rootnode-academy/platform/api/migrate-sqlite.js

#!/usr/bin/env node
// SQLite -> Postgres migration — the one-time cutover for Phase 2B. Run once against the real
// data volume before treating Postgres as authoritative. See
// docs/PHASE2B_ADMIN_AND_PLATFORM_DATA.md "Migration process / rollback" for the full
// procedure and how to roll back.
//
// Usage:
//   SQLITE_PATH=/path/to/progress.db PLATFORM_DB_URL=postgres://... node migrate-sqlite.js
//
// Guarantees:
//   - Backs up the SQLite file before touching anything (never modifies or deletes the source).
//   - Idempotent: safe to re-run. Every write uses ON CONFLICT — re-running re-syncs from the
//     SQLite source, it does not duplicate rows or fail on a second run.
//   - Never silently drops data: prints (and writes to a report file) exactly what was read,
//     what was written, and anything that couldn't be migrated and why.
//   - Preserves the old students.cohort column as a real cohort + enrollment, not just a
//     dropped string.

const fs = require('fs');
const path = require('path');
const Database = require('better-sqlite3');
const { getPool, runMigrations } = require('./src/db');

const SQLITE_PATH = process.env.SQLITE_PATH || '/data/progress.db';

async function main() {
  const report = { students: 0, cohorts_created: 0, enrollments: 0, progress_rows: 0, exam_sessions: 0, skipped: [] };

  if (!fs.existsSync(SQLITE_PATH)) {
    console.error(`SQLite file not found at ${SQLITE_PATH} — nothing to migrate. Set SQLITE_PATH if it's elsewhere.`);
    process.exit(1);
  }

  // 1. Backup — never touch the only copy of pre-migration data.
  const backupPath = `${SQLITE_PATH}.backup-${new Date().toISOString().replace(/[:.]/g, '-')}`;
  fs.copyFileSync(SQLITE_PATH, backupPath);
  console.log(`Backed up ${SQLITE_PATH} -> ${backupPath}`);

  const sqlite = new Database(SQLITE_PATH, { readonly: true });
  await runMigrations();
  const pool = getPool();

  // 2. Students + cohorts + enrollments
  const studentRows = safeAll(sqlite, `SELECT * FROM students`, report, 'students');
  const cohortsSeen = new Set();
  for (const s of studentRows) {
    await pool.query(
      `INSERT INTO students (student_id, display_name, created_at, last_active_at)
       VALUES ($1, $2, $3, $4)
       ON CONFLICT (student_id) DO UPDATE SET display_name = excluded.display_name, last_active_at = excluded.last_active_at`,
      [s.student_id, s.display_name, s.created_at || new Date().toISOString(), s.last_active_at || null]
    );
    report.students++;

    if (s.cohort && !cohortsSeen.has(s.cohort)) {
      cohortsSeen.add(s.cohort);
      const { rowCount } = await pool.query(
        `INSERT INTO cohorts (cohort_id, name) VALUES ($1, $1) ON CONFLICT (cohort_id) DO NOTHING`,
        [s.cohort]
      );
      if (rowCount) report.cohorts_created++;
    }
    if (s.cohort) {
      await pool.query(
        `INSERT INTO cohort_enrollments (cohort_id, student_id) VALUES ($1, $2) ON CONFLICT DO NOTHING`,
        [s.cohort, s.student_id]
      );
      report.enrollments++;
    }
  }

  // 3. Progress
  const progressRows = safeAll(sqlite, `SELECT * FROM progress`, report, 'progress');
  for (const p of progressRows) {
    await pool.query(
      `INSERT INTO progress (student_id, scenario_id, status, attempts, last_validated, completed_at, started_at, notes, time_spent_seconds)
       VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9)
       ON CONFLICT (student_id, scenario_id) DO UPDATE SET
         status = excluded.status, attempts = excluded.attempts, last_validated = excluded.last_validated,
         completed_at = excluded.completed_at, started_at = excluded.started_at, notes = excluded.notes,
         time_spent_seconds = excluded.time_spent_seconds`,
      [p.student_id, p.scenario_id, p.status, p.attempts || 0, p.last_validated, p.completed_at, p.started_at, p.notes, p.time_spent_seconds || 0]
    );
    report.progress_rows++;
  }

  // 4. Exam sessions (historical import only — see migrations/006_exam_sessions.sql note)
  const sessionRows = safeAll(sqlite, `SELECT * FROM sessions`, report, 'sessions');
  for (const s of sessionRows) {
    await pool.query(
      `INSERT INTO exam_sessions (session_id, student_id, bundle_id, started_at, submitted_at, status, exam_minutes, duration_secs, snapshot, scenario_ids)
       VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10)
       ON CONFLICT (session_id) DO UPDATE SET status = excluded.status, submitted_at = excluded.submitted_at`,
      [s.id, s.student_id || null, s.bundle_id, s.started_at, s.submitted_at, s.status, s.exam_minutes, s.duration_secs,
        s.snapshot ? JSON.parse(s.snapshot) : null, s.scenario_ids ? JSON.parse(s.scenario_ids) : null]
    );
    report.exam_sessions++;
  }

  sqlite.close();

  const reportPath = path.join(path.dirname(SQLITE_PATH), `migration-report-${Date.now()}.json`);
  fs.writeFileSync(reportPath, JSON.stringify(report, null, 2));
  console.log('\nMigration report:');
  console.log(JSON.stringify(report, null, 2));
  console.log(`\nFull report written to ${reportPath}`);
  console.log(`Backup preserved at ${backupPath} — not deleted, not modified.`);
  await pool.end();
}

function safeAll(sqlite, sql, report, label) {
  try {
    return sqlite.prepare(sql).all();
  } catch (e) {
    report.skipped.push({ table: label, reason: e.message });
    console.warn(`Could not read ${label}: ${e.message} (table may not exist in this SQLite file — recorded in report, not fatal)`);
    return [];
  }
}

main().catch((e) => { console.error('Migration failed:', e); process.exit(1); });