@dogsbay/search 0.0.0-stage → 0.2.0-beta.117

package/dist/html.js

@@ -0,0 +1,351 @@
+const VOID = /* @__PURE__ */ new Set(["area", "base", "br", "col", "embed", "hr", "img", "input", "link", "meta", "param", "source", "track", "wbr"]);
+const SKIP = /* @__PURE__ */ new Set(["script", "style", "noscript", "template", "svg", "nav", "aside", "form", "button", "select", "pre", "textarea", "footer"]);
+const BLOCK = /^(p|div|li|ul|ol|dl|dt|dd|h[1-6]|tr|td|th|table|section|article|main|header|blockquote|figure|figcaption|details|summary|br|hr)$/;
+const CLOSES_P = /^(p|div|ul|ol|dl|h[1-6]|table|section|article|main|header|footer|nav|aside|blockquote|figure|details|pre|hr|form)$/;
+function* tokens(html) {
+  const re = /<!--[\s\S]*?(?:-->|$)|<!\[CDATA\[[\s\S]*?(?:\]\]>|$)|<![^>]*>|<\?[^>]*>|<(script|style|template|textarea)\b((?:[^>"']|"[^"]*"|'[^']*')*)>([\s\S]*?)(?:<\/\1\s*>|$)|<(\/?)([a-zA-Z][\w:.-]*)((?:[^>"']|"[^"]*"|'[^']*')*)>|([^<]+)|</gi;
+  for (const m of html.matchAll(re)) {
+    const [whole, rawName, rawAttrs, , slash, name, attrs, text] = m;
+    if (text !== void 0) yield { kind: "text", text };
+    else if (whole === "<") yield { kind: "text", text: "<" };
+    else if (rawName) {
+      const n = rawName.toLowerCase();
+      yield { kind: "open", name: n, attrs: parseAttrs(rawAttrs ?? ""), selfClosing: false };
+      yield { kind: "close", name: n };
+    } else if (name) {
+      const n = name.toLowerCase();
+      if (slash) yield { kind: "close", name: n };
+      else yield { kind: "open", name: n, attrs: parseAttrs(attrs ?? ""), selfClosing: /(^|\s|["'])\/$/.test((attrs ?? "").trim()) };
+    }
+  }
+}
+function parseAttrs(s) {
+  const out = /* @__PURE__ */ new Map();
+  for (const m of s.matchAll(/([^\s"'>\/=]+)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|([^\s"'>]+)))?/g)) {
+    const name = m[1].toLowerCase();
+    if (!out.has(name)) out.set(name, decode(m[2] ?? m[3] ?? m[4] ?? ""));
+  }
+  return out;
+}
+function extractHtml(html, url, opts = {}) {
+  const all = [...tokens(html)];
+  let lang;
+  let titleTag = "";
+  const metas = /* @__PURE__ */ new Map();
+  let refresh = false;
+  let inTitle = false;
+  let bodyAt = all.length;
+  for (let i = 0; i < all.length; i++) {
+    const t = all[i];
+    if (t.kind === "open" && t.name === "body") {
+      bodyAt = i;
+      break;
+    }
+    if (t.kind === "open" && t.name === "html") lang = t.attrs.get("lang") || void 0;
+    else if (t.kind === "open" && t.name === "title") inTitle = true;
+    else if (t.kind === "close" && t.name === "title") inTitle = false;
+    else if (t.kind === "text" && inTitle) titleTag += t.text;
+    else if (t.kind === "open" && t.name === "meta") {
+      if ((t.attrs.get("http-equiv") ?? "").toLowerCase() === "refresh") refresh = true;
+      const name = (t.attrs.get("name") ?? "").toLowerCase();
+      if (name) metas.set(name, t.attrs.get("content") ?? "");
+    }
+  }
+  if (refresh || /\bnoindex\b/i.test(metas.get("robots") ?? "")) return void 0;
+  const marker = all.some((t) => t.kind === "open" && t.attrs.has("data-search-body")) ? "data-search-body" : all.some((t) => t.kind === "open" && t.attrs.has("data-pagefind-body")) ? "data-pagefind-body" : void 0;
+  const has = (n) => all.some((t) => t.kind === "open" && t.name === n);
+  const rootTag = marker ? void 0 : has("main") ? "main" : has("article") ? "article" : "body";
+  const stack = [];
+  let skipDepth = 0;
+  let rootDepth = 0;
+  let rootSeen = false;
+  let heading;
+  let tagItem;
+  let tagListDepth = 0;
+  const body = [];
+  const headings = [];
+  const anchors = [];
+  const tags = [];
+  let h1;
+  const push = (o) => {
+    stack.push(o);
+    if (o.skip) skipDepth++;
+    if (o.root) rootDepth++;
+    if (o.heading) heading = o.heading;
+    if (o.tagList) tagListDepth++;
+    if (o.tagItem) tagItem = o.tagItem;
+  };
+  const pop = () => {
+    const o = stack.pop();
+    if (o.skip) skipDepth--;
+    if (o.root) rootDepth--;
+    if (o.tagList) tagListDepth--;
+    if (BLOCK.test(o.name)) (heading ? heading.text : body).push(" ");
+    if (o.permalink && /^\s*[#¶🔗§]?\s*$/u.test(o.permalink.heading.slice(o.permalink.from).join(""))) o.permalink.heading.length = o.permalink.from;
+    if (o.tagItem) {
+      const t = squash(o.tagItem.join("")).replace(/^#\s*/, "");
+      if (t) tags.push(t);
+      tagItem = void 0;
+    }
+    if (o.heading) {
+      heading = void 0;
+      const text = squash(o.heading.text.join("")).replace(/\s*[¶🔗]$/u, "");
+      if (!text) return;
+      if (o.heading.level === 1 && h1 === void 0) h1 = text;
+      else {
+        headings.push(text);
+        if (o.heading.level <= 3 && o.heading.id) anchors.push([text, o.heading.id]);
+      }
+    }
+  };
+  const closeTo = (names, limit = Infinity) => {
+    for (let i = stack.length - 1, n = 0; i >= 0 && n < limit; i--, n++) {
+      if (names.test(stack[i].name)) {
+        while (stack.length > i) pop();
+        return;
+      }
+      if (/^(ul|ol|dl|table|body|main|article|section|div)$/.test(stack[i].name)) return;
+    }
+  };
+  for (let i = bodyAt; i < all.length; i++) {
+    const t = all[i];
+    if (t.kind === "text") {
+      if (!rootDepth || skipDepth) continue;
+      const text = decode(t.text);
+      (heading ? heading.text : body).push(text);
+      tagItem?.push(text);
+      continue;
+    }
+    if (t.kind === "close") {
+      let j = stack.length - 1;
+      while (j >= 0 && stack[j].name !== t.name) j--;
+      if (j >= 0) while (stack.length > j) pop();
+      continue;
+    }
+    const { name, attrs } = t;
+    if (CLOSES_P.test(name)) closeTo(/^p$/, 64);
+    if (name === "li") closeTo(/^li$/);
+    else if (name === "dt" || name === "dd") closeTo(/^d[td]$/);
+    else if (name === "tr") closeTo(/^tr$/);
+    else if (name === "td" || name === "th") closeTo(/^t[dh]$/);
+    else if (name === "option") closeTo(/^option$/);
+    if (name === "img" && rootDepth && !skipDepth) {
+      const alt = attrs.get("alt");
+      if (alt) (heading ? heading.text : body).push(` ${alt} `);
+    }
+    if (VOID.has(name) || t.selfClosing) {
+      if ((name === "br" || name === "hr") && rootDepth) (heading ? heading.text : body).push(" ");
+      continue;
+    }
+    const ignore = attrs.has("data-search-ignore") || attrs.has("data-pagefind-ignore") || attrs.get("aria-hidden") === "true";
+    const isRoot = !rootSeen && !skipDepth && !ignore && (marker ? attrs.has(marker) : name === rootTag);
+    if (isRoot) rootSeen = true;
+    const skip = (SKIP.has(name) || name === "header" && !rootDepth && !isRoot || ignore) && !isRoot;
+    const o = { name, root: isRoot, skip };
+    const level = /^h([1-6])$/.exec(name)?.[1];
+    if (level && !heading) o.heading = { level: Number(level), id: attrs.get("id") || void 0, text: [] };
+    if (name === "a" && heading && (attrs.get("href") ?? "").startsWith("#")) o.permalink = { heading: heading.text, from: heading.text.length };
+    if (attrs.has("data-tag-list")) o.tagList = true;
+    if (name === "li" && tagListDepth) o.tagItem = [];
+    if (BLOCK.test(name)) (heading ? heading.text : body).push(" ");
+    push(o);
+  }
+  while (stack.length) pop();
+  const bodyText = squash(body.join(""));
+  const title = h1 ?? squash(decode(titleTag));
+  if (!bodyText && !title) return void 0;
+  const keywords = (metas.get("keywords") ?? "").split(",").map((k) => k.trim()).filter(Boolean);
+  return {
+    doc: { title: title || url, url, section: opts.section ?? folderName(url), excerpt: excerpt(metas.get("description") || bodyText), anchors, tags },
+    parts: { title: title || url, headings, body: bodyText, keywords: [...keywords, ...tags] },
+    lang
+  };
+}
+function squash(s) {
+  return s.replace(/\s+/g, " ").trim();
+}
+function excerpt(s) {
+  if (s.length <= 240) return s;
+  const cut = s.slice(0, 240);
+  return cut.slice(0, cut.lastIndexOf(" ") > 180 ? cut.lastIndexOf(" ") : 240) + "\u2026";
+}
+function folderName(url) {
+  const parts = url.split("/");
+  if (parts.length < 2) return "";
+  let f = parts[0];
+  try {
+    f = decodeURIComponent(f);
+  } catch {
+  }
+  f = f.replace(/[-_]+/g, " ");
+  return f.charAt(0).toUpperCase() + f.slice(1);
+}
+const NAMED = {
+  quot: 34,
+  amp: 38,
+  apos: 39,
+  lt: 60,
+  gt: 62,
+  nbsp: 160,
+  iexcl: 161,
+  cent: 162,
+  pound: 163,
+  curren: 164,
+  yen: 165,
+  brvbar: 166,
+  sect: 167,
+  uml: 168,
+  copy: 169,
+  ordf: 170,
+  laquo: 171,
+  not: 172,
+  shy: 173,
+  reg: 174,
+  macr: 175,
+  deg: 176,
+  plusmn: 177,
+  sup2: 178,
+  sup3: 179,
+  acute: 180,
+  micro: 181,
+  para: 182,
+  middot: 183,
+  cedil: 184,
+  sup1: 185,
+  ordm: 186,
+  raquo: 187,
+  frac14: 188,
+  frac12: 189,
+  frac34: 190,
+  iquest: 191,
+  Agrave: 192,
+  Aacute: 193,
+  Acirc: 194,
+  Atilde: 195,
+  Auml: 196,
+  Aring: 197,
+  AElig: 198,
+  Ccedil: 199,
+  Egrave: 200,
+  Eacute: 201,
+  Ecirc: 202,
+  Euml: 203,
+  Igrave: 204,
+  Iacute: 205,
+  Icirc: 206,
+  Iuml: 207,
+  ETH: 208,
+  Ntilde: 209,
+  Ograve: 210,
+  Oacute: 211,
+  Ocirc: 212,
+  Otilde: 213,
+  Ouml: 214,
+  times: 215,
+  Oslash: 216,
+  Ugrave: 217,
+  Uacute: 218,
+  Ucirc: 219,
+  Uuml: 220,
+  Yacute: 221,
+  THORN: 222,
+  szlig: 223,
+  agrave: 224,
+  aacute: 225,
+  acirc: 226,
+  atilde: 227,
+  auml: 228,
+  aring: 229,
+  aelig: 230,
+  ccedil: 231,
+  egrave: 232,
+  eacute: 233,
+  ecirc: 234,
+  euml: 235,
+  igrave: 236,
+  iacute: 237,
+  icirc: 238,
+  iuml: 239,
+  eth: 240,
+  ntilde: 241,
+  ograve: 242,
+  oacute: 243,
+  ocirc: 244,
+  otilde: 245,
+  ouml: 246,
+  divide: 247,
+  oslash: 248,
+  ugrave: 249,
+  uacute: 250,
+  ucirc: 251,
+  uuml: 252,
+  yacute: 253,
+  thorn: 254,
+  yuml: 255,
+  OElig: 338,
+  oelig: 339,
+  Scaron: 352,
+  scaron: 353,
+  Yuml: 376,
+  fnof: 402,
+  circ: 710,
+  tilde: 732,
+  ensp: 8194,
+  emsp: 8195,
+  thinsp: 8201,
+  zwnj: 8204,
+  zwj: 8205,
+  lrm: 8206,
+  rlm: 8207,
+  ndash: 8211,
+  mdash: 8212,
+  lsquo: 8216,
+  rsquo: 8217,
+  sbquo: 8218,
+  ldquo: 8220,
+  rdquo: 8221,
+  bdquo: 8222,
+  dagger: 8224,
+  Dagger: 8225,
+  bull: 8226,
+  hellip: 8230,
+  permil: 8240,
+  prime: 8242,
+  Prime: 8243,
+  lsaquo: 8249,
+  rsaquo: 8250,
+  euro: 8364,
+  trade: 8482,
+  larr: 8592,
+  uarr: 8593,
+  rarr: 8594,
+  darr: 8595,
+  harr: 8596,
+  rArr: 8658,
+  lArr: 8656,
+  hArr: 8660,
+  minus: 8722,
+  le: 8804,
+  ge: 8805,
+  ne: 8800,
+  asymp: 8776,
+  infin: 8734,
+  check: 10003,
+  hearts: 9829
+};
+function decode(s) {
+  if (!s.includes("&")) return s;
+  return s.replace(/&(#x[0-9a-f]+|#\d+|[a-z][a-z0-9]*);/gi, (m, e) => {
+    if (e[0] === "#") {
+      const n2 = e[1] === "x" || e[1] === "X" ? parseInt(e.slice(2), 16) : Number(e.slice(1));
+      return Number.isFinite(n2) && n2 > 0 && n2 < 1114112 ? String.fromCodePoint(n2) : m;
+    }
+    const n = NAMED[e] ?? NAMED[e.toLowerCase()];
+    return n === void 0 ? m : String.fromCodePoint(n);
+  });
+}
+export {
+  extractHtml,
+  folderName
+};
+//# sourceMappingURL=html.js.map

package/dist/html.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/html.ts"],"sourcesContent":["/**\n * Index built HTML, from any site generator: extract a page's title,\n * headings, body text, keywords and tags from its markup, for\n * SearchIndexBuilder.add(). A small tokenizer, not a DOM, so it runs in a\n * Worker and on large sites without a parser dependency.\n *\n * What is indexed:\n *   root     the element with a `data-search-body` attribute (or Pagefind's\n *            `data-pagefind-body`), else the first <main>, else <article>,\n *            else <body>; never one inside a skipped element\n *   skipped  script, style, template, noscript, svg, nav, aside, form,\n *            button, select, textarea, pre (code blocks), footer, header\n *            outside the root, aria-hidden content, and anything with\n *            `data-search-ignore` or `data-pagefind-ignore`\n *   title    the first <h1> in the root, else <title>; later h1s are headings\n *   headings h2–h6 (and later h1s) in the root; h2/h3 with an id are anchors\n *   keywords <meta name=\"keywords\">, and tag lists (`data-tag-list`)\n * A page with a meta refresh or a robots `noindex`, or with neither a\n * title nor body text, is not indexed.\n */\nimport type { SearchDoc } from \"./index.js\";\n\nexport interface HtmlPage {\n  doc: SearchDoc;\n  parts: { title: string; headings: string[]; body: string; keywords: string[] };\n  lang?: string;\n}\n\ntype Attrs = Map<string, string>;\ntype Token = { kind: \"open\"; name: string; attrs: Attrs; selfClosing: boolean } | { kind: \"close\"; name: string } | { kind: \"text\"; text: string };\n\nconst VOID = new Set([\"area\", \"base\", \"br\", \"col\", \"embed\", \"hr\", \"img\", \"input\", \"link\", \"meta\", \"param\", \"source\", \"track\", \"wbr\"]);\nconst SKIP = new Set([\"script\", \"style\", \"noscript\", \"template\", \"svg\", \"nav\", \"aside\", \"form\", \"button\", \"select\", \"pre\", \"textarea\", \"footer\"]);\nconst BLOCK = /^(p|div|li|ul|ol|dl|dt|dd|h[1-6]|tr|td|th|table|section|article|main|header|blockquote|figure|figcaption|details|summary|br|hr)$/;\n/** Elements that close an open <p> (the HTML parser's implied end tag). */\nconst CLOSES_P = /^(p|div|ul|ol|dl|h[1-6]|table|section|article|main|header|footer|nav|aside|blockquote|figure|details|pre|hr|form)$/;\n\n/**\n * Tokens of a document: tags with parsed attributes, and text. Comments,\n * CDATA, doctypes and the contents of script, style and template are left\n * out here, so nothing inside them is ever mistaken for markup.\n */\nfunction* tokens(html: string): Generator<Token> {\n  const re = /<!--[\\s\\S]*?(?:-->|$)|<!\\[CDATA\\[[\\s\\S]*?(?:\\]\\]>|$)|<![^>]*>|<\\?[^>]*>|<(script|style|template|textarea)\\b((?:[^>\"']|\"[^\"]*\"|'[^']*')*)>([\\s\\S]*?)(?:<\\/\\1\\s*>|$)|<(\\/?)([a-zA-Z][\\w:.-]*)((?:[^>\"']|\"[^\"]*\"|'[^']*')*)>|([^<]+)|</gi;\n  for (const m of html.matchAll(re)) {\n    const [whole, rawName, rawAttrs, , slash, name, attrs, text] = m;\n    if (text !== undefined) yield { kind: \"text\", text };\n    else if (whole === \"<\") yield { kind: \"text\", text: \"<\" };\n    else if (rawName) {\n      // Kept as an element with no content, so skip rules and structure still apply.\n      const n = rawName.toLowerCase();\n      yield { kind: \"open\", name: n, attrs: parseAttrs(rawAttrs ?? \"\"), selfClosing: false };\n      yield { kind: \"close\", name: n };\n    } else if (name) {\n      const n = name.toLowerCase();\n      if (slash) yield { kind: \"close\", name: n };\n      else yield { kind: \"open\", name: n, attrs: parseAttrs(attrs ?? \"\"), selfClosing: /(^|\\s|[\"'])\\/$/.test((attrs ?? \"\").trim()) };\n    }\n  }\n}\n\nfunction parseAttrs(s: string): Attrs {\n  const out: Attrs = new Map();\n  for (const m of s.matchAll(/([^\\s\"'>\\/=]+)(?:\\s*=\\s*(?:\"([^\"]*)\"|'([^']*)'|([^\\s\"'>]+)))?/g)) {\n    const name = m[1].toLowerCase();\n    if (!out.has(name)) out.set(name, decode(m[2] ?? m[3] ?? m[4] ?? \"\"));\n  }\n  return out;\n}\n\ninterface Open {\n  name: string;\n  skip: boolean;\n  root: boolean;\n  heading?: { level: number; id?: string; text: string[] };\n  /** A heading's permalink (<a href=\"#…\">): its text is dropped when it is only a mark. */\n  permalink?: { heading: string[]; from: number };\n  tagList?: boolean;\n  tagItem?: string[];\n}\n\n/**\n * Extract one page. `url` is the page's URL relative to the site (what\n * results link to); `section` defaults to its first folder.\n */\nexport function extractHtml(html: string, url: string, opts: { section?: string } = {}): HtmlPage | undefined {\n  const all = [...tokens(html)];\n\n  // Head: <title>, <meta>, <html lang>. Everything before <body>, by token, not by text search.\n  let lang: string | undefined;\n  let titleTag = \"\";\n  const metas = new Map<string, string>();\n  let refresh = false;\n  let inTitle = false;\n  let bodyAt = all.length;\n  for (let i = 0; i < all.length; i++) {\n    const t = all[i];\n    if (t.kind === \"open\" && t.name === \"body\") {\n      bodyAt = i;\n      break;\n    }\n    if (t.kind === \"open\" && t.name === \"html\") lang = t.attrs.get(\"lang\") || undefined;\n    else if (t.kind === \"open\" && t.name === \"title\") inTitle = true;\n    else if (t.kind === \"close\" && t.name === \"title\") inTitle = false;\n    else if (t.kind === \"text\" && inTitle) titleTag += t.text;\n    else if (t.kind === \"open\" && t.name === \"meta\") {\n      if ((t.attrs.get(\"http-equiv\") ?? \"\").toLowerCase() === \"refresh\") refresh = true;\n      const name = (t.attrs.get(\"name\") ?? \"\").toLowerCase();\n      if (name) metas.set(name, t.attrs.get(\"content\") ?? \"\");\n    }\n  }\n  if (refresh || /\\bnoindex\\b/i.test(metas.get(\"robots\") ?? \"\")) return undefined;\n\n  // The root: the marked element, else the first <main>, <article> or <body> that is not inside a skipped element.\n  const marker = all.some((t) => t.kind === \"open\" && t.attrs.has(\"data-search-body\"))\n    ? \"data-search-body\"\n    : all.some((t) => t.kind === \"open\" && t.attrs.has(\"data-pagefind-body\"))\n      ? \"data-pagefind-body\"\n      : undefined;\n  const has = (n: string) => all.some((t) => t.kind === \"open\" && t.name === n);\n  const rootTag = marker ? undefined : has(\"main\") ? \"main\" : has(\"article\") ? \"article\" : \"body\";\n\n  const stack: Open[] = [];\n  let skipDepth = 0; // open skipped elements\n  let rootDepth = 0; // open root elements (0 or 1)\n  let rootSeen = false;\n  let heading: Open[\"heading\"];\n  let tagItem: string[] | undefined;\n  let tagListDepth = 0;\n  const body: string[] = [];\n  const headings: string[] = [];\n  const anchors: [string, string][] = [];\n  const tags: string[] = [];\n  let h1: string | undefined;\n\n  const push = (o: Open) => {\n    stack.push(o);\n    if (o.skip) skipDepth++;\n    if (o.root) rootDepth++;\n    if (o.heading) heading = o.heading;\n    if (o.tagList) tagListDepth++;\n    if (o.tagItem) tagItem = o.tagItem;\n  };\n  const pop = () => {\n    const o = stack.pop()!;\n    if (o.skip) skipDepth--;\n    if (o.root) rootDepth--;\n    if (o.tagList) tagListDepth--;\n    if (BLOCK.test(o.name)) (heading ? heading.text : body).push(\" \");\n    if (o.permalink && /^\\s*[#¶🔗§]?\\s*$/u.test(o.permalink.heading.slice(o.permalink.from).join(\"\"))) o.permalink.heading.length = o.permalink.from;\n    if (o.tagItem) {\n      const t = squash(o.tagItem.join(\"\")).replace(/^#\\s*/, \"\");\n      if (t) tags.push(t);\n      tagItem = undefined;\n    }\n    if (o.heading) {\n      heading = undefined;\n      // A trailing ¶ or 🔗 is a permalink mark left as text; a trailing # is kept (it ends \"C#\").\n      const text = squash(o.heading.text.join(\"\")).replace(/\\s*[¶🔗]$/u, \"\");\n      if (!text) return;\n      if (o.heading.level === 1 && h1 === undefined) h1 = text;\n      else {\n        headings.push(text);\n        if (o.heading.level <= 3 && o.heading.id) anchors.push([text, o.heading.id]);\n      }\n    }\n  };\n  /** Close open elements down to and including the last one named `name`, if it is open (within `limit` elements). */\n  const closeTo = (names: RegExp, limit = Infinity) => {\n    for (let i = stack.length - 1, n = 0; i >= 0 && n < limit; i--, n++) {\n      if (names.test(stack[i].name)) {\n        while (stack.length > i) pop();\n        return;\n      }\n      if (/^(ul|ol|dl|table|body|main|article|section|div)$/.test(stack[i].name)) return; // not across a container\n    }\n  };\n\n  for (let i = bodyAt; i < all.length; i++) {\n    const t = all[i];\n    if (t.kind === \"text\") {\n      if (!rootDepth || skipDepth) continue;\n      const text = decode(t.text);\n      (heading ? heading.text : body).push(text);\n      tagItem?.push(text);\n      continue;\n    }\n    if (t.kind === \"close\") {\n      let j = stack.length - 1;\n      while (j >= 0 && stack[j].name !== t.name) j--;\n      if (j >= 0) while (stack.length > j) pop();\n      continue;\n    }\n    const { name, attrs } = t;\n    // Implied end tags, as the HTML parser applies them, so unclosed <p> and <li> do not nest without limit.\n    if (CLOSES_P.test(name)) closeTo(/^p$/, 64);\n    if (name === \"li\") closeTo(/^li$/);\n    else if (name === \"dt\" || name === \"dd\") closeTo(/^d[td]$/);\n    else if (name === \"tr\") closeTo(/^tr$/);\n    else if (name === \"td\" || name === \"th\") closeTo(/^t[dh]$/);\n    else if (name === \"option\") closeTo(/^option$/);\n\n    if (name === \"img\" && rootDepth && !skipDepth) {\n      const alt = attrs.get(\"alt\");\n      if (alt) (heading ? heading.text : body).push(` ${alt} `);\n    }\n    if (VOID.has(name) || t.selfClosing) {\n      if ((name === \"br\" || name === \"hr\") && rootDepth) (heading ? heading.text : body).push(\" \");\n      continue;\n    }\n    const ignore = attrs.has(\"data-search-ignore\") || attrs.has(\"data-pagefind-ignore\") || attrs.get(\"aria-hidden\") === \"true\";\n    const isRoot = !rootSeen && !skipDepth && !ignore && (marker ? attrs.has(marker) : name === rootTag);\n    if (isRoot) rootSeen = true;\n    // A <header> inside the root is the page's own header (it often holds the <h1>); outside, it is the site's.\n    const skip = (SKIP.has(name) || (name === \"header\" && !rootDepth && !isRoot) || ignore) && !isRoot;\n    const o: Open = { name, root: isRoot, skip };\n    const level = /^h([1-6])$/.exec(name)?.[1];\n    if (level && !heading) o.heading = { level: Number(level), id: attrs.get(\"id\") || undefined, text: [] };\n    if (name === \"a\" && heading && (attrs.get(\"href\") ?? \"\").startsWith(\"#\")) o.permalink = { heading: heading.text, from: heading.text.length };\n    if (attrs.has(\"data-tag-list\")) o.tagList = true;\n    if (name === \"li\" && tagListDepth) o.tagItem = [];\n    if (BLOCK.test(name)) (heading ? heading.text : body).push(\" \");\n    push(o);\n  }\n  while (stack.length) pop();\n\n  const bodyText = squash(body.join(\"\"));\n  const title = h1 ?? squash(decode(titleTag));\n  if (!bodyText && !title) return undefined;\n  const keywords = (metas.get(\"keywords\") ?? \"\").split(\",\").map((k) => k.trim()).filter(Boolean);\n  return {\n    doc: { title: title || url, url, section: opts.section ?? folderName(url), excerpt: excerpt(metas.get(\"description\") || bodyText), anchors, tags },\n    parts: { title: title || url, headings, body: bodyText, keywords: [...keywords, ...tags] },\n    lang,\n  };\n}\n\nfunction squash(s: string): string {\n  return s.replace(/\\s+/g, \" \").trim();\n}\n\nfunction excerpt(s: string): string {\n  if (s.length <= 240) return s;\n  const cut = s.slice(0, 240);\n  return cut.slice(0, cut.lastIndexOf(\" \") > 180 ? cut.lastIndexOf(\" \") : 240) + \"…\";\n}\n\n/** \"guide/install.html\" → \"Guide\"; top-level pages have no section. Works on encoded and raw paths. */\nexport function folderName(url: string): string {\n  const parts = url.split(\"/\");\n  if (parts.length < 2) return \"\";\n  let f = parts[0];\n  try {\n    f = decodeURIComponent(f);\n  } catch {\n    /* a raw name with \"%\" */\n  }\n  f = f.replace(/[-_]+/g, \" \");\n  return f.charAt(0).toUpperCase() + f.slice(1);\n}\n\n// Named entities: all of Latin-1 plus the common typographic and symbol ones. Numeric references are always decoded.\nconst NAMED: Record<string, number> = {\n  quot: 34, amp: 38, apos: 39, lt: 60, gt: 62, nbsp: 160, iexcl: 161, cent: 162, pound: 163, curren: 164, yen: 165, brvbar: 166, sect: 167, uml: 168, copy: 169,\n  ordf: 170, laquo: 171, not: 172, shy: 173, reg: 174, macr: 175, deg: 176, plusmn: 177, sup2: 178, sup3: 179, acute: 180, micro: 181, para: 182, middot: 183,\n  cedil: 184, sup1: 185, ordm: 186, raquo: 187, frac14: 188, frac12: 189, frac34: 190, iquest: 191, Agrave: 192, Aacute: 193, Acirc: 194, Atilde: 195, Auml: 196,\n  Aring: 197, AElig: 198, Ccedil: 199, Egrave: 200, Eacute: 201, Ecirc: 202, Euml: 203, Igrave: 204, Iacute: 205, Icirc: 206, Iuml: 207, ETH: 208, Ntilde: 209,\n  Ograve: 210, Oacute: 211, Ocirc: 212, Otilde: 213, Ouml: 214, times: 215, Oslash: 216, Ugrave: 217, Uacute: 218, Ucirc: 219, Uuml: 220, Yacute: 221, THORN: 222,\n  szlig: 223, agrave: 224, aacute: 225, acirc: 226, atilde: 227, auml: 228, aring: 229, aelig: 230, ccedil: 231, egrave: 232, eacute: 233, ecirc: 234, euml: 235,\n  igrave: 236, iacute: 237, icirc: 238, iuml: 239, eth: 240, ntilde: 241, ograve: 242, oacute: 243, ocirc: 244, otilde: 245, ouml: 246, divide: 247, oslash: 248,\n  ugrave: 249, uacute: 250, ucirc: 251, uuml: 252, yacute: 253, thorn: 254, yuml: 255, OElig: 338, oelig: 339, Scaron: 352, scaron: 353, Yuml: 376, fnof: 402,\n  circ: 710, tilde: 732, ensp: 8194, emsp: 8195, thinsp: 8201, zwnj: 8204, zwj: 8205, lrm: 8206, rlm: 8207, ndash: 8211, mdash: 8212, lsquo: 8216, rsquo: 8217,\n  sbquo: 8218, ldquo: 8220, rdquo: 8221, bdquo: 8222, dagger: 8224, Dagger: 8225, bull: 8226, hellip: 8230, permil: 8240, prime: 8242, Prime: 8243, lsaquo: 8249,\n  rsaquo: 8250, euro: 8364, trade: 8482, larr: 8592, uarr: 8593, rarr: 8594, darr: 8595, harr: 8596, rArr: 8658, lArr: 8656, hArr: 8660, minus: 8722, le: 8804,\n  ge: 8805, ne: 8800, asymp: 8776, infin: 8734, check: 10003, hearts: 9829,\n};\n\nfunction decode(s: string): string {\n  if (!s.includes(\"&\")) return s;\n  return s.replace(/&(#x[0-9a-f]+|#\\d+|[a-z][a-z0-9]*);/gi, (m, e: string) => {\n    if (e[0] === \"#\") {\n      const n = e[1] === \"x\" || e[1] === \"X\" ? parseInt(e.slice(2), 16) : Number(e.slice(1));\n      return Number.isFinite(n) && n > 0 && n < 0x110000 ? String.fromCodePoint(n) : m;\n    }\n    const n = NAMED[e] ?? NAMED[e.toLowerCase()];\n    return n === undefined ? m : String.fromCodePoint(n);\n  });\n}\n"],"mappings":"AA+BA,MAAM,OAAO,oBAAI,IAAI,CAAC,QAAQ,QAAQ,MAAM,OAAO,SAAS,MAAM,OAAO,SAAS,QAAQ,QAAQ,SAAS,UAAU,SAAS,KAAK,CAAC;AACpI,MAAM,OAAO,oBAAI,IAAI,CAAC,UAAU,SAAS,YAAY,YAAY,OAAO,OAAO,SAAS,QAAQ,UAAU,UAAU,OAAO,YAAY,QAAQ,CAAC;AAChJ,MAAM,QAAQ;AAEd,MAAM,WAAW;AAOjB,UAAU,OAAO,MAAgC;AAC/C,QAAM,KAAK;AACX,aAAW,KAAK,KAAK,SAAS,EAAE,GAAG;AACjC,UAAM,CAAC,OAAO,SAAS,UAAU,EAAE,OAAO,MAAM,OAAO,IAAI,IAAI;AAC/D,QAAI,SAAS,OAAW,OAAM,EAAE,MAAM,QAAQ,KAAK;AAAA,aAC1C,UAAU,IAAK,OAAM,EAAE,MAAM,QAAQ,MAAM,IAAI;AAAA,aAC/C,SAAS;AAEhB,YAAM,IAAI,QAAQ,YAAY;AAC9B,YAAM,EAAE,MAAM,QAAQ,MAAM,GAAG,OAAO,WAAW,YAAY,EAAE,GAAG,aAAa,MAAM;AACrF,YAAM,EAAE,MAAM,SAAS,MAAM,EAAE;AAAA,IACjC,WAAW,MAAM;AACf,YAAM,IAAI,KAAK,YAAY;AAC3B,UAAI,MAAO,OAAM,EAAE,MAAM,SAAS,MAAM,EAAE;AAAA,UACrC,OAAM,EAAE,MAAM,QAAQ,MAAM,GAAG,OAAO,WAAW,SAAS,EAAE,GAAG,aAAa,iBAAiB,MAAM,SAAS,IAAI,KAAK,CAAC,EAAE;AAAA,IAC/H;AAAA,EACF;AACF;AAEA,SAAS,WAAW,GAAkB;AACpC,QAAM,MAAa,oBAAI,IAAI;AAC3B,aAAW,KAAK,EAAE,SAAS,gEAAgE,GAAG;AAC5F,UAAM,OAAO,EAAE,CAAC,EAAE,YAAY;AAC9B,QAAI,CAAC,IAAI,IAAI,IAAI,EAAG,KAAI,IAAI,MAAM,OAAO,EAAE,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,CAAC;AAAA,EACtE;AACA,SAAO;AACT;AAiBO,SAAS,YAAY,MAAc,KAAa,OAA6B,CAAC,GAAyB;AAC5G,QAAM,MAAM,CAAC,GAAG,OAAO,IAAI,CAAC;AAG5B,MAAI;AACJ,MAAI,WAAW;AACf,QAAM,QAAQ,oBAAI,IAAoB;AACtC,MAAI,UAAU;AACd,MAAI,UAAU;AACd,MAAI,SAAS,IAAI;AACjB,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,KAAK;AACnC,UAAM,IAAI,IAAI,CAAC;AACf,QAAI,EAAE,SAAS,UAAU,EAAE,SAAS,QAAQ;AAC1C,eAAS;AACT;AAAA,IACF;AACA,QAAI,EAAE,SAAS,UAAU,EAAE,SAAS,OAAQ,QAAO,EAAE,MAAM,IAAI,MAAM,KAAK;AAAA,aACjE,EAAE,SAAS,UAAU,EAAE,SAAS,QAAS,WAAU;AAAA,aACnD,EAAE,SAAS,WAAW,EAAE,SAAS,QAAS,WAAU;AAAA,aACpD,EAAE,SAAS,UAAU,QAAS,aAAY,EAAE;AAAA,aAC5C,EAAE,SAAS,UAAU,EAAE,SAAS,QAAQ;AAC/C,WAAK,EAAE,MAAM,IAAI,YAAY,KAAK,IAAI,YAAY,MAAM,UAAW,WAAU;AAC7E,YAAM,QAAQ,EAAE,MAAM,IAAI,MAAM,KAAK,IAAI,YAAY;AACrD,UAAI,KAAM,OAAM,IAAI,MAAM,EAAE,MAAM,IAAI,SAAS,KAAK,EAAE;AAAA,IACxD;AAAA,EACF;AACA,MAAI,WAAW,eAAe,KAAK,MAAM,IAAI,QAAQ,KAAK,EAAE,EAAG,QAAO;AAGtE,QAAM,SAAS,IAAI,KAAK,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,MAAM,IAAI,kBAAkB,CAAC,IAC/E,qBACA,IAAI,KAAK,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,MAAM,IAAI,oBAAoB,CAAC,IACpE,uBACA;AACN,QAAM,MAAM,CAAC,MAAc,IAAI,KAAK,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,SAAS,CAAC;AAC5E,QAAM,UAAU,SAAS,SAAY,IAAI,MAAM,IAAI,SAAS,IAAI,SAAS,IAAI,YAAY;AAEzF,QAAM,QAAgB,CAAC;AACvB,MAAI,YAAY;AAChB,MAAI,YAAY;AAChB,MAAI,WAAW;AACf,MAAI;AACJ,MAAI;AACJ,MAAI,eAAe;AACnB,QAAM,OAAiB,CAAC;AACxB,QAAM,WAAqB,CAAC;AAC5B,QAAM,UAA8B,CAAC;AACrC,QAAM,OAAiB,CAAC;AACxB,MAAI;AAEJ,QAAM,OAAO,CAAC,MAAY;AACxB,UAAM,KAAK,CAAC;AACZ,QAAI,EAAE,KAAM;AACZ,QAAI,EAAE,KAAM;AACZ,QAAI,EAAE,QAAS,WAAU,EAAE;AAC3B,QAAI,EAAE,QAAS;AACf,QAAI,EAAE,QAAS,WAAU,EAAE;AAAA,EAC7B;AACA,QAAM,MAAM,MAAM;AAChB,UAAM,IAAI,MAAM,IAAI;AACpB,QAAI,EAAE,KAAM;AACZ,QAAI,EAAE,KAAM;AACZ,QAAI,EAAE,QAAS;AACf,QAAI,MAAM,KAAK,EAAE,IAAI,EAAG,EAAC,UAAU,QAAQ,OAAO,MAAM,KAAK,GAAG;AAChE,QAAI,EAAE,aAAa,oBAAoB,KAAK,EAAE,UAAU,QAAQ,MAAM,EAAE,UAAU,IAAI,EAAE,KAAK,EAAE,CAAC,EAAG,GAAE,UAAU,QAAQ,SAAS,EAAE,UAAU;AAC5I,QAAI,EAAE,SAAS;AACb,YAAM,IAAI,OAAO,EAAE,QAAQ,KAAK,EAAE,CAAC,EAAE,QAAQ,SAAS,EAAE;AACxD,UAAI,EAAG,MAAK,KAAK,CAAC;AAClB,gBAAU;AAAA,IACZ;AACA,QAAI,EAAE,SAAS;AACb,gBAAU;AAEV,YAAM,OAAO,OAAO,EAAE,QAAQ,KAAK,KAAK,EAAE,CAAC,EAAE,QAAQ,cAAc,EAAE;AACrE,UAAI,CAAC,KAAM;AACX,UAAI,EAAE,QAAQ,UAAU,KAAK,OAAO,OAAW,MAAK;AAAA,WAC/C;AACH,iBAAS,KAAK,IAAI;AAClB,YAAI,EAAE,QAAQ,SAAS,KAAK,EAAE,QAAQ,GAAI,SAAQ,KAAK,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC;AAAA,MAC7E;AAAA,IACF;AAAA,EACF;AAEA,QAAM,UAAU,CAAC,OAAe,QAAQ,aAAa;AACnD,aAAS,IAAI,MAAM,SAAS,GAAG,IAAI,GAAG,KAAK,KAAK,IAAI,OAAO,KAAK,KAAK;AACnE,UAAI,MAAM,KAAK,MAAM,CAAC,EAAE,IAAI,GAAG;AAC7B,eAAO,MAAM,SAAS,EAAG,KAAI;AAC7B;AAAA,MACF;AACA,UAAI,mDAAmD,KAAK,MAAM,CAAC,EAAE,IAAI,EAAG;AAAA,IAC9E;AAAA,EACF;AAEA,WAAS,IAAI,QAAQ,IAAI,IAAI,QAAQ,KAAK;AACxC,UAAM,IAAI,IAAI,CAAC;AACf,QAAI,EAAE,SAAS,QAAQ;AACrB,UAAI,CAAC,aAAa,UAAW;AAC7B,YAAM,OAAO,OAAO,EAAE,IAAI;AAC1B,OAAC,UAAU,QAAQ,OAAO,MAAM,KAAK,IAAI;AACzC,eAAS,KAAK,IAAI;AAClB;AAAA,IACF;AACA,QAAI,EAAE,SAAS,SAAS;AACtB,UAAI,IAAI,MAAM,SAAS;AACvB,aAAO,KAAK,KAAK,MAAM,CAAC,EAAE,SAAS,EAAE,KAAM;AAC3C,UAAI,KAAK,EAAG,QAAO,MAAM,SAAS,EAAG,KAAI;AACzC;AAAA,IACF;AACA,UAAM,EAAE,MAAM,MAAM,IAAI;AAExB,QAAI,SAAS,KAAK,IAAI,EAAG,SAAQ,OAAO,EAAE;AAC1C,QAAI,SAAS,KAAM,SAAQ,MAAM;AAAA,aACxB,SAAS,QAAQ,SAAS,KAAM,SAAQ,SAAS;AAAA,aACjD,SAAS,KAAM,SAAQ,MAAM;AAAA,aAC7B,SAAS,QAAQ,SAAS,KAAM,SAAQ,SAAS;AAAA,aACjD,SAAS,SAAU,SAAQ,UAAU;AAE9C,QAAI,SAAS,SAAS,aAAa,CAAC,WAAW;AAC7C,YAAM,MAAM,MAAM,IAAI,KAAK;AAC3B,UAAI,IAAK,EAAC,UAAU,QAAQ,OAAO,MAAM,KAAK,IAAI,GAAG,GAAG;AAAA,IAC1D;AACA,QAAI,KAAK,IAAI,IAAI,KAAK,EAAE,aAAa;AACnC,WAAK,SAAS,QAAQ,SAAS,SAAS,UAAW,EAAC,UAAU,QAAQ,OAAO,MAAM,KAAK,GAAG;AAC3F;AAAA,IACF;AACA,UAAM,SAAS,MAAM,IAAI,oBAAoB,KAAK,MAAM,IAAI,sBAAsB,KAAK,MAAM,IAAI,aAAa,MAAM;AACpH,UAAM,SAAS,CAAC,YAAY,CAAC,aAAa,CAAC,WAAW,SAAS,MAAM,IAAI,MAAM,IAAI,SAAS;AAC5F,QAAI,OAAQ,YAAW;AAEvB,UAAM,QAAQ,KAAK,IAAI,IAAI,KAAM,SAAS,YAAY,CAAC,aAAa,CAAC,UAAW,WAAW,CAAC;AAC5F,UAAM,IAAU,EAAE,MAAM,MAAM,QAAQ,KAAK;AAC3C,UAAM,QAAQ,aAAa,KAAK,IAAI,IAAI,CAAC;AACzC,QAAI,SAAS,CAAC,QAAS,GAAE,UAAU,EAAE,OAAO,OAAO,KAAK,GAAG,IAAI,MAAM,IAAI,IAAI,KAAK,QAAW,MAAM,CAAC,EAAE;AACtG,QAAI,SAAS,OAAO,YAAY,MAAM,IAAI,MAAM,KAAK,IAAI,WAAW,GAAG,EAAG,GAAE,YAAY,EAAE,SAAS,QAAQ,MAAM,MAAM,QAAQ,KAAK,OAAO;AAC3I,QAAI,MAAM,IAAI,eAAe,EAAG,GAAE,UAAU;AAC5C,QAAI,SAAS,QAAQ,aAAc,GAAE,UAAU,CAAC;AAChD,QAAI,MAAM,KAAK,IAAI,EAAG,EAAC,UAAU,QAAQ,OAAO,MAAM,KAAK,GAAG;AAC9D,SAAK,CAAC;AAAA,EACR;AACA,SAAO,MAAM,OAAQ,KAAI;AAEzB,QAAM,WAAW,OAAO,KAAK,KAAK,EAAE,CAAC;AACrC,QAAM,QAAQ,MAAM,OAAO,OAAO,QAAQ,CAAC;AAC3C,MAAI,CAAC,YAAY,CAAC,MAAO,QAAO;AAChC,QAAM,YAAY,MAAM,IAAI,UAAU,KAAK,IAAI,MAAM,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,OAAO,OAAO;AAC7F,SAAO;AAAA,IACL,KAAK,EAAE,OAAO,SAAS,KAAK,KAAK,SAAS,KAAK,WAAW,WAAW,GAAG,GAAG,SAAS,QAAQ,MAAM,IAAI,aAAa,KAAK,QAAQ,GAAG,SAAS,KAAK;AAAA,IACjJ,OAAO,EAAE,OAAO,SAAS,KAAK,UAAU,MAAM,UAAU,UAAU,CAAC,GAAG,UAAU,GAAG,IAAI,EAAE;AAAA,IACzF;AAAA,EACF;AACF;AAEA,SAAS,OAAO,GAAmB;AACjC,SAAO,EAAE,QAAQ,QAAQ,GAAG,EAAE,KAAK;AACrC;AAEA,SAAS,QAAQ,GAAmB;AAClC,MAAI,EAAE,UAAU,IAAK,QAAO;AAC5B,QAAM,MAAM,EAAE,MAAM,GAAG,GAAG;AAC1B,SAAO,IAAI,MAAM,GAAG,IAAI,YAAY,GAAG,IAAI,MAAM,IAAI,YAAY,GAAG,IAAI,GAAG,IAAI;AACjF;AAGO,SAAS,WAAW,KAAqB;AAC9C,QAAM,QAAQ,IAAI,MAAM,GAAG;AAC3B,MAAI,MAAM,SAAS,EAAG,QAAO;AAC7B,MAAI,IAAI,MAAM,CAAC;AACf,MAAI;AACF,QAAI,mBAAmB,CAAC;AAAA,EAC1B,QAAQ;AAAA,EAER;AACA,MAAI,EAAE,QAAQ,UAAU,GAAG;AAC3B,SAAO,EAAE,OAAO,CAAC,EAAE,YAAY,IAAI,EAAE,MAAM,CAAC;AAC9C;AAGA,MAAM,QAAgC;AAAA,EACpC,MAAM;AAAA,EAAI,KAAK;AAAA,EAAI,MAAM;AAAA,EAAI,IAAI;AAAA,EAAI,IAAI;AAAA,EAAI,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,KAAK;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,KAAK;AAAA,EAAK,MAAM;AAAA,EAC1J,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,KAAK;AAAA,EAAK,KAAK;AAAA,EAAK,KAAK;AAAA,EAAK,MAAM;AAAA,EAAK,KAAK;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,QAAQ;AAAA,EACxJ,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAC3J,OAAO;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,KAAK;AAAA,EAAK,QAAQ;AAAA,EACzJ,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAC5J,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAC3J,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,KAAK;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAC3J,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,QAAQ;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,OAAO;AAAA,EAAK,QAAQ;AAAA,EAAK,QAAQ;AAAA,EAAK,MAAM;AAAA,EAAK,MAAM;AAAA,EACxJ,MAAM;AAAA,EAAK,OAAO;AAAA,EAAK,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,QAAQ;AAAA,EAAM,MAAM;AAAA,EAAM,KAAK;AAAA,EAAM,KAAK;AAAA,EAAM,KAAK;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EACxJ,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,QAAQ;AAAA,EAAM,QAAQ;AAAA,EAAM,MAAM;AAAA,EAAM,QAAQ;AAAA,EAAM,QAAQ;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,QAAQ;AAAA,EAC1J,QAAQ;AAAA,EAAM,MAAM;AAAA,EAAM,OAAO;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,MAAM;AAAA,EAAM,OAAO;AAAA,EAAM,IAAI;AAAA,EACxJ,IAAI;AAAA,EAAM,IAAI;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAM,OAAO;AAAA,EAAO,QAAQ;AACtE;AAEA,SAAS,OAAO,GAAmB;AACjC,MAAI,CAAC,EAAE,SAAS,GAAG,EAAG,QAAO;AAC7B,SAAO,EAAE,QAAQ,yCAAyC,CAAC,GAAG,MAAc;AAC1E,QAAI,EAAE,CAAC,MAAM,KAAK;AAChB,YAAMA,KAAI,EAAE,CAAC,MAAM,OAAO,EAAE,CAAC,MAAM,MAAM,SAAS,EAAE,MAAM,CAAC,GAAG,EAAE,IAAI,OAAO,EAAE,MAAM,CAAC,CAAC;AACrF,aAAO,OAAO,SAASA,EAAC,KAAKA,KAAI,KAAKA,KAAI,UAAW,OAAO,cAAcA,EAAC,IAAI;AAAA,IACjF;AACA,UAAM,IAAI,MAAM,CAAC,KAAK,MAAM,EAAE,YAAY,CAAC;AAC3C,WAAO,MAAM,SAAY,IAAI,OAAO,cAAc,CAAC;AAAA,EACrD,CAAC;AACH;","names":["n"]}

package/dist/index.d.ts

@@ -0,0 +1,94 @@
+import { TreeNode } from '@dogsbay/types';
+
+/**
+ * The indexer: documents in, index files out (see format.ts). Pure and
+ * Worker-safe: no Node built-ins, and memory for one page at a time plus
+ * the postings.
+ */
+
+interface SearchDoc {
+    title: string;
+    url: string;
+    section: string;
+    excerpt: string;
+    /** h2/h3 headings with their anchors, for results within the page. */
+    anchors?: [text: string, id: string][];
+    /** Tags, for filters (`tag:x`). */
+    tags?: string[];
+}
+/**
+ * Ranking: each field is scored on its own and the fields are combined
+ * afterwards, BM25-style: term frequency saturates (k1), and a field's
+ * frequency is divided by its length relative to the corpus average (b),
+ * so a long page that mentions a word fifty times does not outrank a page
+ * with the word in its title. The inverse document frequency is applied at
+ * query time from the length of the term's posting list.
+ */
+declare const RANK: {
+    k1: number;
+    fields: {
+        title: {
+            weight: number;
+            b: number;
+        };
+        keywords: {
+            weight: number;
+            b: number;
+        };
+        headings: {
+            weight: number;
+            b: number;
+        };
+        body: {
+            weight: number;
+            b: number;
+        };
+    };
+};
+declare class SearchIndexBuilder {
+    private opts;
+    constructor(opts?: {
+        lang?: string;
+    });
+    private docs;
+    /** Token counts per document and field (title, headings, body), for length normalisation. */
+    private lengths;
+    /**
+     * term → flat [doc, counts, doc, counts, …]. Documents arrive in order, so
+     * each page's counts are summed locally and appended once. Flat arrays
+     * cost a fraction of a Map per term, which matters inside a Worker.
+     */
+    private postings;
+    add(doc: SearchDoc, parts: {
+        title: string;
+        headings: string[];
+        body: string;
+        keywords?: string[];
+    }): void;
+    /** The stored score of one posting: saturated, length-normalised field scores, weighted and summed. */
+    private scorer;
+    get size(): number;
+    /**
+     * Terms grouped into shards: by first character, and a group larger than
+     * SHARD_CHARS split on the next character (up to MAX_DEPTH), so a query
+     * loads small files on large sites. Terms no longer than a group's key stay
+     * in the group's own file.
+     */
+    private shardTerms;
+    /** Emit the index files, keyed by path relative to the unit's `_dogsbay/search/` dir. */
+    emit(): Map<string, string>;
+    /**
+     * The index files one at a time: the document table, the fragments, then
+     * one shard per term prefix. Each shard is built, serialised and handed
+     * over before the next, and its postings are dropped, so writing the index
+     * needs memory for one shard rather than for every shard and its JSON.
+     */
+    emitFiles(): Generator<[string, string]>;
+}
+/** Text for indexing: headings separated from body, code blocks skipped. */
+declare function extractText(tree: TreeNode[]): {
+    headings: string[];
+    body: string;
+};
+
+export { RANK, type SearchDoc, SearchIndexBuilder, extractText };