitter:title" content="TW">
    <meta name="og:title" content="wrong attribute">`;
  assertEquals(extractOgTag(page, "title"), "Tom & Jerry — ❤");
  assertEquals(extractOgTag(page, "description"), "Desc");
  assertEquals(extractOgTag(page, "image"), undefined);
  assertEquals(extractMetaName(page, "twitter:title"), "TW");
});

Deno.test("extractOgTag: content never spans into the next tag", () => {
  const page = `<meta property="og:title"><meta name="x" content="leak">`;
  assertEquals(extractOgTag(page, "title"), undefined);
});

Deno.test("extractPageTitle", () => {
  assertEquals(
    extractPageTitle("<title lang=en>  A &lt;b&gt;  </title>"),
    "A <b>",
  );
  assertEquals(extractPageTitle("<p>none</p>"), undefined);
});

Deno.test("extractThemeColor: unscoped, then light, then any", () => {
  assertEquals(
    extractThemeColor(`
      <meta name="theme-color" media="(prefers-color-scheme: dark)" content="#000">
      <meta name="theme-color" media="(prefers-color-scheme: light)" content="#fff">`),
    "#fff",
  );
  assertEquals(
    extractThemeColor(`
      <meta name="theme-color" media="(prefers-color-scheme: dark)" content="#000">
      <meta name="theme-color" content="#123">`),
    "#123",
  );
  assertEquals(
    extractThemeColor(
      `<meta name="theme-color" media="(prefers-color-scheme: dark)" content="#000">`,
    ),
    "#000",
  );
  assertEquals(
    extractMaskIconColor(
      `<link rel="mask-icon" href="/m.svg" color="#5bbad5">`,
    ),
    "#5bbad5",
  );
});

Deno.test("normalizeCssColor", () => {
  assertEquals(normalizeCssColor("#ABC"), "#aabbcc");
  assertEquals(normalizeCssColor("#abcd"), "#aabbcc");
  assertEquals(normalizeCssColor("#112233"), "#112233");
  assertEquals(normalizeCssColor("#11223380"), "#112233");
  assertEquals(normalizeCssColor(" Navy "), "#000080");
  assertEquals(normalizeCssColor("rgb(255, 0, 128)"), "#ff0080");
  assertEquals(normalizeCssColor("rgba(300 -5 50% / .5)"), "#ff0080");
  assertEquals(normalizeCssColor("#12345"), undefined);
  assertEquals(normalizeCssColor("hsl(0 100% 50%)"), undefined);
  assertEquals(normalizeCssColor("rgb(1,2)"), undefined);
  assertEquals(normalizeCssColor("transparent"), undefined);
  assertEquals(normalizeCssColor(undefined), undefined);
});

Deno.test("extractJsonLd: @graph, arrays and image shapes", () => {
  const ld = (data: unknown) =>
    `<script type="application/ld+json">${JSON.stringify(data)}</script>`;
  assertEquals(
    extractJsonLd(
      ld({
        "@graph": [{ "@type": "WebSite" }, {
          headline: "Head",
          image: { url: "https://i/1.jpg" },
        }],
      }),
    ),
    { title: "Head", description: undefined, thumbnailUrl: "https://i/1.jpg" },
  );
  assertEquals(
    extractJsonLd(ld({ name: ["N"], image: ["https://i/2.jpg"] })).thumbnailUrl,
    "https://i/2.jpg",
  );
  assertEquals(
    extractJsonLd(ld({ name: "N", image: "/relative.jpg" })).thumbnailUrl,
    undefined,
  );
  // Invalid JSON is skipped for the next block.
  assertEquals(
    extractJsonLd(
      `<script type="application/ld+json">{oops</script>${
        ld({ name: "Second" })
      }`,
    ).title,
    "Second",
  );
  assertEquals(extractJsonLd("<p></p>"), {});
});

Deno.test("image extraction: large images, content images, relative URLs", () => {
  const page = `
    <img src="data:image/png;base64,xx" width="500" height="500">
    <img src="/small.jpg" width="50" height="50">
    <img src="/img/site-logo.png">
    <img src="/icons/sprite.svg">
    <img src="/photos/cat.jpg" width="640" height="480">`;
  assertEquals(
    extractLargeImage(page, BASE),
    "https://example.com/photos/cat.jpg",
  );
  assertEquals(
    extractFirstContentImage(page, BASE),
    "https://example.com/small.jpg",
  );
  assertEquals(
    extractFirstContentImage(`<img src="//cdn.test/a.jpg">`, BASE),
    "https://cdn.test/a.jpg",
  );
});

Deno.test("extractBestIcon: SVG beats everything, else the largest", () => {
  assertEquals(
    extractBestIcon(
      `<link rel="icon" href="/16.png" sizes="16x16">
       <link rel="apple-touch-icon" href="/180.png" sizes="180x180">
       <link rel="stylesheet" href="/x.css">`,
      BASE,
    ),
    "https://example.com/180.png",
  );
  assertEquals(
    extractBestIcon(
      `<link rel="icon" href="/180.png" sizes="180x180">
       <link rel="icon" type="image/svg+xml" href="/icon">`,
      BASE,
    ),
    "https://example.com/icon",
  );
  assertEquals(extractBestIcon("<p></p>", BASE), undefined);
});