chunker and ui

tests/chunker/test_worker.py

@@ -0,0 +1,127 @@
+"""
+Tests for Worker — processing, retry with backoff, error handling.
+
+Demonstrates: TDD (Interview Topic 8) — mocking processors, testing retry logic.
+"""
+
+from unittest.mock import MagicMock
+
+import pytest
+
+from core.chunker.models import Chunk, ChunkResult
+from core.chunker.processor import Processor
+from core.chunker.queue import ChunkQueue
+from core.chunker.worker import Worker
+
+
+class FailNTimesProcessor(Processor):
+    """Test processor that fails N times then succeeds."""
+
+    def __init__(self, fail_count: int):
+        self.fail_count = fail_count
+        self.call_count = 0
+
+    def process(self, chunk: Chunk) -> ChunkResult:
+        self.call_count += 1
+        if self.call_count <= self.fail_count:
+            raise RuntimeError(f"Simulated failure #{self.call_count}")
+        return ChunkResult(
+            sequence=chunk.sequence,
+            success=True,
+            processing_time=0.001,
+        )
+
+
+class AlwaysFailProcessor(Processor):
+    """Test processor that always fails."""
+
+    def process(self, chunk: Chunk) -> ChunkResult:
+        raise RuntimeError("Always fails")
+
+
+class TestWorker:
+    def test_processes_chunks(self, make_chunk):
+        """Worker processes all chunks from queue."""
+        q = ChunkQueue(maxsize=5)
+        for i in range(3):
+            q.put(make_chunk(i))
+        q.close()
+
+        from core.chunker.processor import ChecksumProcessor
+        worker = Worker("w-0", q, ChecksumProcessor(), max_retries=0)
+        results = worker.run()
+
+        assert len(results) == 3
+        assert all(r.success for r in results)
+
+    def test_retry_on_failure(self, make_chunk):
+        """Worker retries on processor failure."""
+        q = ChunkQueue(maxsize=5)
+        q.put(make_chunk(0))
+        q.close()
+
+        proc = FailNTimesProcessor(fail_count=2)
+        worker = Worker("w-0", q, proc, max_retries=3)
+        results = worker.run()
+
+        assert len(results) == 1
+        assert results[0].success is True
+        assert results[0].retries == 2
+        assert proc.call_count == 3  # 2 failures + 1 success
+
+    def test_max_retries_exceeded(self, make_chunk):
+        """Worker gives up after max retries."""
+        q = ChunkQueue(maxsize=5)
+        q.put(make_chunk(0))
+        q.close()
+
+        worker = Worker("w-0", q, AlwaysFailProcessor(), max_retries=2)
+        results = worker.run()
+
+        assert len(results) == 1
+        assert results[0].success is False
+        assert results[0].error is not None
+        assert worker.error_count == 1
+
+    def test_worker_id_on_results(self, make_chunk):
+        """Worker stamps its ID on results."""
+        q = ChunkQueue(maxsize=5)
+        q.put(make_chunk(0))
+        q.close()
+
+        from core.chunker.processor import ChecksumProcessor
+        worker = Worker("worker-7", q, ChecksumProcessor())
+        results = worker.run()
+
+        assert results[0].worker_id == "worker-7"
+
+    def test_event_callback(self, make_chunk):
+        """Worker emits events via callback."""
+        q = ChunkQueue(maxsize=5)
+        q.put(make_chunk(0))
+        q.close()
+
+        events = []
+        callback = MagicMock(side_effect=lambda t, d: events.append((t, d)))
+
+        from core.chunker.processor import ChecksumProcessor
+        worker = Worker("w-0", q, ChecksumProcessor(), event_callback=callback)
+        worker.run()
+
+        event_types = [e[0] for e in events]
+        assert "worker_status" in event_types
+        assert "chunk_processing" in event_types
+        assert "chunk_done" in event_types
+
+    def test_processed_count(self, make_chunk):
+        """Worker tracks processed count."""
+        q = ChunkQueue(maxsize=10)
+        for i in range(5):
+            q.put(make_chunk(i))
+        q.close()
+
+        from core.chunker.processor import ChecksumProcessor
+        worker = Worker("w-0", q, ChecksumProcessor())
+        worker.run()
+
+        assert worker.processed_count == 5