perf: batch scheduling reconciliation and add widget

src/govoplan_scheduling/backend/router.py

@@ -297,6 +297,7 @@ def api_search_scheduling_people(
 @router.get("/requests", response_model=SchedulingRequestListResponse)
 def api_list_scheduling_requests(
     status_filter: str | None = Query(default=None, alias="status"),
+    limit: int = 100,
     session: Session = Depends(get_session),
     principal: ApiPrincipal = Depends(get_api_principal),
 ) -> SchedulingRequestListResponse:
@@ -307,21 +308,11 @@ def api_list_scheduling_requests(
         actor_ids=_principal_actor_ids(principal),
         can_manage=_can_manage_scheduling(principal),
         status=status_filter,
+        limit=limit,
     )
-    actor_ids = _principal_actor_ids(principal)
-    for request in requests:
-        refresh_participant_response_state(
-            session,
-            request=request,
-            actor_ids=actor_ids,
-        )
-    response = SchedulingRequestListResponse(
+    return SchedulingRequestListResponse(
         requests=[_request_response(request, principal=principal) for request in requests]
     )
-    # Poll responses are authoritative, while Scheduling keeps a durable
-    # participant projection used by its task-oriented list.
-    session.commit()
-    return response
 
 
 @router.post("/requests", response_model=SchedulingRequestResponse, status_code=status.HTTP_201_CREATED)
@@ -407,16 +398,9 @@ def api_get_scheduling_request(
             actor_ids=_principal_actor_ids(principal),
             can_manage=_can_manage_scheduling(principal),
         )
-        refresh_participant_response_state(
-            session,
-            request=request,
-            actor_ids=_principal_actor_ids(principal),
-        )
     except SchedulingError as exc:
         raise _scheduling_http_error(exc) from exc
-    response = _request_response(request, principal=principal)
-    session.commit()
-    return response
+    return _request_response(request, principal=principal)
 
 
 @router.patch("/requests/{request_id}", response_model=SchedulingRequestResponse)