Add main.py

main.py

@@ -1,16 +1,13 @@
 import os
+import re
 import asyncio
 from typing import TypedDict, Annotated
 
 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
 
-# ---------- LLM ----------
+# LLM configuration (OpenRouter)
 llm = ChatOpenAI(
     model="openai/gpt-oss-20b:free",
     base_url="https://openrouter.ai/api/v1",
@@ -18,91 +15,78 @@ llm = ChatOpenAI(
     temperature=0.0,
 )
 
-# ---------- Backend for deepagents (not used directly in graph but required by create_deep_agent) ----------
-backend = CompositeBackend([
-    LocalShellBackend(workspace_dir="./workspace"),
-    FilesystemBackend(),
-])
-
 # ---------- State definition ----------
 class ReflectState(TypedDict):
     question: str
     draft: str
     critique: str
-    verdict: str  # "ok" | "needs_revision"
+    verdict: str  # ok | needs_revision
     round: int
     max_rounds: int
 
-# ---------- Nodes ----------
+# ---------- Node implementations ----------
-async def draft_answer(state: ReflectState) -> ReflectState:
+async def draft_answer(state: ReflectState) -> dict:
-    prompt = f"Write a concise answer (5–10 sentences) to the following question:\n\n{state['question']}"
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
-    state["draft"] = response.content.strip()
-    return state
-
-async def reflect(state: ReflectState) -> ReflectState:
     prompt = (
-        f"You are a critic. Evaluate the following draft answer for completeness, specificity, and lack of filler.\n\nDraft:\n{state['draft']}\n\nProvide a verdict (ok or needs_revision) and 2–3 bullet points of critique."
+        f"Write a concise answer (5–10 sentences) to the following question:\n\n"
+        f"Question: {state['question']}"
     )
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
+    response = await llm.ainvoke([{"role": "user", "content": prompt}])
+    draft = response.content.strip()
+    return {"draft": draft, "round": 0}
+
+async def reflect(state: ReflectState) -> dict:
+    prompt = (
+        f"You are a critical reviewer. Evaluate the following draft answer for completeness, specificity, and lack of filler.\n\n"
+        f"Draft: {state['draft']}\n\n"
+        f"Provide a verdict (ok or needs_revision) and 2–3 bullet points of critique."
+    )
+    response = await llm.ainvoke([{"role": "user", "content": prompt}])
     text = response.content.strip()
-    # Simple parsing: first line verdict, rest critique
+    verdict_match = re.search(r"(ok|needs_revision)", text, re.IGNORECASE)
-    lines = text.splitlines()
+    verdict = verdict_match.group(1).lower() if verdict_match else "needs_revision"
-    verdict_line = lines[0].lower()
+    return {"critique": text, "verdict": verdict}
-    verdict = "ok" if "ok" in verdict_line else "needs_revision"
-    critique = "\n".join(lines[1:]).strip()
-    state["verdict"] = verdict
-    state["critique"] = critique
-    return state
 
-async def rewrite(state: ReflectState) -> ReflectState:
+async def rewrite(state: ReflectState) -> dict:
     prompt = (
-        f"Rewrite the draft answer taking into account the following critique. Keep the answer concise (5–10 sentences).\n\nCritique:\n{state['critique']}\n\nOriginal Draft:\n{state['draft']}"
+        f"Rewrite the draft answer incorporating the following critique. Keep the answer concise (5–10 sentences).\n\n"
+        f"Critique: {state['critique']}\n\n"
+        f"Original Draft: {state['draft']}"
     )
-    response = await llm.ainvoke([HumanMessage(content=prompt)])
+    response = await llm.ainvoke([{"role": "user", "content": prompt}])
-    state["draft"] = response.content.strip()
+    new_draft = response.content.strip()
-    state["round"] += 1
+    return {"draft": new_draft, "round": state['round'] + 1}
-    return state
 
-# ---------- Graph ----------
+# ---------- Graph construction ----------
-graph = StateGraph(ReflectState)
+builder = StateGraph(ReflectState)
-graph.add_node("draft_answer", draft_answer)
+builder.add_node("draft_answer", draft_answer)
-graph.add_node("reflect", reflect)
+builder.add_node("reflect", reflect)
-graph.add_node("rewrite", rewrite)
+builder.add_node("rewrite", rewrite)
 
-# Entry point
+builder.set_entry_point("draft_answer")
-graph.set_entry_point("draft_answer")
+builder.add_edge("draft_answer", "reflect")
+builder.add_conditional_edges(
+    "reflect",
+    lambda x: x["verdict"],
+    {
+        "ok": END,
+        "needs_revision": "rewrite",
+    },
+)
+builder.add_edge("rewrite", "reflect")
 
-# Transition logic
+# Limit rounds
-# After draft_answer -> reflect
+async def limit_rounds(state: ReflectState) -> str:
-graph.add_edge("draft_answer", "reflect")
+    if state["round"] >= state["max_rounds"] and state["verdict"] == "needs_revision":
-# After reflect
+        return END
-# if ok -> END
+    return "reflect"
-# if needs_revision and round < max_rounds -> rewrite
-# else -> END
 
-def reflect_conditional(state: ReflectState):
+builder.add_conditional_edges("rewrite", limit_rounds, {"reflect": "reflect", END: END})
-    if state["verdict"] == "ok":
-        return "END"
-    if state["round"] < state["max_rounds"]:
-        return "rewrite"
-    return "END"
 
-graph.add_conditional_edges("reflect", reflect_conditional, {
+graph = builder.compile()
-    "rewrite": "rewrite",
-    "END": "END",
-})
 
-# After rewrite -> reflect
+# ---------- Demo execution ----------
-graph.add_edge("rewrite", "reflect")
+async def main():
-
+    question = "Объясни студенту разницу между tool и resource в MCP."
-app = graph.compile()
-
-# ---------- DeepAgent wrapper ----------
-# The deepagent will simply forward the human message to the graph and return the final draft.
-@tool
-def run_graph(question: str) -> str:
-    """Run the reflection graph for a given question."""
     initial_state: ReflectState = {
         "question": question,
         "draft": "",
@@ -111,23 +95,8 @@ def run_graph(question: str) -> str:
         "round": 0,
         "max_rounds": 2,
     }
-    result = app.invoke(initial_state)
+    final_state = await graph.ainvoke(initial_state)
-    return result["draft"]
+    print("\nFinal Answer:\n", final_state["draft"])
-
-agent = create_deep_agent(
-    model=llm,
-    tools=[run_graph],
-    backend=backend,
-    system_prompt="You are an assistant that can answer questions using a self‑checking process.",
-)
-
-async def main():
-    question = "Объясни студенту разницу между tool и resource в MCP."
-    response = await agent.ainvoke(
-        {"messages": [HumanMessage(content=question)]},
-        {"configurable": {"thread_id": "session-1"}},
-    )
-    print("Final answer:\n", response["messages"][-1].content)
 
 if __name__ == "__main__":
     asyncio.run(main())