lg:node

lg:node - Node Management

Manage nodes in LangGraph graphs including creation, connection, and modification of various node types.

Workflow Steps

Add Node

1. Validate Input

   • Check project structure
   • Verify node name is unique
   • Validate node type

2. Generate Node Code

   • Create node function/class
   • Setup node configuration
   • Add error handling
   • Configure retries (if enabled)

3. Update Graph

   • Add node to graph
   • Configure interrupts (if specified)
   • Update state schema (if needed)

4. Generate Tests

   • Create node test file
   • Add unit tests
   • Add integration tests

5. Update Documentation

   • Add node to README
   • Document node behavior
   • Update architecture diagram

Connect Nodes

1. Validate Connection

   • Check source and target exist
   • Verify no circular dependencies (for non-conditional edges)
   • Validate edge type

2. Add Edge

   • Add regular edge or conditional edge
   • Update routing logic
   • Configure edge conditions

3. Update Tests

   • Add edge tests
   • Test routing logic

4. Update Documentation

   • Update graph visualization
   • Document flow

Remove Node

1. Analyze Dependencies

   • Find all edges to/from node
   • Check for dependent nodes
   • Identify routing impacts

2. Remove Components

   • Remove edges
   • Remove node from graph
   • Remove node file
   • Clean up routing

3. Update Tests

   • Remove node tests
   • Update integration tests

4. Update Documentation

   • Remove from README
   • Update diagrams

Node Types

LLM Node

Language model processing node.

from langchain_anthropic import ChatAnthropic
from langchain_core.messages import SystemMessage

def llm_node(state: State):
    """LLM processing node."""
    llm = ChatAnthropic(model="claude-sonnet-5")

    system_prompt = "You are a helpful assistant."
    messages = [SystemMessage(content=system_prompt)] + state["messages"]

    response = llm.invoke(messages)
    return {"messages": [response]}