Xu_BrowserBookmark/apps/server/src/lib/bookmarkHtmlNode.js

import * as cheerio from "cheerio";

export function parseNetscapeBookmarkHtmlNode(html) {
  const $ = cheerio.load(html, { decodeEntities: false });
  const rootDl = $("dl").first();
  if (!rootDl.length) return { folders: [], bookmarks: [] };

  const folders = [];
  const bookmarks = [];

  function normText(s) {
    return String(s || "").replace(/\s+/g, " ").trim();
  }

  function collectLevelDt(node) {
    const out = [];
    const children = $(node).contents().toArray();
    for (const child of children) {
      if (!child || child.type !== "tag") continue;
      const tag = child.tagName?.toLowerCase();
      if (tag === "dt") {
        out.push(child);
        continue;
      }
      if (tag === "dl") {
        // nested list belongs to the previous <DT>
        continue;
      }
      out.push(...collectLevelDt(child));
    }
    return out;
  }

  function findNextDlForDt(dtNode, stopDlNode) {
    let cur = dtNode;
    while (cur && cur !== stopDlNode) {
      let next = cur.nextSibling;
      while (next && next.type !== "tag") next = next.nextSibling;
      if (next && next.type === "tag" && next.tagName?.toLowerCase() === "dl") return $(next);
      cur = cur.parent;
    }
    return null;
  }

  function walkDl($dl, parentTempId) {
    // Netscape format: <DL><p> contains repeating <DT> items and nested <DL>.
    // When parsed, <DT> may be wrapped (e.g. inside <p>), so we must be robust.
    const dts = collectLevelDt($dl[0]);
    for (const node of dts) {
      const $dt = $(node);
      const $h3 = $dt.children("h3").first().length ? $dt.children("h3").first() : $dt.find("h3").first();
      const $a = $dt.children("a").first().length ? $dt.children("a").first() : $dt.find("a").first();
      const $nestedDl = $dt.children("dl").first();
      const $nextDl = $nestedDl.length ? $nestedDl : findNextDlForDt(node, $dl[0]);

      if ($h3.length) {
        const tempId = `${folders.length + 1}`;
        const name = normText($h3.text() || "");
        folders.push({ tempId, parentTempId: parentTempId ?? null, name });
        if ($nextDl?.length) walkDl($nextDl, tempId);
      } else if ($a.length) {
        const title = normText($a.text() || "");
        const url = $a.attr("href") || "";
        bookmarks.push({ parentTempId: parentTempId ?? null, title, url });
      }
    }
  }

  walkDl(rootDl, null);
  return { folders, bookmarks };
}

export function buildNetscapeBookmarkHtml({ folders, bookmarks }) {
  // folders: [{id, parentId, name}]
  // bookmarks: [{folderId, title, url}]
  const folderChildren = new Map();
  const bookmarkChildren = new Map();

  for (const f of folders) {
    const key = f.parentId ?? "root";
    if (!folderChildren.has(key)) folderChildren.set(key, []);
    folderChildren.get(key).push(f);
  }
  for (const b of bookmarks) {
    const key = b.folderId ?? "root";
    if (!bookmarkChildren.has(key)) bookmarkChildren.set(key, []);
    bookmarkChildren.get(key).push(b);
  }

  function esc(s) {
    return String(s)
      .replaceAll("&", "&amp;")
      .replaceAll("<", "&lt;")
      .replaceAll(">", "&gt;")
      .replaceAll('"', "&quot;");
  }

  function renderFolder(parentId) {
    const key = parentId ?? "root";
    const subFolders = (folderChildren.get(key) || []).slice().sort((a, b) => a.name.localeCompare(b.name));
    const subBookmarks = (bookmarkChildren.get(key) || []).slice().sort((a, b) => a.title.localeCompare(b.title));

    let out = "<DL><p>\n";

    for (const f of subFolders) {
      out += `    <DT><H3>${esc(f.name)}</H3>\n`;
      out += renderFolder(f.id)
        .split("\n")
        .map((line) => (line ? `    ${line}` : line))
        .join("\n");
      out += "\n";
    }

    for (const b of subBookmarks) {
      out += `    <DT><A HREF=\"${esc(b.url)}\">${esc(b.title)}</A>\n`;
    }

    out += "</DL><p>";
    return out;
  }

  const header = `<!DOCTYPE NETSCAPE-Bookmark-file-1>\n<!-- This is an automatically generated file. -->\n<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html; charset=UTF-8\">\n<TITLE>Bookmarks</TITLE>\n<H1>Bookmarks</H1>\n`;
  const body = renderFolder(null);
  return header + body + "\n";
}