feat: config-driven multi-scenario + gcda accumulation fix + gcov merge fix

Config-driven architecture:
- coverage_dates: YAML-driven date replacement for LEAVE_RECORDS
- row_overrides: per-scenario field overrides (e.g. STATUS='9')
- delete_all_rows: per-scenario table emptying for empty-cursor tests

gcda accumulation fix (V3 bug #7):
- Delete .gcda before each scenario run to prevent GnuCOBOL accumulation
- Force-copy scenario gcda (skip mtime check)
- Copy scenario DB to CWD data/kin.db for CONNECT TO path resolution

Multi-run gcov merge fix:
- Always merge multi-run gcov data regardless of generate_coverage flag
- Fix v3_root UnboundLocalError in cleanup path

Other fixes:
- _make_key_unique: skip WHERE-constrained columns to avoid PK conflict
- incremental_supplement: support fields_dict for base record generation
- check_coverage: use structure coverage data if available
- orchestrator.py: filter _-prefixed fields in TestCase; merge Agent2Data cases

orchestrator.py

@@ -49,7 +49,8 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
             # 转换为 TestCase 列表（增强管线的基础数据集）
             complete_tests = []
             for i, rec in enumerate(base_records):
-                complete_tests.append(TestCase(id=f"CTG-{i+1:04d}", fields=dict(rec)))
+                fields = {k: v for k, v in rec.items() if not k.startswith('_')}
+                complete_tests.append(TestCase(id=f"CTG-{i+1:04d}", fields=fields))
 
             # HINA 完整类型判定管道（Keyword / 规则引擎 / LLM 辅助三路径）
             classification: dict = {}
@@ -84,13 +85,15 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
                     break
                 gaps = gate_result.get("issues", {}).get("decision_gaps", [])
                 if gaps and structure.get("branch_tree_obj"):
-                    delta = incremental_supplement(structure["branch_tree_obj"], gaps)
+                    data_fields = structure.get("data_fields", [])
+                    delta = incremental_supplement(structure["branch_tree_obj"], gaps, data_fields)
                     base_records.extend(delta)
                     # 同步更新 complete_tests
                     for i, d in enumerate(delta):
+                        d_fields = {k: v for k, v in d.items() if not k.startswith('_')}
                         complete_tests.append(TestCase(
                             id=f"CTG-S{attempt+1}-{i+1:04d}",
-                            fields=dict(d),
+                            fields=d_fields,
                         ))
                     cov = check_coverage(structure, base_records)
                 else:
@@ -109,7 +112,13 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
 
         suite = Agent2Data(llm).design(tree, cfg.coverage_default, cfg.runner_mode == "spark")
         vr.llm_cost += 0.002
-        suite.test_cases = complete_tests  # 替换为增强管线数据（P1/P2 修复）
+        # 合并: cobol_testgen 基线 + 策略标记 + Agent2Data LLM 场景数据
+        merged = list(complete_tests)
+        existing_ids = {tc.id for tc in merged}
+        for tc in suite.test_cases:
+            if tc.id not in existing_ids:
+                merged.append(tc)
+        suite.test_cases = merged
         vr.debug["test_cases"] = [{"id":tc.id,"fields":tc.fields,"targets":tc.coverage_targets} for tc in suite.test_cases]
         vr.debug["spark_config"] = {"records":suite.spark_config.num_records} if suite.has_spark else None