Update utils/langgraph_pipeline.py

utils/langgraph_pipeline.py

@@ -8,15 +8,13 @@ from agents import (
     quality_assurance_agent,
 )
 from langchain_core.messages import HumanMessage, AIMessage
-from langchain_core.messages.base import BaseMessage
 from typing import TypedDict, List
-
+from langchain_core.messages.base import BaseMessage
 
 class InputState(TypedDict):
     messages: List[BaseMessage]
     chat_log: list
 
-
 class OutputState(TypedDict):
     pm_output: str
     proj_output: str
@@ -25,10 +23,10 @@ class OutputState(TypedDict):
     qa_output: str
     chat_log: list
 
-
-# ✅ DO NOT USE @tool here — this bypasses LangChain's strict typing enforcement
 def bridge_to_product_manager(state: dict) -> dict:
-    """Convert HumanMessage into structured AIMessage for Product Manager agent."""
+    """
+    Convert HumanMessage into structured AIMessage for Product Manager agent.
+    """
     user_prompt = state["messages"][-1].content
 
     system_spec = f"""# Stakeholder Prompt
@@ -45,17 +43,16 @@ Please analyze and generate a structured product specification including goals,
         "chat_log": state["chat_log"] + [{"role": "System", "content": system_spec}],
     }
 
+# Define the bridge node as a simple function node
+def bridge_node(state: dict) -> dict:
+    return bridge_to_product_manager(state)
 
-# ✅ Use ToolNode.from_function to avoid StructuredTool issues
-bridge_node = ToolNode.from_function(bridge_to_product_manager)
 pm_node = ToolNode([product_manager_agent.run])
 proj_node = ToolNode([project_manager_agent.run])
 arch_node = ToolNode([software_architect_agent.run])
 dev_node = ToolNode([software_engineer_agent.run])
 qa_node = ToolNode([quality_assurance_agent.run])
 
-
-# LangGraph setup
 graph = StateGraph(input=InputState, output=OutputState)
 
 graph.add_node("Bridge", bridge_node)
@@ -75,13 +72,10 @@ graph.add_edge("QualityAssurance", END)
 
 compiled_graph = graph.compile()
 
-
 def run_pipeline_and_save(prompt: str):
     initial_state = {
         "messages": [HumanMessage(content=prompt)],
         "chat_log": [],
     }
-
     final_state = compiled_graph.invoke(initial_state)
-
-    return final_state["chat_log"], final_state["qa_output"]
+    return final_state["chat_log"], final_state["qa_output"]