clawdie-ai/scripts/backup.ts

/**
 * Clawdie backup — exports all critical state to a portable tarball.
 *
 * What's included:
 *   messages.db       — SQLite: all chats, messages, tasks, sessions
 *   memory_db.sql     — pg_dump: agent memory + embeddings
 *   skills_db.sql     — pg_dump: built-in knowledge (skip with --skip-skills)
 *   env               — .env config (API keys, db passwords, all settings)
 *   groups/           — registered group configurations
 *   mount-allowlist.json — mount security allowlist
 *   manifest.json     — backup metadata
 *
 * ZFS snapshots: taken automatically via hostd before export (skip with --no-snapshot).
 * Sanoid handles ongoing automated retention; this script does on-demand export.
 *
 * Usage:
 *   npx tsx scripts/backup.ts
 *   npx tsx scripts/backup.ts --skip-skills --output /mnt/backup
 *   npx tsx scripts/backup.ts --no-snapshot
 *
 * Output: ~/clawdie-backup-DD.mmm.YYYY-HHMM.tar.gz
 *
 * Requires: pg_dump (install postgresql17-client on host if missing)
 * Restore:  see docs/sessions/2026-03-16-backup-restore.md
 */
import { spawnSync } from 'child_process';
import fs from 'fs';
import os from 'os';
import path from 'path';

import {
  AGENT_NAME,
  MEMORY_DB_NAME,
  MEMORY_DB_URL,
  MOUNT_ALLOWLIST_PATH,
  SKILLS_DB_NAME,
  SKILLS_DB_URL,
  STORE_DIR,
  GROUPS_DIR,
  ZFS_PREFIX,
} from '../src/config.js';
import { formatDisplayDate, formatSnapshotStamp } from '../src/display-date.js';
import { hostd } from '../src/hostd/client.js';

// ── CLI args ──────────────────────────────────────────────────────────────────

const args = process.argv.slice(2);
const SKIP_SKILLS = args.includes('--skip-skills');
const NO_SNAPSHOT = args.includes('--no-snapshot');
const outputIdx = args.indexOf('--output');
const OUTPUT_DIR = outputIdx !== -1 ? args[outputIdx + 1] : os.homedir();

// ── Timestamp helpers ─────────────────────────────────────────────────────────

const now = new Date();
const STAMP = formatSnapshotStamp(now);
const BACKUP_NAME = `${AGENT_NAME}-backup-${STAMP}`;

// ── Helpers ───────────────────────────────────────────────────────────────────

function ok(msg: string): void { console.log(`  ✓  ${msg}`); }
function warn(msg: string): void { console.log(`  ⚠  ${msg}`); }
function fail(msg: string): never { console.error(`  ✗  ${msg}`); process.exit(1); }

function run(cmd: string, args: string[], env?: NodeJS.ProcessEnv): { ok: boolean; output: string } {
  const result = spawnSync(cmd, args, {
    encoding: 'utf-8',
    stdio: ['ignore', 'pipe', 'pipe'],
    env: { ...process.env, ...env },
  });
  const output = [result.stdout, result.stderr].filter(Boolean).join('\n').trim();
  return { ok: result.status === 0 && !result.error, output };
}

function commandExists(cmd: string): boolean {
  return run('which', [cmd]).ok;
}

// ── ZFS snapshots (via hostd) ──────────────────────────────────────────────────

async function takeZfsSnapshots(snapshotTag: string): Promise<void> {
  const datasets = [
    `zroot/${ZFS_PREFIX}/jails/${AGENT_NAME}-db`,
    `zroot/${ZFS_PREFIX}/jails/${AGENT_NAME}-git`,
    `zroot/${ZFS_PREFIX}/jails/${AGENT_NAME}-cms`,
  ];

  for (const dataset of datasets) {
    try {
      const res = await hostd('zfs-snapshot', { dataset, name: snapshotTag });
      if (res.ok) {
        ok(`ZFS snapshot: ${dataset}@${snapshotTag}`);
      } else {
        warn(`ZFS snapshot failed for ${dataset}: ${res.error ?? res.output}`);
      }
    } catch {
      warn(`hostd unreachable — ZFS snapshot skipped for ${dataset}`);
      break;
    }
  }
}

// ── pg_dump ───────────────────────────────────────────────────────────────────

function pgDump(url: string, dbName: string, destFile: string): boolean {
  if (!commandExists('pg_dump')) {
    warn(`pg_dump not found — skipping ${dbName} (install postgresql-client on host)`);
    return false;
  }

  const result = run('pg_dump', ['--dbname', url, '--no-password', '--clean', '--if-exists', '-f', destFile]);
  if (result.ok) {
    const sizeKb = Math.round(fs.statSync(destFile).size / 1024);
    ok(`pg_dump ${dbName} → ${path.basename(destFile)} (${sizeKb} KB)`);
    return true;
  } else {
    warn(`pg_dump ${dbName} failed: ${result.output.slice(0, 200)}`);
    return false;
  }
}

// ── Main ──────────────────────────────────────────────────────────────────────

async function main(): Promise<void> {
  console.log('');
  console.log(`Clawdie backup — ${formatDisplayDate(now)}`);
  console.log(`Output: ${OUTPUT_DIR}/${BACKUP_NAME}.tar.gz`);
  console.log('');

  // ── Staging dir ─────────────────────────────────────────────────────────────

  const tmpBase = path.join(process.cwd(), 'tmp', 'backup');
  fs.mkdirSync(tmpBase, { recursive: true });
  const stagingBase = fs.mkdtempSync(path.join(tmpBase, 'clawdie-backup-'));
  const stagingDir = path.join(stagingBase, BACKUP_NAME);
  fs.mkdirSync(stagingDir, { recursive: true });

  const manifest: Record<string, unknown> = {
    agent: AGENT_NAME,
    created_at: now.toISOString(),
    created_display: formatDisplayDate(now),
    stamp: STAMP,
    items: [] as string[],
  };

  const items = manifest.items as string[];

  // ── ZFS snapshots ────────────────────────────────────────────────────────────

  if (!NO_SNAPSHOT) {
    console.log('Taking ZFS snapshots...');
    // Use snapshot stamp format per AGENTS.md: DD.mmm.YYYY-HHMM
    const snapshotTag = `pre-backup-${STAMP}`;
    await takeZfsSnapshots(snapshotTag);
    console.log('');
  }

  // ── SQLite ────────────────────────────────────────────────────────────────────

  console.log('Exporting SQLite...');
  const sqliteSrc = path.join(STORE_DIR, 'messages.db');
  if (fs.existsSync(sqliteSrc)) {
    fs.copyFileSync(sqliteSrc, path.join(stagingDir, 'messages.db'));
    const sizeKb = Math.round(fs.statSync(sqliteSrc).size / 1024);
    ok(`messages.db (${sizeKb} KB)`);
    items.push('messages.db');
  } else {
    warn('messages.db not found — skipping');
  }
  console.log('');

  // ── PostgreSQL dumps ──────────────────────────────────────────────────────────

  console.log('Exporting PostgreSQL...');
  if (pgDump(MEMORY_DB_URL, MEMORY_DB_NAME, path.join(stagingDir, 'memory_db.sql'))) {
    items.push('memory_db.sql');
  }
  if (!SKIP_SKILLS) {
    if (pgDump(SKILLS_DB_URL, SKILLS_DB_NAME, path.join(stagingDir, 'skills_db.sql'))) {
      items.push('skills_db.sql');
    }
  } else {
    ok('skills_db skipped (--skip-skills)');
  }
  console.log('');

  // ── Config ────────────────────────────────────────────────────────────────────

  console.log('Exporting config...');
  const envSrc = path.join(process.cwd(), '.env');
  if (fs.existsSync(envSrc)) {
    fs.copyFileSync(envSrc, path.join(stagingDir, 'env'));
    ok('.env');
    items.push('env');
  } else {
    warn('.env not found');
  }

  const groupsSrc = GROUPS_DIR;
  if (fs.existsSync(groupsSrc)) {
    const dest = path.join(stagingDir, 'groups');
    fs.mkdirSync(dest);
    for (const f of fs.readdirSync(groupsSrc)) {
      fs.copyFileSync(path.join(groupsSrc, f), path.join(dest, f));
    }
    ok('groups/');
    items.push('groups/');
  }

  if (fs.existsSync(MOUNT_ALLOWLIST_PATH)) {
    fs.copyFileSync(MOUNT_ALLOWLIST_PATH, path.join(stagingDir, 'mount-allowlist.json'));
    ok('mount-allowlist.json');
    items.push('mount-allowlist.json');
  } else {
    ok('mount-allowlist.json not found — skipping');
  }
  console.log('');

  // ── Manifest ──────────────────────────────────────────────────────────────────

  fs.writeFileSync(path.join(stagingDir, 'manifest.json'), JSON.stringify(manifest, null, 2));

  // ── Tarball ───────────────────────────────────────────────────────────────────

  console.log('Creating archive...');
  if (!fs.existsSync(OUTPUT_DIR)) fs.mkdirSync(OUTPUT_DIR, { recursive: true });
  const tarPath = path.join(OUTPUT_DIR, `${BACKUP_NAME}.tar.gz`);
  const tarResult = run('tar', ['-czf', tarPath, '-C', stagingBase, BACKUP_NAME]);

  if (!tarResult.ok) fail(`tar failed: ${tarResult.output}`);

  const sizeMb = (fs.statSync(tarPath).size / 1024 / 1024).toFixed(2);
  ok(`${BACKUP_NAME}.tar.gz (${sizeMb} MB)`);

  // ── Cleanup ───────────────────────────────────────────────────────────────────

  fs.rmSync(stagingBase, { recursive: true, force: true });

  console.log('');
  console.log('─'.repeat(60));
  console.log(`  Backup complete: ${tarPath}`);
  console.log(`  Items: ${items.join(', ')}`);
  console.log(`  Restore: see docs/sessions/2026-03-16-backup-restore.md`);
  console.log('─'.repeat(60));
  console.log('');
}

main().catch((err) => {
  console.error('Backup failed:', err instanceof Error ? err.message : String(err));
  process.exit(1);
});