init: Prefect ETL sync project (RIIN/SDS → PG) with 12 sync flows

2026-07-23 08:13:09 +00:00
commit 4cba1da4e1
41 changed files with 6219 additions and 0 deletions
@@ -0,0 +1,77 @@
+"""RIIN 订单同步:按时间窗口(+可选筛选)拉取订单与商品。
+
+对应 riin-order-sync。timeType 固定 3(商家推送时间)。
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from prefect import flow, task
+
+from api.riin.models import OrderPageRequest
+from flows.riin._pipeline import sync_riin_records
+from flows.shared import SyncDB, create_riin_client, fetch_riin_all_pages, format_date
+
+
+@task(retries=3, retry_delay_seconds=[3, 6, 9])
+def extract_orders(params: dict[str, Any]) -> list[Any]:
+    client = create_riin_client()
+    try:
+        return fetch_riin_all_pages(client, params)
+    finally:
+        client.close()
+
+
+@task
+def upsert_records(raw_orders: list[Any]) -> dict[str, int]:
+    client = create_riin_client()
+    try:
+        return sync_riin_records(raw_orders, client, SyncDB())
+    finally:
+        client.close()
+
+
+@flow(name="riin-order-sync", log_prints=True)
+def riin_order_sync(
+    order_date_begin: str,
+    order_date_end: str,
+    platform_type: int | None = None,
+    order_status: int | None = None,
+    factory_ids: list[str] | None = None,
+    express_source: int | None = None,
+    urgent_order: int | None = None,
+    has_refund: bool | None = None,
+    page_size: int = 200,
+) -> dict[str, Any]:
+    if not order_date_begin or not order_date_end:
+        raise ValueError("order_date_begin / order_date_end 为必填")
+
+    params: dict[str, Any] = {
+        "time_type": 3,
+        "start_time": order_date_begin,
+        "end_time": order_date_end,
+        "page_size": page_size,
+    }
+    if platform_type is not None:
+        params["platform_type"] = platform_type
+    if order_status is not None:
+        params["order_status"] = order_status
+    if factory_ids:
+        params["factory_ids"] = factory_ids
+    if express_source is not None:
+        params["express_source"] = express_source
+    if urgent_order is not None:
+        params["urgent_order"] = urgent_order
+    if has_refund is not None:
+        params["has_refund"] = has_refund
+
+    print(f"RIIN 订单同步 {order_date_begin} ~ {order_date_end}")
+    raw = extract_orders(params)
+    stats = upsert_records(raw)
+    print(f"完成: 订单 {stats['orderCount']} / 商品 {stats['goodsCount']} | {stats}")
+    return {"syncRange": {"from": order_date_begin, "to": order_date_end}, **stats}
+
+
+if __name__ == "__main__":
+    riin_order_sync.serve(name="riin-order-sync")