Langchain to LangGraph Migration - Need Help in Prompts and other things...

[mraguth]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

My LangGraph working example:

# First we initialize the model we want to use.
from langgraph.prebuilt import create_react_agent
from langgraph.prebuilt.chat_agent_executor import AgentState
from langgraph.errors import GraphRecursionError

model = llm

# For this tutorial we will use custom tool that returns pre-defined values for weather in two cities (NYC & SF)

from typing import Literal

from langchain_core.tools import tool
# Define the graph

from langgraph.prebuilt import create_react_agent

prompt1 = ChatPromptTemplate.from_messages(
    [
        ("system", template),
        ("placeholder", "{messages}"),
        MessagesPlaceholder("chat_history", optional=True),
    ]
)

def _modify_state_messages(state: AgentState):
    return prompt.invoke({"messages": state["messages"]}).to_messages() + [
        ("user", "I am AI Agent.")
    ]

graph = create_react_agent(model, tools=tools,state_modifier=prompt1)

from IPython.display import Image, display

display(Image(graph.get_graph().draw_mermaid_png()))

def print_stream(stream):
    for s in stream:
        message = s["messages"][-1]
        if isinstance(message, tuple):
            print(message)
        else:
            message.pretty_print()

inputs = {"messages": [("user", output_query_str)]}
print_stream(graph.stream(inputs, stream_mode="values"))

Description

I am attempting to migrate my Langchain code to LangGraph and have partially succeeded. However, I need assistance with migrating max_iterations, adding additional prompts such as tune_prompt and full_prompt, and setting allow_dangerous_code=True, as shown in the agent executor below. Is it possible to incorporate these features in LangGraph? Please advise.

My Langchain code (migrated to Langgraph vs. Not migrated to Langgraph) ....

agent_executor = AgentExecutor(
    agent=agent,      # migrated
    tools=tools,         # migrated
    max_iterations=10, # not migrated
    early_stopping_method="force",    # not migrated
    allow_dangerous_code=True, # not migrated
    allow_dangerous_requests=True, # not migrated
    agent_executor_kwargs={"handle_parsing_errors": True}, # not migrated
    tune_prompt = tune_prompt, # not migrated
    full_prompt = full_prompt, # not migrated
    metadata_prompt = metadata_prompt, # not migrated
) | (lambda x: x["output"])

Working LangGraph code with above migrations. I like to migrate others in to this below code:

# First we initialize the model we want to use.
from langgraph.prebuilt import create_react_agent
from langgraph.prebuilt.chat_agent_executor import AgentState
from langgraph.errors import GraphRecursionError

model = llm

# For this tutorial we will use custom tool that returns pre-defined values for weather in two cities (NYC & SF)

from typing import Literal

from langchain_core.tools import tool
# Define the graph

from langgraph.prebuilt import create_react_agent

prompt1 = ChatPromptTemplate.from_messages(
    [
        ("system", template),
        ("placeholder", "{messages}"),
        MessagesPlaceholder("chat_history", optional=True),
    ]
)

def _modify_state_messages(state: AgentState):
    return prompt.invoke({"messages": state["messages"]}).to_messages() + [
        ("user", "I am AI Agent.")
    ]

graph = create_react_agent(model, tools=tools,state_modifier=prompt1)

from IPython.display import Image, display

display(Image(graph.get_graph().draw_mermaid_png()))

def print_stream(stream):
    for s in stream:
        message = s["messages"][-1]
        if isinstance(message, tuple):
            print(message)
        else:
            message.pretty_print()

inputs = {"messages": [("user", output_query_str)]}
print_stream(graph.stream(inputs, stream_mode="values"))

System Info

Name: langgraph
Version: 0.2.34

[mraguth]

Can you help me how I convert this langchain prompt to langgraph prompt? Here I am re-shaping the prompt to make it cleaner. Any help would be appeciated

template = """
You are a helpful assistant tasked with generating multiple sub-questions related to: {question}. Each subsequent query should be derived from the top one.
The goal is to systematically break down the input question into a series of interconnected sub-problems/sub-questions to arrive at a comprehensive answer.
Ensure that each step is logically derived from the previous step, and all steps are interconnected.
output: Please do not create more than 5 steps. Do not write how and why to do steps! just write what LLMs need to do. Output also nice to have a table formatted data along with narration.
"""

prompt_decomposition = ChatPromptTemplate.from_template(template)

# LLM
llm = llm

# Chain
#generate_queries_decomposition = (prompt_decomposition | llm | StrOutputParser() | (lambda x: x.split("\n")))
generate_queries_decomposition = (prompt_decomposition | llm | StrOutputParser() | (lambda x: "\n".join(x.split("\n"))))

# Run - Main Question
#question =question

output_query_str1 = generate_queries_decomposition.invoke({"question":output_query_str})