Migrate to SQLite (#174)

* Migrating Fredy from LowDb to SqLite 🎉

* adding new sql migration system for future sql migrations

* adding setting to change  sqlite path for db files

* create migration plan for graceful migration lowdb -> sqlite

* Improving Documentation

* adding test for sqliteconnection

* upgrading dependencies

* making nodejs 22 as min version

* improve scraper

* adding overwrite ability for db migra

test/db/migrations/migrate.test.js

@@ -0,0 +1,329 @@
+import { expect } from 'chai';
+import esmock from 'esmock';
+
+// We will fully mock fs, crypto, SqliteConnection, and dynamic import of migration modules
+
+describe('db/migrations/migrate.js - runMigrations', () => {
+  let calls;
+  let runMigrations;
+  let prevExitCode;
+
+  beforeEach(async () => {
+    calls = {
+      fs: { existsSync: [], mkdirSync: [], readdirSync: [], readFileSync: [] },
+      sql: {
+        getConnection: 0,
+        tableExists: false,
+        query: [],
+        execute: [],
+        withTransaction: [],
+        optimize: 0,
+      },
+      logs: { info: [], warn: [], error: [] },
+    };
+
+    // Mock fs to avoid touching disk
+    const fsMock = {
+      existsSync: (p) => {
+        calls.fs.existsSync.push(p);
+        return true;
+      },
+      mkdirSync: (p, opts) => {
+        calls.fs.mkdirSync.push({ p, opts });
+      },
+      readdirSync: (p) => {
+        calls.fs.readdirSync.push(p);
+        return [];
+      },
+      readFileSync: (p) => {
+        calls.fs.readFileSync.push(p);
+        return Buffer.from('dummy');
+      },
+    };
+
+    // Mock crypto sha256
+    const cryptoMock = {
+      createHash: () => ({ update: () => ({ digest: () => 'sha256sum' }) }),
+    };
+
+    // Mock logger
+    const loggerMock = {
+      info: (...a) => calls.logs.info.push(a),
+      warn: (...a) => calls.logs.warn.push(a),
+      error: (...a) => calls.logs.error.push(a),
+    };
+
+    // Mock SqliteConnection
+    const sqlMock = {
+      getConnection: () => {
+        calls.sql.getConnection += 1;
+        return {};
+      },
+      tableExists: () => calls.sql.tableExists,
+      query: (sql) => {
+        calls.sql.query.push(sql);
+        return [];
+      },
+      execute: (sql, params) => {
+        calls.sql.execute.push({ sql, params });
+        return { changes: 1 };
+      },
+      withTransaction: (cb) => {
+        calls.sql.withTransaction.push(true);
+        const db = {
+          prepare: (s) => ({ run: (...args) => calls.sql.execute.push({ sql: s, params: args }) }),
+        };
+        return cb(db);
+      },
+      optimize: () => {
+        calls.sql.optimize += 1;
+      },
+    };
+
+    // esmock with dependency replacements
+    const path = await import('node:path');
+    const ROOT = path.resolve('.');
+    const sqlPath = path.join(ROOT, 'lib', 'services', 'storage', 'SqliteConnection.js');
+    const loggerPath = path.join(ROOT, 'lib', 'services', 'logger.js');
+    const mod = await esmock(
+      '../../../db/migrations/migrate.js',
+      {},
+      {
+        fs: fsMock,
+        crypto: cryptoMock,
+        [sqlPath]: sqlMock,
+        [loggerPath]: loggerMock,
+      },
+    );
+
+    runMigrations = mod.runMigrations;
+
+    // remember original exitCode to restore later
+    prevExitCode = process.exitCode;
+  });
+
+  afterEach(() => {
+    // restore original process.exitCode
+    process.exitCode = prevExitCode;
+  });
+
+  it('logs and returns when no migration files are found', async () => {
+    await runMigrations();
+    expect(calls.logs.info.some((a) => String(a[0]).includes('No migration files'))).to.equal(true);
+    expect(calls.sql.getConnection).to.equal(0);
+    expect(calls.sql.optimize).to.equal(0);
+  });
+
+  it('applies a single new migration inside a transaction and records it', async () => {
+    // Re-mock with one file and module loader
+    const fsMock = {
+      existsSync: () => true,
+      mkdirSync: () => {},
+      readdirSync: () => ['1.init.js'],
+      readFileSync: () => Buffer.from('dummy'),
+    };
+    const cryptoMock = { createHash: () => ({ update: () => ({ digest: () => 'abc' }) }) };
+    const loggerMock = {
+      info: (...a) => calls.logs.info.push(a),
+      warn: (...a) => calls.logs.warn.push(a),
+      error: (...a) => calls.logs.error.push(a),
+    };
+
+    const sqlMock = {
+      getConnection: () => {
+        calls.sql.getConnection += 1;
+        return {};
+      },
+      tableExists: () => false, // schema_migrations not present yet
+      query: () => [],
+      execute: (sql, params) => {
+        calls.sql.execute.push({ sql, params });
+        return { changes: 1 };
+      },
+      withTransaction: (cb) => {
+        calls.sql.withTransaction.push(true);
+        const db = {
+          exec: () => {},
+          prepare: (s) => ({ run: (...args) => calls.sql.execute.push({ sql: s, params: args }) }),
+        };
+        return cb(db);
+      },
+      optimize: () => {
+        calls.sql.optimize += 1;
+      },
+    };
+
+    // The migration module: exports up(db)
+    const migrationModule = {
+      up: (db) => {
+        db.exec && db.exec('CREATE TABLE schema_migrations(name TEXT)');
+      },
+    };
+
+    // We need to intercept dynamic import by esmock: provide a stub for import(url)
+    // esmock supports mocking via a virtual module using URL matching, but simpler approach:
+    // place the file path that migrate.js will compute and make Node import resolve to our stub
+    // We simulate by mocking url.pathToFileURL is still used, but dynamic import will be handled by esmock when we map the computed path.
+
+    const path = await import('node:path');
+    const ROOT = path.resolve('.');
+
+    const sqlPath = path.join(ROOT, 'lib', 'services', 'storage', 'SqliteConnection.js');
+    const loggerPath = path.join(ROOT, 'lib', 'services', 'logger.js');
+    // Use global importer hook to bypass dynamic import
+    globalThis.__TEST_MIGRATE_IMPORT__ = async () => migrationModule;
+
+    const mod = await esmock(
+      '../../../db/migrations/migrate.js',
+      {},
+      {
+        fs: fsMock,
+        crypto: cryptoMock,
+        [sqlPath]: sqlMock,
+        [loggerPath]: loggerMock,
+      },
+    );
+
+    runMigrations = mod.runMigrations;
+
+    await runMigrations();
+
+    // Should have started a transaction and inserted into schema_migrations
+    expect(calls.sql.withTransaction.length).to.equal(1);
+    const inserted = calls.sql.execute.find((e) => String(e.sql).includes('INSERT INTO schema_migrations'));
+    expect(!!inserted).to.equal(true);
+    expect(calls.sql.optimize).to.equal(1);
+  });
+
+  it('skips already executed migration with same checksum', async () => {
+    const fsMock = {
+      existsSync: () => true,
+      mkdirSync: () => {},
+      readdirSync: () => ['1.init.js'],
+      readFileSync: () => Buffer.from('dummy'),
+    };
+    const cryptoMock = { createHash: () => ({ update: () => ({ digest: () => 'same' }) }) };
+    const loggerMock = {
+      info: (...a) => calls.logs.info.push(a),
+      warn: (...a) => calls.logs.warn.push(a),
+      error: (...a) => calls.logs.error.push(a),
+    };
+
+    const sqlMock = {
+      getConnection: () => {
+        calls.sql.getConnection += 1;
+        return {};
+      },
+      tableExists: () => true,
+      query: () => [{ name: '1.init.js', checksum: 'same' }],
+      execute: (sql, params) => {
+        calls.sql.execute.push({ sql, params });
+        return { changes: 1 };
+      },
+      withTransaction: (cb) => {
+        calls.sql.withTransaction.push(true);
+        const db = { prepare: (s) => ({ run: (...args) => calls.sql.execute.push({ sql: s, params: args }) }) };
+        return cb(db);
+      },
+      optimize: () => {
+        calls.sql.optimize += 1;
+      },
+    };
+
+    const path = await import('node:path');
+    const ROOT = path.resolve('.');
+    const sqlPath = path.join(ROOT, 'lib', 'services', 'storage', 'SqliteConnection.js');
+    const loggerPath = path.join(ROOT, 'lib', 'services', 'logger.js');
+
+    globalThis.__TEST_MIGRATE_IMPORT__ = async () => ({ up: () => {} });
+
+    const mod = await esmock(
+      '../../../db/migrations/migrate.js',
+      {},
+      {
+        fs: fsMock,
+        crypto: cryptoMock,
+        [sqlPath]: sqlMock,
+        [loggerPath]: loggerMock,
+      },
+    );
+
+    runMigrations = mod.runMigrations;
+
+    await runMigrations();
+
+    // Should not run transaction because it's skipped
+    expect(calls.sql.withTransaction.length).to.equal(0);
+    expect(calls.sql.optimize).to.equal(1);
+  });
+
+  it('aborts with exitCode=1 when a migration throws, without applying insert', async () => {
+    const fsMock = {
+      existsSync: () => true,
+      mkdirSync: () => {},
+      readdirSync: () => ['1.bad.js'],
+      readFileSync: () => Buffer.from('dummy'),
+    };
+    const cryptoMock = { createHash: () => ({ update: () => ({ digest: () => 'bad' }) }) };
+    const loggerMock = {
+      info: (...a) => calls.logs.info.push(a),
+      warn: (...a) => calls.logs.warn.push(a),
+      error: (...a) => calls.logs.error.push(a),
+    };
+
+    const sqlMock = {
+      getConnection: () => {
+        calls.sql.getConnection += 1;
+        return {};
+      },
+      tableExists: () => false,
+      query: () => [],
+      execute: (sql, params) => {
+        calls.sql.execute.push({ sql, params });
+        return { changes: 1 };
+      },
+      withTransaction: (cb) => {
+        calls.sql.withTransaction.push(true);
+        const db = {
+          exec: () => {},
+          prepare: (s) => ({ run: (...args) => calls.sql.execute.push({ sql: s, params: args }) }),
+        };
+        return cb(db);
+      },
+      optimize: () => {
+        calls.sql.optimize += 1;
+      },
+    };
+
+    const path = await import('node:path');
+    const ROOT = path.resolve('.');
+
+    globalThis.__TEST_MIGRATE_IMPORT__ = async () => ({
+      up: () => {
+        throw new Error('boom');
+      },
+    });
+    const sqlPath = path.join(ROOT, 'lib', 'services', 'storage', 'SqliteConnection.js');
+    const loggerPath = path.join(ROOT, 'lib', 'services', 'logger.js');
+
+    const mod = await esmock(
+      '../../../db/migrations/migrate.js',
+      {},
+      {
+        fs: fsMock,
+        crypto: cryptoMock,
+        [sqlPath]: sqlMock,
+        [loggerPath]: loggerMock,
+      },
+    );
+
+    runMigrations = mod.runMigrations;
+
+    await runMigrations();
+
+    expect(process.exitCode).to.equal(1);
+    // No insert into schema_migrations should be recorded since transaction failed
+    const inserted = calls.sql.execute.find((e) => String(e.sql).includes('INSERT INTO schema_migrations'));
+    expect(inserted).to.equal(undefined);
+  });
+});

test/mocks/mockStore.js

@@ -1,8 +1,8 @@
 const db = {};
-export const setKnownListings = (jobKey, providerId, listings) => {
+export const storeListings = (jobKey, providerId, listings) => {
   if (!Array.isArray(listings)) throw Error('Not a valid array');
   db[providerId] = listings;
 };
-export const getKnownListings = (jobKey, providerId) => {
+export const getKnownListingHashesForJobAndProvider = (jobKey, providerId) => {
   return db[providerId] || [];
 };

test/provider/einsAImmobilien.test.js

@@ -25,6 +25,7 @@ describe('#einsAImmobilien testsuite()', () => {
           expect(notify.size).to.be.a('string');
           expect(notify.title).to.be.a('string');
           expect(notify.link).to.be.a('string');
+          expect(notify.address).to.be.a('string');
           /** check the values if possible **/
           expect(notify.size).to.be.not.empty;
           expect(notify.title).to.be.not.empty;

test/storage/SqliteConnection.test.js

@@ -0,0 +1,142 @@
+import { expect } from 'chai';
+import esmock from 'esmock';
+
+// We explicitly avoid touching the real filesystem or creating a real DB file.
+// better-sqlite3 is fully mocked and operates in-memory via our stubs.
+
+describe('SqliteConnection', () => {
+  let SqliteConnection;
+  let calls;
+
+  beforeEach(async () => {
+    calls = {
+      fs: { existsSync: [], mkdirSync: [] },
+      db: { pragma: [], prepare: [], transactionWraps: 0, close: 0 },
+      prepareAll: [],
+      prepareRun: [],
+      prepareGet: [],
+      processOnce: [],
+      logs: { warn: [], debug: [] },
+    };
+
+    // stub for fs
+    const fsMock = {
+      existsSync: (dir) => {
+        calls.fs.existsSync.push(dir);
+        // Pretend directory always exists to avoid mkdir
+        return true;
+      },
+      mkdirSync: (dir, opts) => {
+        calls.fs.mkdirSync.push({ dir, opts });
+      },
+    };
+
+    // Prepare object returned from db.prepare()
+    const prepareObj = {
+      all: (params) => {
+        calls.prepareAll.push(params);
+        return [{ x: 1 }];
+      },
+      run: (params) => {
+        calls.prepareRun.push(params);
+        return { changes: 1 };
+      },
+      get: (param) => {
+        calls.prepareGet.push(param);
+        // return truthy by default
+        return { one: 1 };
+      },
+    };
+
+    // Database mock constructor
+    const BetterSqlite3Mock = function (filepath, options) {
+      // expose on instance
+      this.filepath = filepath;
+      this.options = options;
+      this.pragma = (p) => {
+        calls.db.pragma.push(p);
+        return undefined;
+      };
+      this.prepare = (sql) => {
+        calls.db.prepare.push(sql);
+        return prepareObj;
+      };
+      this.transaction = (fn) => {
+        // better-sqlite3 returns a function that executes inside a transaction
+        return (cb) => {
+          calls.db.transactionWraps += 1;
+          return fn(cb);
+        };
+      };
+      this.close = () => {
+        calls.db.close += 1;
+      };
+    };
+
+    // esmock the module with our stubs
+    SqliteConnection = await esmock(
+      '../../lib/services/storage/SqliteConnection.js',
+      {},
+      {
+        fs: fsMock,
+        'better-sqlite3': { default: BetterSqlite3Mock },
+      },
+    );
+  });
+
+  afterEach(() => {
+    // ensure we can close between tests
+    SqliteConnection.close();
+  });
+
+  it('creates singleton connection and applies PRAGMAs without touching disk', () => {
+    const db1 = SqliteConnection.getConnection();
+    const db2 = SqliteConnection.getConnection();
+
+    expect(db1).to.equal(db2);
+    // journal_mode, synchronous, cache_size, foreign_keys, optimize
+    expect(calls.db.pragma).to.deep.equal([
+      'journal_mode = WAL',
+      'synchronous = NORMAL',
+      'cache_size = -64000',
+      'foreign_keys = ON',
+      'optimize',
+    ]);
+    // mkdirSync should not be called because existsSync returned true
+    expect(calls.fs.mkdirSync).to.have.length(0);
+  });
+
+  it('executes query and execute helpers', () => {
+    const rows = SqliteConnection.query('SELECT 1', {});
+    expect(rows).to.be.an('array');
+    expect(rows[0]).to.deep.equal({ x: 1 });
+
+    const info = SqliteConnection.execute('UPDATE x SET y=1 WHERE id=@id', { id: 5 });
+    expect(info).to.have.property('changes', 1);
+  });
+
+  it('tableExists uses sqlite_master get()', () => {
+    const exists = SqliteConnection.tableExists('users');
+    expect(exists).to.equal(true);
+  });
+
+  it('withTransaction wraps callback', () => {
+    const result = SqliteConnection.withTransaction((db) => {
+      // ensure we can use the db to prepare
+      db.prepare('SELECT inside').all({});
+      return 42;
+    });
+    expect(result).to.equal(42);
+    expect(calls.db.prepare).to.include('SELECT inside');
+  });
+
+  it('optimize() delegates to PRAGMA optimize and close() calls it again then closes', () => {
+    SqliteConnection.optimize();
+    // It will use the existing connection and call pragma('optimize')
+    expect(calls.db.pragma).to.include('optimize');
+
+    SqliteConnection.close();
+    // close increments close counter
+    expect(calls.db.close).to.equal(1);
+  });
+});