refactor(db)!: clean up business SQL access

- make realtime replacement and analysis result writes transactional\n- consolidate SCADA repositories and remove process-global project state\n- validate SCADA batches and use indexed GIS-backed business queries\n\nBREAKING CHANGE: remove the public analysis result writer and the pipeline-health network_name query parameter.

app/infra/db/timescaledb/internal_queries.py

@@ -88,16 +88,15 @@ class InternalQueries:
                     rows = ScadaRepository.get_scada_by_ids_time_range_sync(
                         conn, device_ids, start_time, end_time
                     )
-                    # 处理结果，返回每个 device_id 的第一个值
-                    result = {}
-                    for device_id in device_ids:
-                        device_rows = [
-                            row for row in rows if row["device_id"] == device_id
-                        ]
-                        if device_rows:
-                            result[device_id] = device_rows[0]["monitored_value"]
-                        else:
-                            result[device_id] = None
+                    # Rows are ordered by device/time; retain the first sample
+                    # for each requested device in one pass.
+                    result = {device_id: None for device_id in device_ids}
+                    seen: set[str] = set()
+                    for row in rows:
+                        device_id = str(row["device_id"])
+                        if device_id in result and device_id not in seen:
+                            result[device_id] = row["monitored_value"]
+                            seen.add(device_id)
                     return result
             except Exception as e:
                 logger.error(f"查询尝试 {attempt + 1} 失败: {e}")
@@ -135,8 +134,6 @@ class InternalQueries:
                         result.setdefault(device_id, []).append(
                             {"time": row["time"].isoformat(), "value": value}
                         )
-                    for device_id in result:
-                        result[device_id].sort(key=lambda item: item["time"])
                     return result
             except Exception as e:
                 logger.error(f"查询尝试 {attempt + 1} 失败: {e}")
@@ -307,16 +304,7 @@ class InternalQueries:
     def _resolve_simulation_table(element_type: str) -> tuple[str, str, set[str]]:
         normalized_type = element_type.lower()
         if normalized_type == "node":
-            return "node_results", "node_id", {"actual_demand", "total_head", "pressure", "quality"}
+            return "node_results", "node_id", set(RealtimeRepository.NODE_FIELDS)
         if normalized_type == "link":
-            return "link_results", "link_id", {
-                "flow",
-                "friction",
-                "headloss",
-                "quality",
-                "reaction",
-                "setting",
-                "status",
-                "velocity",
-            }
+            return "link_results", "link_id", set(RealtimeRepository.LINK_FIELDS)
         raise ValueError(f"Unsupported element_type: {element_type}")