ink-note-app/electron/importer.cjs

'use strict';
// Bringing notes in from another app: an Obsidian vault or a Notion "Markdown & CSV" export, as a folder or a .zip.
// Everything goes into one new top-level folder of the notes, so nothing existing is touched.
//   - Notion's 32-character ids are taken out of file and folder names, and its relative Markdown links
//     ([Title](Folder/Page%20abc….md)) become [[wiki links]] to the imported pages.
//   - Obsidian's .obsidian folder and Notion's _all.csv files are left out; other files (pictures, PDFs) are copied.
const fs = require('node:fs');
const fsp = fs.promises;
const path = require('node:path');
const zlib = require('node:zlib');

const SKIP_DIR = /^(\.obsidian|\.trash|\.git|\.notes|__MACOSX|node_modules)$/;
const MAX_FILES = 50000;
const MAX_FILE = 200 * 1024 * 1024;

/* zip reading: central directory, stored and deflated entries */
function readZip(file) {
  const buf = fs.readFileSync(file);
  let end = -1;
  for (let i = buf.length - 22; i >= Math.max(0, buf.length - 65557); i--) if (buf.readUInt32LE(i) === 0x06054b50) { end = i; break; }
  if (end < 0) throw new Error('That file is not a zip archive.');
  const count = buf.readUInt16LE(end + 10);
  let at = buf.readUInt32LE(end + 16);
  const out = [];
  for (let n = 0; n < count && n < MAX_FILES; n++) {
    if (buf.readUInt32LE(at) !== 0x02014b50) break;
    const method = buf.readUInt16LE(at + 10);
    const csize = buf.readUInt32LE(at + 20);
    const size = buf.readUInt32LE(at + 24);
    const nameLen = buf.readUInt16LE(at + 28), extraLen = buf.readUInt16LE(at + 30), commentLen = buf.readUInt16LE(at + 32);
    const local = buf.readUInt32LE(at + 42);
    const name = buf.toString('utf8', at + 46, at + 46 + nameLen);
    at += 46 + nameLen + extraLen + commentLen;
    if (name.endsWith('/') || size > MAX_FILE) continue;
    const dataAt = local + 30 + buf.readUInt16LE(local + 26) + buf.readUInt16LE(local + 28);
    out.push({
      rel: name,
      read: () => {
        const raw = buf.subarray(dataAt, dataAt + csize);
        return method === 0 ? raw : zlib.inflateRawSync(raw);
      }
    });
  }
  return out;
}

async function* walkDir(root, rel = '') {
  let entries;
  try { entries = await fsp.readdir(path.join(root, rel), { withFileTypes: true }); } catch { return; }
  for (const e of entries) {
    if (e.isSymbolicLink()) continue;
    const r = rel ? `${rel}/${e.name}` : e.name;
    if (e.isDirectory()) { if (!SKIP_DIR.test(e.name)) yield* walkDir(root, r); }
    else if (e.isFile()) yield r;
  }
}

// "Page 0a1b2c3d4e5f60718293a4b5c6d7e8f9.md" -> "Page.md"; a plain name is left alone.
const NOTION_ID = /\s+[0-9a-f]{32}(?=\.[^.]+$|$)/i;
const stripId = (seg) => seg.replace(NOTION_ID, '');
const badChars = /[<>:"|?*\u0000-\u001f]/g;
const safeSeg = (s) => s.replace(badChars, '-').replace(/^\.+/, '').trim() || 'Untitled';

// Where each source path ends up: { 'a/b abc.md': 'a/b.md' }, names made unique.
function plan(files) {
  const map = new Map();
  const taken = new Set();
  for (const rel of files.slice().sort()) {
    const parts = rel.split('/').map((p) => safeSeg(stripId(p)));
    let target = parts.join('/');
    const ext = path.posix.extname(target);
    const stem = target.slice(0, target.length - ext.length);
    for (let n = 2; taken.has(target.toLowerCase()); n++) target = `${stem} ${n}${ext}`;
    taken.add(target.toLowerCase());
    map.set(rel, target);
  }
  return map;
}

// Rewrite relative links in one note: to other notes -> [[wiki links]], to files -> the new relative path.
function rewriteBody(body, fromRel, map, rootName) {
  const dir = path.posix.dirname(fromRel);
  const resolve = (href) => {
    let h = href.split('#')[0];
    try { h = decodeURIComponent(h); } catch { /* leave as written */ }
    if (!h || /^[a-z][a-z0-9+.-]*:/i.test(h) || h.startsWith('/') || h.startsWith('#')) return null;
    const joined = path.posix.normalize(dir === '.' ? h : `${dir}/${h}`);
    return joined.startsWith('..') ? null : joined;
  };
  return body.replace(/(!?)\[([^\]\n]*)\]\(([^)\n]+)\)/g, (all, bang, text, href) => {
    const target = resolve(href.trim().replace(/^<|>$/g, ''));
    if (!target || !map.has(target)) return all;
    const to = map.get(target);
    if (!bang && /\.md$/i.test(to)) {
      const id = `${rootName}/${to.replace(/\.md$/i, '')}`;
      const title = id.split('/').pop();
      return text && text !== title ? `[[${id}|${text}]]` : `[[${id}]]`;
    }
    // a picture or file: its path from this note's new folder
    const fromDir = path.posix.dirname(map.get(fromRel));
    const rel = path.posix.relative(fromDir === '.' ? '' : fromDir, to);
    return `${bang}[${text}](${encodeURI(rel).replace(/\(/g, '%28').replace(/\)/g, '%29')})`;
  });
}

// files: [{ rel, read() -> Buffer }]. Writes them under <vaultRoot>/<name>. Returns { folder, notes, files }.
async function importFiles(vaultRoot, name, files) {
  const kept = files.filter((f) => !f.rel.split('/').some((p) => SKIP_DIR.test(p)) && !/_all\.csv$/i.test(f.rel) && !/(^|\/)\.DS_Store$/.test(f.rel));
  // a zip often has one folder around everything: look inside it
  const first = new Set(kept.map((f) => f.rel.split('/')[0]));
  const strip = first.size === 1 && kept.every((f) => f.rel.includes('/')) ? [...first][0] + '/' : '';
  const items = kept.map((f) => ({ ...f, rel: f.rel.slice(strip.length) })).filter((f) => f.rel);
  if (!items.length) throw new Error('There was nothing to import in that.');
  const map = plan(items.map((f) => f.rel));

  let folder = safeSeg(name);
  for (let n = 2; fs.existsSync(path.join(vaultRoot, folder)) || fs.existsSync(path.join(vaultRoot, folder + '.md')); n++) folder = `${safeSeg(name)} ${n}`;

  let notes = 0, others = 0;
  for (const f of items) {
    const to = path.join(vaultRoot, folder, ...map.get(f.rel).split('/'));
    await fsp.mkdir(path.dirname(to), { recursive: true });
    let data = f.read();
    if (/\.md$/i.test(f.rel)) {
      data = Buffer.from(rewriteBody(data.toString('utf8'), f.rel, map, folder), 'utf8');
      notes++;
    } else others++;
    await fsp.writeFile(to, data);
  }
  return { folder, notes, files: others };
}

async function importFolder(vaultRoot, dir) {
  const files = [];
  for await (const rel of walkDir(dir)) {
    if (files.length >= MAX_FILES) break;
    const full = path.join(dir, rel);
    files.push({ rel, read: () => fs.readFileSync(full) });
  }
  return importFiles(vaultRoot, path.basename(dir), files);
}

async function importZip(vaultRoot, file) {
  const entries = readZip(file).filter((e) => !e.rel.includes('..'));
  return importFiles(vaultRoot, path.basename(file).replace(/\.zip$/i, ''), entries);
}

module.exports = { importFolder, importZip, importFiles, readZip, stripId, plan, rewriteBody };