Model-config web apps: describe the app in configuration instead of writing code. Built on Polars and DuckDB.
test_dbbasic_channels.py · 18.2 KB · 588 lines
| 1 | #!/usr/bin/env python3 |
| 2 | """ |
| 3 | Comprehensive test suite for DBBasic Channels and Task Queue |
| 4 | Tests performance, concurrency, reliability, and compares with PostgreSQL benchmarks |
| 5 | """ |
| 6 | |
| 7 | import pytest |
| 8 | import asyncio |
| 9 | import time |
| 10 | import json |
| 11 | import uuid |
| 12 | from typing import List, Dict, Any |
| 13 | import duckdb |
| 14 | import statistics |
| 15 | from datetime import datetime, timedelta |
| 16 | |
| 17 | # Import our systems |
| 18 | from dbbasic_channels import DBBasicChannels, Channel |
| 19 | from dbbasic_task_queue import DBBasicTaskQueue, TaskWorker |
| 20 | |
| 21 | |
| 22 | class TestDBBasicChannels: |
| 23 | """Test suite for DBBasic Channels pub/sub system""" |
| 24 | |
| 25 | @pytest.fixture |
| 26 | async def channels(self): |
| 27 | """Create a test channels instance""" |
| 28 | channels = DBBasicChannels(db_path=":memory:") # In-memory for tests |
| 29 | yield channels |
| 30 | # Cleanup |
| 31 | channels.conn.close() |
| 32 | |
| 33 | @pytest.mark.asyncio |
| 34 | async def test_publish_single_message(self, channels): |
| 35 | """Test publishing a single message""" |
| 36 | message_id = await channels.publish( |
| 37 | Channel.ORDERS.value, |
| 38 | "new_order", |
| 39 | {"order_id": "TEST123", "total": 100.00} |
| 40 | ) |
| 41 | |
| 42 | assert message_id is not None |
| 43 | assert len(message_id) == 36 # UUID length |
| 44 | |
| 45 | # Verify in database |
| 46 | result = channels.conn.execute( |
| 47 | "SELECT * FROM messages WHERE id = ?", [message_id] |
| 48 | ).fetchone() |
| 49 | |
| 50 | assert result is not None |
| 51 | assert result[1] == Channel.ORDERS.value # channel |
| 52 | assert result[2] == "new_order" # message_type |
| 53 | |
| 54 | @pytest.mark.asyncio |
| 55 | async def test_subscribe_and_receive(self, channels): |
| 56 | """Test subscribing to a channel and receiving messages""" |
| 57 | received_messages = [] |
| 58 | |
| 59 | async def test_callback(message): |
| 60 | received_messages.append(message) |
| 61 | |
| 62 | # Subscribe to orders channel |
| 63 | subscriber_id = await channels.subscribe( |
| 64 | Channel.ORDERS.value, |
| 65 | test_callback, |
| 66 | "test_subscriber" |
| 67 | ) |
| 68 | |
| 69 | # Publish a message |
| 70 | await channels.publish( |
| 71 | Channel.ORDERS.value, |
| 72 | "test_message", |
| 73 | {"data": "test"} |
| 74 | ) |
| 75 | |
| 76 | # Give it time to process |
| 77 | await asyncio.sleep(0.1) |
| 78 | |
| 79 | assert len(received_messages) == 1 |
| 80 | assert received_messages[0]['channel'] == Channel.ORDERS.value |
| 81 | assert received_messages[0]['message_type'] == "test_message" |
| 82 | |
| 83 | @pytest.mark.asyncio |
| 84 | async def test_wildcard_subscription(self, channels): |
| 85 | """Test subscribing to all channels with wildcard""" |
| 86 | all_messages = [] |
| 87 | |
| 88 | async def monitor_callback(message): |
| 89 | all_messages.append(message) |
| 90 | |
| 91 | # Subscribe to all channels |
| 92 | await channels.subscribe("*", monitor_callback, "global_monitor") |
| 93 | |
| 94 | # Publish to different channels |
| 95 | await channels.publish(Channel.ORDERS.value, "order", {"id": 1}) |
| 96 | await channels.publish(Channel.SHIPPING.value, "ship", {"id": 2}) |
| 97 | await channels.publish(Channel.FRAUD.value, "check", {"id": 3}) |
| 98 | |
| 99 | await asyncio.sleep(0.1) |
| 100 | |
| 101 | # Should receive all messages |
| 102 | assert len(all_messages) == 3 |
| 103 | channels_received = [msg['channel'] for msg in all_messages] |
| 104 | assert Channel.ORDERS.value in channels_received |
| 105 | assert Channel.SHIPPING.value in channels_received |
| 106 | assert Channel.FRAUD.value in channels_received |
| 107 | |
| 108 | @pytest.mark.asyncio |
| 109 | async def test_channel_worker_processing(self, channels): |
| 110 | """Test worker watching and processing channel messages""" |
| 111 | processed = [] |
| 112 | |
| 113 | async def test_processor(message): |
| 114 | processed.append(message) |
| 115 | return {"status": "processed", "id": message['payload'].get('id')} |
| 116 | |
| 117 | # Start worker in background |
| 118 | worker_task = asyncio.create_task( |
| 119 | channels.watch_channel( |
| 120 | Channel.ORDERS.value, |
| 121 | "test_worker", |
| 122 | test_processor |
| 123 | ) |
| 124 | ) |
| 125 | |
| 126 | # Give worker time to start |
| 127 | await asyncio.sleep(0.1) |
| 128 | |
| 129 | # Publish messages |
| 130 | for i in range(5): |
| 131 | await channels.publish( |
| 132 | Channel.ORDERS.value, |
| 133 | "process_me", |
| 134 | {"id": i, "data": f"test_{i}"} |
| 135 | ) |
| 136 | |
| 137 | # Wait for processing |
| 138 | await asyncio.sleep(0.5) |
| 139 | |
| 140 | # Cancel worker |
| 141 | worker_task.cancel() |
| 142 | try: |
| 143 | await worker_task |
| 144 | except asyncio.CancelledError: |
| 145 | pass |
| 146 | |
| 147 | # Check all were processed |
| 148 | assert len(processed) == 5 |
| 149 | assert all(msg['message_type'] == "process_me" for msg in processed) |
| 150 | |
| 151 | @pytest.mark.asyncio |
| 152 | async def test_message_ttl_expiration(self, channels): |
| 153 | """Test that messages expire after TTL""" |
| 154 | # Publish with short TTL |
| 155 | message_id = await channels.publish( |
| 156 | Channel.ORDERS.value, |
| 157 | "expires_soon", |
| 158 | {"data": "test"}, |
| 159 | ttl_seconds=1 |
| 160 | ) |
| 161 | |
| 162 | # Wait for expiration |
| 163 | await asyncio.sleep(2) |
| 164 | |
| 165 | # Try to get the message - should be expired |
| 166 | result = channels.conn.execute(""" |
| 167 | SELECT * FROM messages |
| 168 | WHERE id = ? |
| 169 | AND (created_at + INTERVAL '1 second' * ttl_seconds) > CURRENT_TIMESTAMP |
| 170 | """, [message_id]).fetchone() |
| 171 | |
| 172 | assert result is None # Message should be expired |
| 173 | |
| 174 | @pytest.mark.asyncio |
| 175 | async def test_concurrent_publishing(self, channels): |
| 176 | """Test concurrent publishing from multiple coroutines""" |
| 177 | |
| 178 | async def publisher(channel, count): |
| 179 | for i in range(count): |
| 180 | await channels.publish( |
| 181 | channel, |
| 182 | f"msg_{i}", |
| 183 | {"index": i} |
| 184 | ) |
| 185 | |
| 186 | # Run multiple publishers concurrently |
| 187 | tasks = [ |
| 188 | publisher(Channel.ORDERS.value, 10), |
| 189 | publisher(Channel.SHIPPING.value, 10), |
| 190 | publisher(Channel.FRAUD.value, 10) |
| 191 | ] |
| 192 | |
| 193 | start_time = time.time() |
| 194 | await asyncio.gather(*tasks) |
| 195 | elapsed = time.time() - start_time |
| 196 | |
| 197 | # Check all messages were inserted |
| 198 | count = channels.conn.execute( |
| 199 | "SELECT COUNT(*) FROM messages" |
| 200 | ).fetchone()[0] |
| 201 | |
| 202 | assert count == 30 |
| 203 | assert elapsed < 1.0 # Should be very fast |
| 204 | |
| 205 | @pytest.mark.asyncio |
| 206 | async def test_channel_statistics(self, channels): |
| 207 | """Test channel statistics tracking""" |
| 208 | # Publish to various channels |
| 209 | await channels.publish(Channel.ORDERS.value, "order", {}) |
| 210 | await channels.publish(Channel.ORDERS.value, "order", {}) |
| 211 | await channels.publish(Channel.SHIPPING.value, "ship", {}) |
| 212 | |
| 213 | stats = channels.get_channel_stats() |
| 214 | |
| 215 | assert stats[Channel.ORDERS.value]['total_messages'] >= 2 |
| 216 | assert stats[Channel.SHIPPING.value]['total_messages'] >= 1 |
| 217 | |
| 218 | |
| 219 | class TestDBBasicTaskQueue: |
| 220 | """Test suite for DBBasic Task Queue""" |
| 221 | |
| 222 | @pytest.fixture |
| 223 | def queue(self): |
| 224 | """Create a test task queue instance""" |
| 225 | queue = DBBasicTaskQueue(db_path=":memory:") |
| 226 | yield queue |
| 227 | queue.conn.close() |
| 228 | |
| 229 | def test_add_single_task(self, queue): |
| 230 | """Test adding a single task""" |
| 231 | task_id = queue.add_task( |
| 232 | task_type="test_task", |
| 233 | payload={"key": "value"}, |
| 234 | priority=5 |
| 235 | ) |
| 236 | |
| 237 | assert task_id is not None |
| 238 | |
| 239 | # Verify in database |
| 240 | result = queue.conn.execute( |
| 241 | "SELECT * FROM task_queue WHERE id = ?", [task_id] |
| 242 | ).fetchone() |
| 243 | |
| 244 | assert result is not None |
| 245 | assert result[1] == "test_task" # task_type |
| 246 | assert result[4] == "pending" # status |
| 247 | |
| 248 | def test_bulk_add_tasks(self, queue): |
| 249 | """Test bulk adding tasks - this tests our speed claims!""" |
| 250 | tasks = [ |
| 251 | { |
| 252 | 'task_type': 'bulk_test', |
| 253 | 'service_name': 'test_service', |
| 254 | 'payload': {'index': i}, |
| 255 | 'priority': i % 10 |
| 256 | } |
| 257 | for i in range(10000) |
| 258 | ] |
| 259 | |
| 260 | start_time = time.time() |
| 261 | count = queue.bulk_add_tasks(tasks) |
| 262 | elapsed = time.time() - start_time |
| 263 | |
| 264 | tasks_per_second = count / elapsed |
| 265 | |
| 266 | assert count == 10000 |
| 267 | assert elapsed < 1.0 # Should be very fast |
| 268 | print(f"\n⚡ Bulk insert rate: {tasks_per_second:,.0f} tasks/second") |
| 269 | |
| 270 | # Verify all were inserted |
| 271 | db_count = queue.conn.execute( |
| 272 | "SELECT COUNT(*) FROM task_queue" |
| 273 | ).fetchone()[0] |
| 274 | |
| 275 | assert db_count == 10000 |
| 276 | |
| 277 | def test_get_next_task_priority(self, queue): |
| 278 | """Test that tasks are retrieved by priority""" |
| 279 | # Add tasks with different priorities |
| 280 | low_id = queue.add_task("low", {}, priority=1) |
| 281 | high_id = queue.add_task("high", {}, priority=10) |
| 282 | medium_id = queue.add_task("medium", {}, priority=5) |
| 283 | |
| 284 | # Should get high priority first |
| 285 | task = queue.get_next_task("worker1") |
| 286 | assert task is not None |
| 287 | assert task['task_type'] == "high" |
| 288 | |
| 289 | # Then medium |
| 290 | task = queue.get_next_task("worker1") |
| 291 | assert task['task_type'] == "medium" |
| 292 | |
| 293 | # Then low |
| 294 | task = queue.get_next_task("worker1") |
| 295 | assert task['task_type'] == "low" |
| 296 | |
| 297 | def test_task_locking(self, queue): |
| 298 | """Test that locked tasks aren't given to other workers""" |
| 299 | task_id = queue.add_task("test", {}) |
| 300 | |
| 301 | # Worker 1 gets the task |
| 302 | task1 = queue.get_next_task("worker1", lock_duration_seconds=60) |
| 303 | assert task1 is not None |
| 304 | |
| 305 | # Worker 2 shouldn't get the same task |
| 306 | task2 = queue.get_next_task("worker2") |
| 307 | assert task2 is None # No tasks available |
| 308 | |
| 309 | def test_complete_task(self, queue): |
| 310 | """Test completing a task""" |
| 311 | task_id = queue.add_task("test", {"data": "test"}) |
| 312 | |
| 313 | # Get and complete the task |
| 314 | task = queue.get_next_task("worker1") |
| 315 | queue.complete_task(task['id'], result={"status": "done"}) |
| 316 | |
| 317 | # Check status |
| 318 | result = queue.conn.execute( |
| 319 | "SELECT status, result FROM task_queue WHERE id = ?", |
| 320 | [task['id']] |
| 321 | ).fetchone() |
| 322 | |
| 323 | assert result[0] == "completed" |
| 324 | assert json.loads(result[1])['status'] == "done" |
| 325 | |
| 326 | def test_fail_task_with_retry(self, queue): |
| 327 | """Test failing a task with retry logic""" |
| 328 | task_id = queue.add_task("test", {}, max_attempts=3) |
| 329 | |
| 330 | # Fail the task twice |
| 331 | for attempt in range(2): |
| 332 | task = queue.get_next_task("worker1") |
| 333 | assert task is not None |
| 334 | queue.fail_task(task['id'], f"Error {attempt}") |
| 335 | |
| 336 | # Check it goes back to pending |
| 337 | status = queue.conn.execute( |
| 338 | "SELECT status FROM task_queue WHERE id = ?", [task_id] |
| 339 | ).fetchone()[0] |
| 340 | assert status == "pending" |
| 341 | |
| 342 | # Third failure should mark as failed |
| 343 | task = queue.get_next_task("worker1") |
| 344 | queue.fail_task(task['id'], "Final error") |
| 345 | |
| 346 | status = queue.conn.execute( |
| 347 | "SELECT status FROM task_queue WHERE id = ?", [task_id] |
| 348 | ).fetchone()[0] |
| 349 | assert status == "failed" |
| 350 | |
| 351 | def test_metrics_calculation(self, queue): |
| 352 | """Test metrics calculation""" |
| 353 | # Add various tasks |
| 354 | for i in range(10): |
| 355 | queue.add_task("test", {"index": i}) |
| 356 | |
| 357 | # Process some |
| 358 | for i in range(5): |
| 359 | task = queue.get_next_task(f"worker_{i}") |
| 360 | if task: |
| 361 | queue.complete_task(task['id']) |
| 362 | |
| 363 | metrics = queue.get_metrics() |
| 364 | |
| 365 | assert metrics['pending_tasks'] == 5 |
| 366 | assert metrics['completed_tasks'] == 5 |
| 367 | assert metrics['theoretical_max_rows_sec'] == 402_000_000 |
| 368 | |
| 369 | @pytest.mark.asyncio |
| 370 | async def test_task_worker(self, queue): |
| 371 | """Test the TaskWorker class""" |
| 372 | processed = [] |
| 373 | |
| 374 | async def test_processor(task): |
| 375 | processed.append(task) |
| 376 | return {"processed": True} |
| 377 | |
| 378 | # Create worker |
| 379 | worker = TaskWorker("test_worker", queue) |
| 380 | worker.process_task = test_processor |
| 381 | |
| 382 | # Add tasks |
| 383 | for i in range(3): |
| 384 | queue.add_task("test", {"index": i}) |
| 385 | |
| 386 | # Run worker briefly |
| 387 | worker_task = asyncio.create_task(worker.start()) |
| 388 | await asyncio.sleep(0.5) |
| 389 | worker.stop() |
| 390 | worker_task.cancel() |
| 391 | try: |
| 392 | await worker_task |
| 393 | except asyncio.CancelledError: |
| 394 | pass |
| 395 | |
| 396 | # Check tasks were processed |
| 397 | assert len(processed) == 3 |
| 398 | |
| 399 | |
| 400 | class TestPerformanceBenchmarks: |
| 401 | """Performance benchmarks comparing with PostgreSQL baseline""" |
| 402 | |
| 403 | @pytest.fixture |
| 404 | def queue(self): |
| 405 | return DBBasicTaskQueue(db_path=":memory:") |
| 406 | |
| 407 | @pytest.fixture |
| 408 | async def channels(self): |
| 409 | return DBBasicChannels(db_path=":memory:") |
| 410 | |
| 411 | def test_benchmark_insert_speed(self, queue): |
| 412 | """Benchmark: Insert speed vs PostgreSQL baseline""" |
| 413 | counts = [100, 1000, 10000] |
| 414 | results = [] |
| 415 | |
| 416 | for count in counts: |
| 417 | tasks = [ |
| 418 | {'task_type': 'bench', 'payload': {'i': i}} |
| 419 | for i in range(count) |
| 420 | ] |
| 421 | |
| 422 | # Use bulk insert |
| 423 | start = time.time() |
| 424 | queue.bulk_add_tasks(tasks) |
| 425 | elapsed = time.time() - start |
| 426 | |
| 427 | rate = count / elapsed |
| 428 | results.append({ |
| 429 | 'count': count, |
| 430 | 'time': elapsed, |
| 431 | 'rate': rate |
| 432 | }) |
| 433 | |
| 434 | print(f"\n📊 {count:,} tasks: {elapsed:.3f}s = {rate:,.0f} tasks/sec") |
| 435 | |
| 436 | # PostgreSQL baseline: ~5,000-10,000 inserts/sec |
| 437 | # Our target: 100,000+ inserts/sec |
| 438 | assert all(r['rate'] > 10000 for r in results), "Should beat PostgreSQL baseline" |
| 439 | |
| 440 | @pytest.mark.asyncio |
| 441 | async def test_benchmark_concurrent_workers(self, queue): |
| 442 | """Benchmark: Concurrent worker processing""" |
| 443 | |
| 444 | # Add many tasks |
| 445 | tasks = [ |
| 446 | {'task_type': 'work', 'payload': {'id': i}} |
| 447 | for i in range(1000) |
| 448 | ] |
| 449 | queue.bulk_add_tasks(tasks) |
| 450 | |
| 451 | processed_count = 0 |
| 452 | |
| 453 | async def worker_proc(worker_id): |
| 454 | nonlocal processed_count |
| 455 | while True: |
| 456 | task = queue.get_next_task(worker_id) |
| 457 | if not task: |
| 458 | break |
| 459 | queue.complete_task(task['id']) |
| 460 | processed_count += 1 |
| 461 | |
| 462 | # Run multiple workers |
| 463 | start = time.time() |
| 464 | workers = [worker_proc(f"worker_{i}") for i in range(10)] |
| 465 | await asyncio.gather(*workers) |
| 466 | elapsed = time.time() - start |
| 467 | |
| 468 | rate = processed_count / elapsed |
| 469 | print(f"\n📊 Processed {processed_count} tasks with 10 workers") |
| 470 | print(f" Time: {elapsed:.3f}s") |
| 471 | print(f" Rate: {rate:,.0f} tasks/sec") |
| 472 | |
| 473 | # Should process very quickly with concurrent workers |
| 474 | assert rate > 500, "Should process >500 tasks/sec with 10 workers" |
| 475 | |
| 476 | @pytest.mark.asyncio |
| 477 | async def test_benchmark_pub_sub_latency(self, channels): |
| 478 | """Benchmark: Pub/Sub message latency""" |
| 479 | latencies = [] |
| 480 | |
| 481 | async def measure_latency(): |
| 482 | received_time = None |
| 483 | |
| 484 | async def callback(msg): |
| 485 | nonlocal received_time |
| 486 | received_time = time.time() |
| 487 | |
| 488 | await channels.subscribe(Channel.ORDERS.value, callback) |
| 489 | |
| 490 | # Measure latency |
| 491 | send_time = time.time() |
| 492 | await channels.publish(Channel.ORDERS.value, "test", {}) |
| 493 | |
| 494 | # Wait for message |
| 495 | await asyncio.sleep(0.01) |
| 496 | |
| 497 | if received_time: |
| 498 | latency = (received_time - send_time) * 1000 # ms |
| 499 | latencies.append(latency) |
| 500 | |
| 501 | # Run multiple measurements |
| 502 | for _ in range(100): |
| 503 | await measure_latency() |
| 504 | |
| 505 | avg_latency = statistics.mean(latencies) |
| 506 | p99_latency = statistics.quantiles(latencies, n=100)[98] |
| 507 | |
| 508 | print(f"\n📊 Pub/Sub Latency:") |
| 509 | print(f" Average: {avg_latency:.3f}ms") |
| 510 | print(f" P99: {p99_latency:.3f}ms") |
| 511 | |
| 512 | # Should have very low latency |
| 513 | assert avg_latency < 10, "Average latency should be <10ms" |
| 514 | assert p99_latency < 50, "P99 latency should be <50ms" |
| 515 | |
| 516 | |
| 517 | def test_integration_full_pipeline(): |
| 518 | """Integration test: Complete pipeline from publish to process""" |
| 519 | |
| 520 | async def run_pipeline(): |
| 521 | channels = DBBasicChannels(db_path=":memory:") |
| 522 | queue = DBBasicTaskQueue(db_path=":memory:") |
| 523 | |
| 524 | results = [] |
| 525 | |
| 526 | # Setup: Channel subscriber that adds to task queue |
| 527 | async def channel_to_queue(message): |
| 528 | if message['message_type'] == 'new_order': |
| 529 | queue.add_task( |
| 530 | 'process_order', |
| 531 | message['payload'], |
| 532 | service_name='order_processor' |
| 533 | ) |
| 534 | |
| 535 | await channels.subscribe(Channel.ORDERS.value, channel_to_queue) |
| 536 | |
| 537 | # Publish order messages |
| 538 | for i in range(10): |
| 539 | await channels.publish( |
| 540 | Channel.ORDERS.value, |
| 541 | 'new_order', |
| 542 | {'order_id': f'ORD{i}', 'amount': i * 100} |
| 543 | ) |
| 544 | |
| 545 | await asyncio.sleep(0.1) # Let messages propagate |
| 546 | |
| 547 | # Process tasks from queue |
| 548 | for _ in range(10): |
| 549 | task = queue.get_next_task('worker1') |
| 550 | if task: |
| 551 | queue.complete_task(task['id'], {'processed': True}) |
| 552 | results.append(task) |
| 553 | |
| 554 | # Verify pipeline worked |
| 555 | assert len(results) == 10 |
| 556 | assert all(t['task_type'] == 'process_order' for t in results) |
| 557 | |
| 558 | # Check metrics |
| 559 | metrics = queue.get_metrics() |
| 560 | assert metrics['completed_tasks'] == 10 |
| 561 | |
| 562 | print(f"\n✅ Full pipeline test: {len(results)} orders processed") |
| 563 | |
| 564 | asyncio.run(run_pipeline()) |
| 565 | |
| 566 | |
| 567 | if __name__ == "__main__": |
| 568 | # Run tests with pytest |
| 569 | pytest.main([__file__, "-v", "--tb=short"]) |
| 570 | |
| 571 | # Or run specific benchmarks |
| 572 | print("\n" + "="*60) |
| 573 | print("Running Performance Benchmarks") |
| 574 | print("="*60) |
| 575 | |
| 576 | queue = DBBasicTaskQueue(db_path=":memory:") |
| 577 | |
| 578 | # Benchmark 1: Bulk insert speed |
| 579 | print("\n📊 Benchmark: Bulk Insert Speed") |
| 580 | tasks = [{'task_type': 'bench', 'payload': {'i': i}} for i in range(100000)] |
| 581 | start = time.time() |
| 582 | queue.bulk_add_tasks(tasks) |
| 583 | elapsed = time.time() - start |
| 584 | print(f" 100,000 tasks inserted in {elapsed:.3f}s") |
| 585 | print(f" Rate: {100000/elapsed:,.0f} tasks/second") |
| 586 | print(f" vs PostgreSQL: ~{(100000/elapsed)/10000:.1f}x faster") |
| 587 | |
| 588 | print("\n🚀 DBBasic is ready to replace PostgreSQL queues!") |