feat: Phase 2 complete — 13 Phases of COBOL type classification and test benchmark

P0.6: gcov infrastructure
P1: extract_structure output expansion (11 new feature fields)
P2: Confusion group rule engine (8 pairs + contradiction + backtrack)
P3: 4-factor confidence calculation + quality gate update
P4: 33+2 COBOL program type test samples (22 files, 7 categories)
P5: parametrized/ test data generation engine
P6: japanese_data.py lookup tables
P7-10: Type-specific test suites (~159 parametrized tests)
P11: Full classification pipeline (classify_program) + orchestrator integration
P12: Documentation (module-interfaces, test-plan v3.0, coverage-matrix)

Architecture decisions:
- classification_pipeline/ merged to hina/pipeline/
- parametrized/ as independent module
- japanese_data.py as root-level file
- hina/__all__ only exports classify_program()

Co-Authored-By: Claude <noreply@anthropic.com>

web/__init__.py

@@ -0,0 +1,10 @@
+"""Web API + Worker 包
+
+公开 API:
+  api.py  →  FastAPI 应用（由 uvicorn 启动）
+  worker.py → 后台任务处理循环
+"""
+
+from __future__ import annotations
+
+__all__ = []

web/worker.py

@@ -10,8 +10,18 @@ def main():
     print("Worker started. Watching tasks/ ...")
     while True:
         for tf in sorted(TASKS_DIR.glob("*.json")):
+            data = {}
             try:
-                data = json.loads(tf.read_text())
+                raw = tf.read_text()
+                if not raw.strip():
+                    data = {"id": tf.stem, "status": "error", "result": "empty file"}
+                    tf.write_text(json.dumps(data))
+                    continue
+                data = json.loads(raw)
+                if not isinstance(data, dict):
+                    data = {"id": tf.stem, "status": "error", "result": "invalid JSON type"}
+                    tf.write_text(json.dumps(data))
+                    continue
                 if data.get("status") != "queued":
                     continue
 
@@ -49,10 +59,12 @@ def main():
                 tf.write_text(json.dumps(data))
 
             except Exception as e:
-                data = json.loads(tf.read_text()) if tf.exists() else {}
                 data["status"] = "error"
                 data["result"] = str(e)[:500]
-                tf.write_text(json.dumps(data))
+                try:
+                    tf.write_text(json.dumps(data))
+                except Exception:
+                    pass  # 无法写入错误状态时静默跳过
 
         time.sleep(2)