改 相对路径 完成客户端

etl_billiards/tasks/dws_build_order_summary_task.py

@@ -7,6 +7,7 @@ from datetime import date
 from typing import Any
 
 from .base_task import BaseTask, TaskContext
+from utils.windowing import build_window_segments, calc_window_minutes
 from scripts.build_dws_order_summary import SQL_BUILD_SUMMARY
 
 
@@ -17,37 +18,86 @@ class DwsBuildOrderSummaryTask(BaseTask):
         return "DWS_BUILD_ORDER_SUMMARY"
 
     def execute(self, cursor_data: dict | None = None) -> dict:
-        context = self._build_context(cursor_data)
+        base_context = self._build_context(cursor_data)
         task_code = self.get_task_code()
-        self.logger.info(
-            "%s: start, window[%s ~ %s]",
-            task_code,
-            context.window_start,
-            context.window_end,
+        segments = build_window_segments(
+            self.config,
+            base_context.window_start,
+            base_context.window_end,
+            tz=self.tz,
+            override_only=True,
         )
+        if not segments:
+            segments = [(base_context.window_start, base_context.window_end)]
 
-        try:
-            extracted = self.extract(context)
-            transformed = self.transform(extracted, context)
-            load_result = self.load(transformed, context) or {}
-            self.db.commit()
-        except Exception:
-            self.db.rollback()
-            self.logger.error("%s: failed", task_code, exc_info=True)
-            raise
+        total_segments = len(segments)
+        if total_segments > 1:
+            self.logger.info("%s: 分段执行 共%s段", task_code, total_segments)
 
-        counts = load_result.get("counts") or {}
-        result = {"status": "SUCCESS", "counts": counts}
+        total_counts: dict = {}
+        segment_results: list[dict] = []
+        request_params_list: list[dict] = []
+        total_deleted = 0
+
+        for idx, (window_start, window_end) in enumerate(segments, start=1):
+            context = self._build_context_for_window(window_start, window_end, cursor_data)
+            self.logger.info(
+                "%s: 开始执行(%s/%s), 窗口[%s ~ %s]",
+                task_code,
+                idx,
+                total_segments,
+                context.window_start,
+                context.window_end,
+            )
+
+            try:
+                extracted = self.extract(context)
+                transformed = self.transform(extracted, context)
+                load_result = self.load(transformed, context) or {}
+                self.db.commit()
+            except Exception:
+                self.db.rollback()
+                self.logger.error("%s: 执行失败", task_code, exc_info=True)
+                raise
+
+            counts = load_result.get("counts") or {}
+            self._accumulate_counts(total_counts, counts)
+
+            extra = load_result.get("extra") or {}
+            deleted = int(extra.get("deleted") or 0)
+            total_deleted += deleted
+            request_params = load_result.get("request_params")
+            if request_params:
+                request_params_list.append(request_params)
+
+            if total_segments > 1:
+                segment_results.append(
+                    {
+                        "window": {
+                            "start": context.window_start,
+                            "end": context.window_end,
+                            "minutes": context.window_minutes,
+                        },
+                        "counts": counts,
+                        "extra": extra,
+                    }
+                )
+
+        overall_start = segments[0][0]
+        overall_end = segments[-1][1]
+        result = {"status": "SUCCESS", "counts": total_counts}
         result["window"] = {
-            "start": context.window_start,
-            "end": context.window_end,
-            "minutes": context.window_minutes,
+            "start": overall_start,
+            "end": overall_end,
+            "minutes": calc_window_minutes(overall_start, overall_end),
         }
-        if "request_params" in load_result:
-            result["request_params"] = load_result["request_params"]
-        if "extra" in load_result:
-            result["extra"] = load_result["extra"]
-        self.logger.info("%s: done, counts=%s", task_code, counts)
+        if segment_results:
+            result["segments"] = segment_results
+        if request_params_list:
+            result["request_params"] = request_params_list[0] if len(request_params_list) == 1 else request_params_list
+        if total_deleted:
+            result["extra"] = {"deleted": total_deleted}
+        self.logger.info("%s: 完成, 统计=%s", task_code, total_counts)
         return result
 
     def extract(self, context: TaskContext) -> dict[str, Any]:
@@ -101,7 +151,7 @@ class DwsBuildOrderSummaryTask(BaseTask):
             if extracted["delete_before_insert"]:
                 if extracted["full_refresh"] and extracted["site_id"] is None:
                     cur.execute("TRUNCATE TABLE billiards_dws.dws_order_summary;")
-                    self.logger.info("DWS_BUILD_ORDER_SUMMARY: truncated billiards_dws.dws_order_summary")
+                    self.logger.info("DWS订单汇总: 已清空 billiards_dws.dws_order_summary")
                 else:
                     delete_sql = "DELETE FROM billiards_dws.dws_order_summary WHERE 1=1"
                     delete_args: list[Any] = []
@@ -116,7 +166,7 @@ class DwsBuildOrderSummaryTask(BaseTask):
                         delete_args.append(_as_date(extracted["end_date"]))
                     cur.execute(delete_sql, delete_args)
                     deleted = cur.rowcount
-                    self.logger.info("DWS_BUILD_ORDER_SUMMARY: deleted=%s sql=%s", deleted, delete_sql)
+                    self.logger.info("DWS订单汇总: 删除=%s 语句=%s", deleted, delete_sql)
 
             cur.execute(SQL_BUILD_SUMMARY, sql_params)
             affected = cur.rowcount