Buffer (Batching)

[prosdev]

Phase: 4 - Processing Pipeline
Estimated Effort: 3 hours
Tasks: Task 8 from specs/core-pipeline/tasks.md

Checklist

Task 8: Implement Buffer (3 hours)

• Create Buffer class
• Implement per-partition buffers (dict[int, list[TypedEvent]])
• Implement enqueue() with size-based flush (default: 100)
• Implement _flush_partition() (batch write to EventStore)
• Implement time-based flusher (async background task, default: 5s)
• Implement start_flusher() / stop_flusher()
• Implement graceful shutdown (flush all buffers)
• Write unit tests (mocked EventStore)

Acceptance Criteria

• Per-partition buffering (no cross-partition blocking)
• Size-based flush works
• Time-based flush works
• Graceful shutdown flushes all buffers
• Unit tests pass

Files to Create

src/eventkit/processing/buffer.py
tests/unit/processing/test_buffer.py

Dependencies: Issue #2 (EventStore)
User Stories: Story 5 (Efficient Batch Processing)