feat: 多轮运行 + GCOV 合并 + JSON 出力 + DesignDataGenerator

cobol_testgen/data_merger.py

@@ -0,0 +1,130 @@
+"""データ統合 — 白盒 + 机能 + 策略データの統合と合併。
+
+generate_all_data() エントリポイント:
+  ① generate_data()         → 白盒（MC/DC パスカバレッジ）
+  ② DesignDataGenerator     → 机能（式样书から LLM 生成）
+  ③ strategy_supplement()   → 策略（HINA 分類に基づく境界条件）
+  ④ 重複除去 + フィールド名正規化
+  ⑤ 統合リスト返却
+"""
+
+import logging
+from pathlib import Path
+from typing import Optional
+
+from cobol_testgen import extract_structure, generate_data
+
+logger = logging.getLogger(__name__)
+
+
+def _dedup(
+    main_records: list[dict],
+    additional_records: list[dict] | None = None,
+    key_fields: list[str] | None = None,
+) -> list[dict]:
+    """合并+去重，additional 优先保留。"""
+    if not additional_records:
+        return list(main_records)
+
+    seen = set()
+    result = []
+
+    def _hash(rec, keys):
+        if keys:
+            return tuple(rec.get(k, "") for k in keys)
+        return tuple(sorted(rec.items()))
+
+    for rec in additional_records:
+        h = _hash(rec, key_fields)
+        if h not in seen:
+            seen.add(h)
+            result.append(rec)
+
+    for rec in main_records:
+        h = _hash(rec, key_fields)
+        if h not in seen:
+            seen.add(h)
+            result.append(rec)
+
+    return result
+
+
+def generate_all_data(
+    program_id: str,
+    src_text: str,
+    st: dict | None = None,
+    copybook_dirs: list[str | Path] | None = None,
+    design_doc_dir: str | Path | None = None,
+    llm_client=None,
+    config=None,
+    merge_strategy: str = "merge_to_normal",
+) -> list[dict]:
+    """白盒 + 机能 + 策略 全量生成と統合。
+
+    Args:
+        program_id: プログラム ID
+        src_text: COBOL ソーステキスト
+        st: extract_structure() 結果（省略時は内部で再解析）
+        copybook_dirs: COPYBOOK 探索パス
+        design_doc_dir: 式样书配置ディレクトリ
+        llm_client: LLMClient インスタンス（None で LLM 系スキップ）
+        config: Config インスタンス
+        merge_strategy: merge_to_normal / as_separate_scenes / auto
+
+    Returns:
+        list[dict]: 統合済みレコードリスト
+    """
+    cbd = [str(d) for d in (copybook_dirs or [])]
+
+    # ① 白盒データ
+    if st is None:
+        st = extract_structure(src_text, copybook_dirs=cbd)
+    whitebox = generate_data(src_text, st, copybook_dirs=cbd)
+    logger.info(f"  White-box records: {len(whitebox)}")
+
+    # ② 机能データ（式样书 + LLM）
+    func_data: list[dict] = []
+    if design_doc_dir and llm_client:
+        design_path = Path(design_doc_dir) / f"詳細設計書_{program_id}.md"
+        if design_path.exists():
+            from agents.design_data import DesignDataGenerator, _extract_replacing_rules
+
+            gen = DesignDataGenerator(llm_client, cbd)
+            v3_names = list(st.get("field_names", [])) if st else None
+            replacing = _extract_replacing_rules(src_text)
+
+            try:
+                func_data = gen.generate(
+                    design_md_text=design_path.read_text(encoding="utf-8"),
+                    source_text=src_text,
+                    replacing_rules=replacing,
+                    v3_field_names=v3_names,
+                )
+            except Exception as e:
+                logger.warning(f"  DesignDataGenerator failed: {e}")
+        else:
+            logger.info(f"  Design doc not found: {design_path}")
+
+    logger.info(f"  Functional records: {len(func_data)}")
+
+    # ③ 策略データ
+    strategy_data: list[dict] = []
+    try:
+        from hina.strategy import supplement
+
+        strat_raw = supplement([], {})
+        for s in strat_raw:
+            if isinstance(s, dict) and "fields" in s:
+                strategy_data.append(s["fields"])
+    except Exception as e:
+        logger.debug(f"  Strategy supplement skipped: {e}")
+
+    logger.info(f"  Strategy records: {len(strategy_data)}")
+
+    # ④ 統合
+    all_records = _dedup(whitebox, func_data)
+    if strategy_data:
+        all_records = _dedup(all_records, strategy_data)
+
+    logger.info(f"  Total merged records: {len(all_records)}")
+    return all_records