Add pass variable to prompt on create_react_agent langgraph

[irufano]

i have prompt, tool, and agent

VISUALIZE_PROMPT = """
{question}
Based on question if ask chart/graph do below task else just return "None".

{ai_response}
Based on AI response above, analyze data and extract to dataframe.
Then generate chart/graph based on dataframe code with matplotlib.
The chart/graph result must be in base64.

base64:
"""
# tool
python_repl = PythonREPL()
repl_tool = Tool(
    name="python_repl",
    description="A Python shell. Use this to execute python commands. Input should be a valid python command. If you want to see the output of a value, you should print it out with `print(...)`.",
    func=python_repl.run,
)

# agent
visualize_system_message = SystemMessage(content=VISUALIZE_PROMPT)
visualize_agent_executor = create_react_agent(
    llm, [repl_tool], state_modifier=visualize_system_message
)

# invoke
result = visualize_agent_executor.invoke(
        {"messages": [HumanMessage(content=question)]},  debug=True
    )

print(result)

How to pass value on variable prompt {question} and {ai_response} on create_react_agent langgraph when invoking?

[dosubot[bot]]

Hey @irufano! I'm here to help you with any bugs, questions, or contributions you have. Let's tackle this together!

To pass values to the variables {question} and {ai_response} in the create_react_agent langgraph, you need to provide the question, user_id, and session_id as part of the input to the chain. The ai_response is handled internally by the chain.

Here's how you can pass the question:

from langchain_core.pydantic_v1 import BaseModel

class Question(BaseModel):
    question: str
    user_id: str
    session_id: str

# Create an instance of the Question with your specific question, user_id, and session_id
input_question = Question(question="Your specific question here", user_id="user123", session_id="session456")

# Invoke the chain with the input
result = chain(input_question)

This code snippet shows how to create an instance of the Question class with your specific question, user_id, and session_id, and then invoke the chain with this input. The chain is set up to handle the question input type and manage the ai_response internally ^[1].

To continue talking to Dosu, mention @dosu.

[irufano]

how invoke with create_react_agent?

result = visualize_agent_executor.invoke(
        {"messages": [HumanMessage(content=question)]},  debug=True
    )

[dosubot[bot]]

To invoke the create_react_agent with your setup, you need to pass the question and handle the ai_response internally. Here's how you can do it:

from langchain_core.messages import HumanMessage

# Assuming 'question' is your input question
result = visualize_agent_executor.invoke(
    {
        "messages": [HumanMessage(content=question)]
    },
    debug=True
)

print(result)

This code snippet shows how to pass the question to the agent. The ai_response will be managed by the agent internally ^[1].

To continue talking to Dosu, mention @dosu.

[irufano]

how to pass variable to prompt?

[dosubot[bot]]

To pass a variable to a prompt in the create_react_agent function, use the prompt.partial method. This allows you to set specific variables in the prompt. For example:

prompt = prompt.partial(
    tools=tools_renderer(list(tools)),
    tool_names=", ".join([t.name for t in tools]),
)

This snippet shows how to pass the tools and tool_names variables to the prompt ^[1].

To continue talking to Dosu, mention @dosu.