添加环境变量映射，支持完整性检查任务和工具类任务的执行

etl_billiards/orchestration/scheduler.py

@@ -129,6 +129,13 @@ class ETLScheduler:
     # ------------------------------------------------------------------ internals
     def _run_single_task(self, task_code: str, run_uuid: str, store_id: int):
         """单个任务的抓取/清洗编排。"""
+        task_code_upper = task_code.upper()
+        
+        # 工具类任务：直接执行，不记录 cursor/run
+        if task_code_upper in self.NO_DB_CONFIG_TASKS:
+            self._run_utility_task(task_code_upper, store_id)
+            return
+        
         task_cfg = self._load_task_config(task_code, store_id)
         if not task_cfg:
             self.logger.warning("任务 %s 未启用或不存在", task_code)
@@ -183,6 +190,7 @@ class ETLScheduler:
                             window_end=window.get("end"),
                             run_id=run_id,
                         )
+                        self._maybe_run_integrity_check(task_code, window)
                 return
 
             if self._flow_includes_fetch():
@@ -221,6 +229,7 @@ class ETLScheduler:
                             window_end=window.get("end"),
                             run_id=run_id,
                         )
+                        self._maybe_run_integrity_check(task_code, window)
 
         except Exception as exc:  # noqa: BLE001
             self.run_tracker.update_run(
@@ -322,6 +331,53 @@ class ETLScheduler:
     def _flow_includes_ingest(self) -> bool:
         return self.pipeline_flow in {"INGEST_ONLY", "FULL"}
 
+    # 不需要数据库配置即可运行的任务（工具类/初始化类）
+    NO_DB_CONFIG_TASKS = {
+        # Schema 初始化任务
+        "INIT_ODS_SCHEMA",
+        "INIT_DWD_SCHEMA",
+        "INIT_DWS_SCHEMA",
+        # 质量检查任务
+        "DATA_INTEGRITY_CHECK",
+        "DWD_QUALITY_CHECK",
+        # 工具任务
+        "CHECK_CUTOFF",
+        "MANUAL_INGEST",
+        "ODS_JSON_ARCHIVE",
+        # DWS 汇总任务
+        "DWS_BUILD_ORDER_SUMMARY",
+    }
+
+    def _run_utility_task(self, task_code: str, store_id: int):
+        """
+        执行工具类任务（不记录 cursor/run，直接执行）。
+        这些任务不需要游标管理和运行跟踪。
+        """
+        self.logger.info("%s: 开始执行工具类任务", task_code)
+        
+        try:
+            # 创建任务实例（不需要 API client，使用 None）
+            task = self.task_registry.create_task(
+                task_code, self.config, self.db_ops, None, self.logger
+            )
+            
+            # 执行任务（工具类任务通常不需要 cursor_data）
+            result = task.execute(None)
+            
+            status = (result.get("status") or "").upper() if isinstance(result, dict) else "SUCCESS"
+            if status == "SUCCESS":
+                self.logger.info("%s: 工具类任务执行成功", task_code)
+                if isinstance(result, dict):
+                    counts = result.get("counts", {})
+                    if counts:
+                        self.logger.info("%s: 结果统计: %s", task_code, counts)
+            else:
+                self.logger.warning("%s: 工具类任务执行结果: %s", task_code, status)
+                
+        except Exception as exc:
+            self.logger.error("%s: 工具类任务执行失败: %s", task_code, exc, exc_info=True)
+            raise
+
     def _load_task_config(self, task_code: str, store_id: int) -> dict | None:
         """从数据库加载任务配置。"""
         sql = """
@@ -334,6 +390,45 @@ class ETLScheduler:
         rows = self.db_conn.query(sql, (store_id, task_code))
         return rows[0] if rows else None
 
+    def _maybe_run_integrity_check(self, task_code: str, window: dict | None) -> None:
+        if not self.config.get("integrity.auto_check", False):
+            return
+        if str(task_code or "").upper() != "DWD_LOAD_FROM_ODS":
+            return
+        if not isinstance(window, dict):
+            return
+        window_start = window.get("start")
+        window_end = window.get("end")
+        if not window_start or not window_end:
+            return
+
+        try:
+            from quality.integrity_checker import IntegrityWindow, run_integrity_window
+
+            include_dimensions = bool(self.config.get("integrity.include_dimensions", False))
+            task_codes = str(self.config.get("integrity.ods_task_codes", "") or "").strip()
+            report = run_integrity_window(
+                cfg=self.config,
+                window=IntegrityWindow(
+                    start=window_start,
+                    end=window_end,
+                    label="etl_window",
+                    granularity="window",
+                ),
+                include_dimensions=include_dimensions,
+                task_codes=task_codes,
+                logger=self.logger,
+                write_report=True,
+            )
+            self.logger.info(
+                "Integrity check done: report=%s missing=%s errors=%s",
+                report.get("report_path"),
+                report.get("api_to_ods", {}).get("total_missing"),
+                report.get("api_to_ods", {}).get("total_errors"),
+            )
+        except Exception as exc:  # noqa: BLE001
+            self.logger.warning("Integrity check failed: %s", exc, exc_info=True)
+
     def close(self):
         """关闭连接。"""
         self.db_conn.close()