修复事件触发智能体时无法传入触发事件的记录的pagetype和name

apps/jingrow/jingrow/ai/pagetype/local_ai_agent/__init__.py

@@ -137,9 +137,12 @@ def enqueue_local_ai_agent(pg, agent_name):
 		from .local_ai_agent import create_agent_job
 
 		# 将触发上下文透传为执行参数的一部分（由下游路由读取）
+		event_inputs = {"pagetype": getattr(pg, 'pagetype', None), "name": getattr(pg, 'name', None)} if (hasattr(pg, 'pagetype') and hasattr(pg, 'name')) else {}
+
 		_ = create_agent_job(
 			agent_id=str(agent_id),
 			agent_name=agent_name_value,
+			inputs=event_inputs,
 			# 使用统一的执行路由；会在本地队列处理器里转发
 			route="jingrow/agents/execute",
 		)

apps/jingrow/jingrow/ai/pagetype/local_ai_agent/local_ai_agent.py

@@ -28,7 +28,8 @@ def create_agent_job(
 	agent_name: Optional[str] = None,
 	session_cookie: Optional[str] = None,
 	route: str = "jingrow/agents/execute",
-	method: Optional[str] = None
+	method: Optional[str] = None,
+	inputs: Optional[Dict[str, Any]] = None
 ) -> str:
 	"""
 	创建智能体执行任务的通用函数
@@ -58,6 +59,10 @@ def create_agent_job(
 		'job_id': job_id,
 		'route': route,
 	}
+
+	# 透传触发时的上下文输入（例如 pagetype/name）
+	if inputs and isinstance(inputs, dict):
+		payload['inputs'] = inputs
 	
 	if method:
 		payload['method'] = method

apps/jingrow/jingrow/ai/pagetype/local_ai_agent/nodes/input_record/input_record.py

@@ -1,7 +1,7 @@
 import json
 import sys
 import os
-
+import jingrow
 # 导入utils
 from jingrow.utils.jingrow_api import get_record
 
@@ -13,7 +13,7 @@ def execute(context=None, inputs=None, config=None, **kwargs):
     if config is None:
         config = kwargs.get("config", {})
  
-    # jingrow.log_error("input_record 节点 inputs汇总", f"inputs={inputs}, config={config}, context={context}")
+    jingrow.log_error(f"inputs={inputs}, config={config}, context={context}")
 
     pagetype = inputs.get("pagetype") if isinstance(inputs, dict) else None
     name = inputs.get("name") if isinstance(inputs, dict) else None

apps/jingrow/jingrow/services/queue.py

@@ -196,8 +196,15 @@ async def _execute_flow_job(payload: Dict[str, Any]) -> Dict[str, Any]:
                 from_id = (ref or {}).get("from")
                 field = (ref or {}).get("field")
                 inputs[key] = (context["node_results"].get(from_id) or {}).get(field)
-            # 可达性判断：无入边（起点）或至少有一条入边被激活
+            # 入口节点合并初始触发入参（对齐 SaaS 端：入口节点直接使用 context["inputs"]）
             incoming_list = incoming_edges_by_target.get(node_id, [])
+            if not incoming_list:
+                initial_inputs = (context or {}).get("inputs") or {}
+                if isinstance(initial_inputs, dict) and initial_inputs:
+                    # 后者优先，保留显式连线映射的结果
+                    merged = {**initial_inputs, **inputs}
+                    inputs = merged
+            # 可达性判断：无入边（起点）或至少有一条入边被激活
             has_activated_incoming = any((e.get("id") in active_edge_ids) for e in incoming_list)
             if incoming_list and not has_activated_incoming:
                 # 暂不可达，留待后轮
@@ -265,10 +272,24 @@ async def _execute_agent_job(payload: Dict[str, Any]) -> Dict[str, Any]:
             nodes = flow_data.get("nodes") or []
             edges = flow_data.get("edges") or []
             
+            # 收集初始触发入参，对齐 SaaS：入口节点可直接取用 context["inputs"]
+            initial_inputs: Dict[str, Any] = {}
+            incoming_inputs = payload.get("inputs") or {}
+            if isinstance(incoming_inputs, dict):
+                initial_inputs.update(incoming_inputs)
+            ctx_inputs = (payload.get("context") or {}).get("inputs")
+            if isinstance(ctx_inputs, dict):
+                initial_inputs.update(ctx_inputs)
+            # 兼容直传 pagetype/name 的场景
+            if payload.get("pagetype") and payload.get("name"):
+                initial_inputs.setdefault("pagetype", payload.get("pagetype"))
+                initial_inputs.setdefault("name", payload.get("name"))
+
             context = {
                 "agent_name": agent.get("agent_name") or agent_id, 
                 "agent_data": agent,
-                "session_cookie": session_cookie
+                "session_cookie": session_cookie,
+                "inputs": initial_inputs,
             }
             result = await _execute_flow_job({"nodes": nodes, "edges": edges, "context": context, "flow_id": agent_id})
             return result

apps/jingrow/jingrow/utils/path.py

@@ -7,7 +7,7 @@ from pathlib import Path
 @lru_cache(maxsize=1)
 def get_root_path() -> Path:
     """返回项目根目录：从 utils/path.py 回退到仓库根 (../../..)。"""
-    return Path(__file__).resolve().parents[3]
+    return Path(__file__).resolve().parents[4]
 
 
 @lru_cache(maxsize=1)