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>

write_result.py

@@ -0,0 +1,47 @@
+import json, os, sys
+sys.path.insert(0, ".")
+os.environ["LLM_API_KEY"] = "sk-ca4961087c7f4aefa8ed0fc6f3d02329"
+os.environ["LLM_API_BASE"] = "https://api.deepseek.com/v1"
+
+from config import Config
+from orchestrator import run_pipeline
+
+cfg = Config()
+cfg.llm_model = "deepseek-chat"
+cfg.runner_mode = "native"
+
+task_id = sys.argv[1] if len(sys.argv) > 1 else "ec17bf32"
+tf = f"tasks/{task_id}.json"
+data = json.load(open(tf))
+
+vr = run_pipeline(cfg, f"uploads/{task_id}/copybook.cpy", f"uploads/{task_id}/program.cbl",
+                  f"uploads/{task_id}/java", f"uploads/{task_id}/mapping.yaml")
+
+fields = [{"name":fr.field_name,"status":fr.status,
+           "cobol":str(fr.cobol_value),"java":str(fr.java_value),
+           "suggestion":fr.suggestion} for fr in vr.field_results]
+
+debug = vr.debug
+if "field_tree" in debug:
+    debug["field_tree"] = [{"name":f["name"],"level":f["level"],"pic":f["pic"],
+        "usage":f["usage"],"offset":f["offset"],"length":f["length"]} for f in debug["field_tree"]]
+if "test_cases" in debug:
+    debug["test_cases"] = [{"id":tc["id"],"fields":tc["fields"],
+        "targets":tc.get("targets",[])} for tc in debug["test_cases"]]
+for k in ("cobol_build","java_build"):
+    if k in debug and debug[k]:
+        debug[k]["log"] = debug[k].get("log","")[-500:]
+
+data["status"] = "done"
+data["fields"] = fields
+data["debug"] = debug
+data["result"] = {
+    "program": vr.program, "status": vr.status,
+    "matched": vr.fields_matched, "mismatched": vr.fields_mismatched,
+    "duration": round(vr.duration_s, 1), "runner": vr.runner,
+}
+json.dump(data, open(tf, "w"))
+print("Task updated!")
+print(f"Status: {vr.status}, Matched: {vr.fields_matched}, Mismatched: {vr.fields_mismatched}")
+for fr in vr.field_results:
+    print(f"  {fr.field_name}: {fr.status} (COBOL={fr.cobol_value}, Java={fr.java_value})")