Continuing checkpoint after intrupption of a node

[continentalGT]

Hello everyone ,
I need help.

I have created a simple workflow to explain humanin the loop .the workflow is in the ss.
the first node will simply greet the user and then the intrupption which will ask the user to provide the topic on which he wants the poem and then conintuing the state again from where it left.

My issue is that it is not contuing from the check point.it is starting again with the same thread.
can anyone help me.

here is the code

from typing import Annotated, TypedDict
from langgraph.graph.message import add_messages, AnyMessage
from IPython.display import Image, display
from langgraph.graph import StateGraph, START, END
from langchain_core.messages import AIMessage,HumanMessage
from langgraph.checkpoint.memory import MemorySaver

class MessagesState(TypedDict):
messages: Annotated[list[AnyMessage], add_messages]

def greeting(state: MessagesState):
return {
"messages": [
AIMessage(content="Hello, Good day, I will be helping you in creating a shayari on your favorite topic.Please mention your topic",name='Lalita')
]
}

def shayar(state: MessagesState):
sys_msg='you are an experienced shayar produce a shayari for the topic mentioned by user in just 2-4 lines'
return {"messages":
[
llm.invoke([sys_msg]+state["messages"])
]
}

Build graph

graph = StateGraph(MessagesState)
graph.add_node("greeting", greeting)
graph.add_node("shayar", shayar)

graph.add_edge(START, "greeting")
graph.add_edge("greeting", "shayar")
graph.add_edge("shayar", END)

View before compiling

memory = MemorySaver()
graph = graph.compile(interrupt_before=['shayar'],checkpointer=memory) #adding intrrupt before shayar
display(Image(graph.get_graph().draw_mermaid_png()))

Input

user_hi = input()
initial_input = {"messages": [HumanMessage(content=user_hi)]}

Thread

thread = {"configurable": {"thread_id": "1"}}

Run the graph until the first interruption

for event in graph.stream(initial_input, thread, stream_mode="values"):
event['messages'][-1].pretty_print()

Get user feedback

user_preference = input("Please tell me the topic on which u want shayari : ")

topic = {"messages": [HumanMessage(content=user_preference)]}

# If approved, continue the graph execution

for event in graph.stream(topic


, thread, stream_mode="values"):
event['messages'][-1].pretty_print()