[question] How to perform tracing when using langgraph dev

[jacy1101]

How to perform tracing when using langgraph dev

deep_researcher = deep_researcher_builder.compile()
# Create the OpikTracer
opik_tracer = OpikTracer(graph=deep_researcher.get_graph(xray=True))

use

uvx --refresh --from "langgraph-cli[inmem]" --with-editable . --python 3.11 langgraph dev --allow-blocking

[YarivHashaiComet]

Hi @jacy1101 - thanks for bringing this up - can you please share some more context of what you are trying to do? which issues did you faced and if possible the full snippet code you trying to run?

[alexkuzmik]

Hi @jacy1101! Would be nice to see a script.
But I assume in your application code you should have a place where you call deep_researcher.invoke or deep_researcher.stream. If that's the case - you should just pass the opik tracer like that via the parameter config={"callbacks": [opik_tracer]}

[jacy1101]

Hi @jacy1101 - thanks for bringing this up - can you please share some more context of what you are trying to do? which issues did you faced and if possible the full snippet code you trying to run?

@YarivHashaiComet like this.

from typing import Annotated

from langchain.chat_models import init_chat_model
from typing_extensions import TypedDict

from langgraph.graph import StateGraph, START, END
from langgraph.graph.message import add_messages


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


graph_builder = StateGraph(State)


llm = init_chat_model("anthropic:claude-3-5-sonnet-latest")


def chatbot(state: State):
    return {"messages": [llm.invoke(state["messages"])]}


# The first argument is the unique node name
# The second argument is the function or object that will be called whenever
# the node is used.
graph_builder.add_node("chatbot", chatbot)
graph_builder.add_edge(START, "chatbot")
graph_builder.add_edge("chatbot", END)
graph = graph_builder.compile()

Then use langgraph dev to start it and use it in langgraph studio.

[jacy1101]

Hi @jacy1101! Would be nice to see a script. But I assume in your application code you should have a place where you call deep_researcher.invoke or deep_researcher.stream. If that's the case - you should just pass the opik tracer like that via the parameter config={"callbacks": [opik_tracer]}

No, I don't have any calls to deep_researcher.invoke or deep_researcher.stream.

[Lothiraldan]

Hi @jacy1101, we have investigated the issue and found the code that's calling the agent deep inside langgraph. We will continue looking into ways to enable tracing when using langgraph dev

[Lothiraldan]

Hi @jacy1101, we identified a potential solution for performing tracing when using langgraph dev, could you try the following agent example?

from typing import Annotated

from langchain.chat_models import init_chat_model
from typing_extensions import TypedDict

from langgraph.graph import StateGraph, START, END
from langgraph.graph.message import add_messages


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


graph_builder = StateGraph(State)


llm = init_chat_model("anthropic:claude-haiku-4-5-20251001")


def chatbot(state: State):
    return {"messages": [llm.invoke(state["messages"])]}


# The first argument is the unique node name
# The second argument is the function or object that will be called whenever
# the node is used.
graph_builder.add_node("chatbot", chatbot)
graph_builder.add_edge(START, "chatbot")
graph_builder.add_edge("chatbot", END)

graph = graph_builder.compile()

# New code
from opik.integrations.langchain import OpikTracer
opik_tracer = OpikTracer(graph=graph.get_graph(xray=True))

# Inject the callback in the graph default config
if graph.config is None:
    graph.config = {}
graph.config.setdefault("callbacks", []).append(opik_tracer)