fix: main.py

main.py

@@ -6,9 +6,7 @@ from langchain_openai import ChatOpenAI
 from langchain_core.messages import HumanMessage
 from langchain.tools import tool
 from deepagents import create_deep_agent
-from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
-from langgraph.graph import StateGraph, START, END
-from langgraph.graph.message import add_messages
+from deepagents.backends import CompositeBackend, LocalShellBackend, FilesystemBackend
 
 # ---------- LLM ----------
 llm = ChatOpenAI(
@@ -18,13 +16,13 @@ llm = ChatOpenAI(
     temperature=0.0,
 )
 
-# ---------- Backend for deepagents ----------
+# ---------- Backend ----------
 backend = CompositeBackend([
     LocalShellBackend(workspace_dir="./workspace"),
     FilesystemBackend(),
 ])
 
-# ---------- State definition ----------
+# ---------- State ----------
 class ReflectState(TypedDict):
     question: str
     draft: str
@@ -33,40 +31,58 @@ class ReflectState(TypedDict):
     round: int
     max_rounds: int
 
-# ---------- LangGraph nodes ----------
+# ---------- Nodes ----------
+from langgraph.graph import StateGraph, START, END
+from langgraph.graph.message import add_messages
+
+# Helper to format the prompt for each node
+DRAFT_PROMPT = """Write a concise answer (5–10 sentences) to the following question:
+
+{question}
+"""
+
+REFLECT_PROMPT = """You are a critic. Given the draft answer below, evaluate its completeness, specificity, and absence of filler. Respond with:
+1. verdict: either "ok" or "needs_revision"
+2. critique: 2–3 bullet points explaining what to improve (if any)
+
+Draft:
+{draft}
+"""
+
+REWRITE_PROMPT = """You are revising the draft answer based on the critique. Produce a new draft that addresses the points. Keep the answer concise (5–10 sentences).
+
+Critique:
+{critique}
+
+Previous draft:
+{draft}
+"""
+
+# Node functions
 async def draft_answer(state: ReflectState) -> ReflectState:
-    prompt = f"Write a concise answer (5–10 sentences) to the following question:\n\n{state['question']}"
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
+    response = await llm.ainvoke([HumanMessage(content=DRAFT_PROMPT.format(question=state["question"]))])
     state["draft"] = response.content.strip()
     return state
 
 async def reflect(state: ReflectState) -> ReflectState:
-    prompt = (
-        "You are a critic. Evaluate the draft answer for completeness, specificity, and lack of filler.\n"
-        "Return a verdict ('ok' or 'needs_revision') and 2–3 bullet points of critique.\n"
-        f"Draft: {state['draft']}"
-    )
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
-    # Simple parsing: first line verdict, rest critique
-    lines = response.content.strip().splitlines()
-    verdict = lines[0].strip().lower()
-    critique = "\n".join(lines[1:]).strip()
+    response = await llm.ainvoke([HumanMessage(content=REFLECT_PROMPT.format(draft=state["draft"]))])
+    # Parse verdict and critique
+    text = response.content.strip()
+    verdict_line = next((l for l in text.splitlines() if l.lower().startswith("verdict:")), "")
+    critique_lines = [l for l in text.splitlines() if l.startswith("-") or l.startswith("•")]
+    verdict = verdict_line.split(":",1)[1].strip().lower() if verdict_line else "needs_revision"
+    critique = "\n".join(critique_lines) if critique_lines else ""
     state["verdict"] = verdict
     state["critique"] = critique
     return state
 
 async def rewrite(state: ReflectState) -> ReflectState:
-    prompt = (
-        "Rewrite the draft answer incorporating the following critique. Keep the answer concise (5–10 sentences).\n"
-        f"Critique: {state['critique']}\n"
-        f"Original draft: {state['draft']}"
-    )
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
+    response = await llm.ainvoke([HumanMessage(content=REWRITE_PROMPT.format(critique=state["critique"], draft=state["draft"]))])
     state["draft"] = response.content.strip()
     state["round"] += 1
     return state
 
-# ---------- Graph construction ----------
+# ---------- Graph ----------
 graph = StateGraph(ReflectState)
 graph.add_node("draft_answer", draft_answer)
 graph.add_node("reflect", reflect)
@@ -76,24 +92,30 @@ graph.add_node("rewrite", rewrite)
 graph.set_entry_point("draft_answer")
 
 # Transitions
+# After draft -> reflect
 graph.add_edge("draft_answer", "reflect")
-# From reflect: if ok -> END, else if needs_revision and round < max_rounds -> rewrite
-graph.add_conditional_edges(
-    "reflect",
-    lambda state: (
-        "END" if state["verdict"] == "ok" else (
-            "rewrite" if state["round"] < state["max_rounds"] else "END"
-        )
-    ),
-)
+# After reflect
+# if ok -> END
+# if needs_revision and round < max_rounds -> rewrite
+# else -> END
+
+def reflect_conditional(state: ReflectState):
+    if state["verdict"] == "ok":
+        return "END"
+    if state["round"] < state["max_rounds"]:
+        return "rewrite"
+    return "END"
+
+graph.add_conditional_edges("reflect", reflect_conditional, {"rewrite": "rewrite", "END": "END"})
+# After rewrite -> reflect
 graph.add_edge("rewrite", "reflect")
 
-app = graph.compile()
+graph.compile()
 
 # ---------- DeepAgent wrapper ----------
 @tool
-def run_graph(question: str, max_rounds: int = 2) -> str:
-    """Run the reflection graph for a given question."""
+def run_reflect_graph(question: str, max_rounds: int = 2) -> str:
+    """Run the reflection graph and return the final draft."""
     initial_state: ReflectState = {
         "question": question,
         "draft": "",
@@ -102,20 +124,20 @@ def run_graph(question: str, max_rounds: int = 2) -> str:
         "round": 0,
         "max_rounds": max_rounds,
     }
-    result = app.invoke(initial_state)
+    result = graph.invoke(initial_state)
     return result["draft"]
 
 agent = create_deep_agent(
     model=llm,
-    tools=[run_graph],
+    tools=[run_reflect_graph],
     backend=backend,
-    system_prompt="You are an assistant that can generate and improve answers using reflection. Use the provided tool to get a refined answer.",
+    system_prompt="You are an assistant that can answer questions and self‑critique using the provided tool.",
 )
 
 async def main():
-    question = "Объясни студенту разницу между tool и resource в MCP"
+    question = "Объясни студенту разницу между tool и resource в MCP."
     response = await agent.ainvoke(
-        {"messages": [HumanMessage(content=question)]},
+        {"messages": [HumanMessage(content=f"Please answer: {question}")]},
         {"configurable": {"thread_id": "session-1"}},
     )
     print("Final answer:\n", response["messages"][-1].content)