Merge pull request #20 from ks6088ts-labs/feature/issue-19_human-in-the-loop

human in the loop

Author: ks6088ts

Makefile

@@ -114,4 +114,4 @@ ci-test-docs: install-deps-docs docs ## run CI test for documentation
 
 .PHONY: langgraph-studio
 langgraph-studio: ## run LangGraph Studio
-	uv run langgraph dev --no-reload
+	uv run langgraph dev

docs/index.md

@@ -25,9 +25,9 @@ uv run python -m template_langgraph.tasks.search_documents_on_elasticsearch
 uv run python -m template_langgraph.tasks.run_kabuto_helpdesk_agent "KABUTOの起動時に、画面全体が紫色に点滅し、システムがフリーズします。"
 uv run python -m template_langgraph.tasks.run_kabuto_helpdesk_agent "KABUTOのマニュアルから禅モードに関する情報を教えて下さい"
 
-# BasicWorkflowAgent
-uv run python -m template_langgraph.tasks.draw_basic_workflow_agent_mermaid_png "data/basic_workflow_agent.png"
-uv run python -m template_langgraph.tasks.run_basic_workflow_agent
+# ChatWithToolsAgent
+uv run python -m template_langgraph.tasks.draw_chat_with_tools_agent_mermaid_png "data/chat_with_tools_agent.png"
+uv run python -m template_langgraph.tasks.run_chat_with_tools_agent
 # KABUTOの起動時に、画面全体が紫色に点滅し、システムがフリーズします。KABUTO のマニュアルから、関連する情報を取得したり過去のシステムのトラブルシュート事例が蓄積されたデータベースから、関連する情報を取得して質問に答えてください
 # 天狗のいたずら という現象について KABUTO のマニュアルから、関連する情報を取得したり過去のシステムのトラブルシュート事例が蓄積されたデータベースから、関連する情報を取得して質問に答えてください
 
@@ -42,6 +42,7 @@ uv run python -m template_langgraph.tasks.run_issue_formatter_agent
 ### LangGraph
 
 - [Build a custom workflow](https://langchain-ai.github.io/langgraph/concepts/why-langgraph/)
+- [LangGraphの(LLMなし)Human-in-the-loopを試してみた](https://qiita.com/te_yama/items/db38201af60dec76384d)
 
 ### Sample Codes
 

langgraph.json

@@ -3,9 +3,10 @@
     "."
   ],
   "graphs": {
-    "basic_workflow_agent": "template_langgraph.agents.basic_workflow_agent.agent:graph",
+    "chat_with_tools_agent": "template_langgraph.agents.chat_with_tools_agent.agent:graph",
     "kabuto_helpdesk_agent": "template_langgraph.agents.kabuto_helpdesk_agent:graph",
-    "issue_formatter_agent": "template_langgraph.agents.issue_formatter_agent.agent:graph"
+    "issue_formatter_agent": "template_langgraph.agents.issue_formatter_agent.agent:graph",
+    "task_decomposer_agent": "template_langgraph.agents.task_decomposer_agent.agent:graph"
   },
   "env": ".env"
 }

template_langgraph/agents/basic_workflow_agent/__init__.py renamed to template_langgraph/agents/chat_with_tools_agent/__init__.py

@@ -3,7 +3,7 @@
 from langchain_core.messages import ToolMessage
 from langgraph.graph import END, StateGraph
 
-from template_langgraph.agents.basic_workflow_agent.models import AgentState
+from template_langgraph.agents.chat_with_tools_agent.models import AgentState
 from template_langgraph.llms.azure_openais import AzureOpenAiWrapper
 from template_langgraph.loggers import get_logger
 from template_langgraph.tools.elasticsearch_tool import search_elasticsearch
@@ -36,7 +36,7 @@ def __call__(self, inputs: dict):
         return {"messages": outputs}
 
 
-class BasicWorkflowAgent:
+class ChatWithToolsAgent:
     def __init__(self):
         self.llm = AzureOpenAiWrapper().chat_model
 
@@ -80,7 +80,7 @@ def create_graph(self):
 
     def initialize(self, state: AgentState) -> AgentState:
         """Initialize the agent with the given state."""
-        logger.info(f"Initializing BasicWorkflowAgent with state: {state}")
+        logger.info(f"Initializing ChatWithToolsAgent with state: {state}")
         # Here you can add any initialization logic if needed
         return state
 
@@ -119,7 +119,7 @@ def route_tools(
 
     def finalize(self, state: AgentState) -> AgentState:
         """Finalize the agent's work and prepare the output."""
-        logger.info(f"Finalizing BasicWorkflowAgent with state: {state}")
+        logger.info(f"Finalizing ChatWithToolsAgent with state: {state}")
         # Here you can add any finalization logic if needed
         return state
 
@@ -128,4 +128,4 @@ def draw_mermaid_png(self) -> bytes:
         return self.create_graph().get_graph().draw_mermaid_png()
 
 
-graph = BasicWorkflowAgent().create_graph()
+graph = ChatWithToolsAgent().create_graph()

template_langgraph/agents/basic_workflow_agent/agent.py renamed to template_langgraph/agents/chat_with_tools_agent/agent.py

@@ -0,0 +1,80 @@
+from langgraph.graph import END, StateGraph
+from langgraph.types import interrupt
+
+from template_langgraph.agents.task_decomposer_agent.models import AgentState, TaskList
+from template_langgraph.llms.azure_openais import AzureOpenAiWrapper
+from template_langgraph.loggers import get_logger
+
+logger = get_logger(__name__)
+
+
+class TaskDecomposerAgent:
+    def __init__(self):
+        self.llm = AzureOpenAiWrapper().chat_model
+
+    def create_graph(self):
+        """Create the main graph for the agent."""
+        # Create the workflow state graph
+        workflow = StateGraph(AgentState)
+
+        # Create nodes
+        workflow.add_node("chat", self.chat)
+        workflow.add_node("human_feedback", self.human_feedback)
+
+        # Create edges
+        workflow.set_entry_point("chat")
+        workflow.add_edge("chat", "human_feedback")
+        workflow.add_conditional_edges(
+            source="human_feedback",
+            path=self.route_human_feedback,
+            path_map={
+                "loopback": "chat",
+                "end": END,
+            },
+        )
+        return workflow.compile()
+
+    def chat(self, state: AgentState) -> AgentState:
+        """Chat with tools using the state."""
+        logger.info(f"Chatting with tools using state: {state}")
+
+        task_list = self.llm.with_structured_output(TaskList).invoke(
+            input=state["messages"],
+        )
+        state["task_list"] = task_list
+        logger.info(f"Decomposed tasks: {task_list}")
+        return state
+
+    def human_feedback(self, state: AgentState) -> AgentState:
+        """Handle human feedback."""
+        logger.info(f"Handling human feedback with state: {state}")
+        feedback = interrupt("Type your feedback. If you want to end the conversation, type 'end'.")
+        state["messages"].append(
+            {
+                "content": feedback,
+                "role": "user",
+            }
+        )
+        return state
+
+    def route_human_feedback(
+        self,
+        state: AgentState,
+    ):
+        """
+        Use in the conditional_edge to route to the HumanFeedbackNode if the last message
+        has human feedback. Otherwise, route to the end.
+        """
+        human_feedback = state["messages"][-1].content.strip().lower()
+        if human_feedback == "end":
+            logger.info("Ending the conversation as per user request.")
+            return "end"
+        logger.info("Looping back to chat for further processing.")
+        return "loopback"
+
+    def draw_mermaid_png(self) -> bytes:
+        """Draw the graph in Mermaid format."""
+        return self.create_graph().get_graph().draw_mermaid_png()
+
+
+graph = TaskDecomposerAgent().create_graph()

template_langgraph/agents/basic_workflow_agent/models.py renamed to template_langgraph/agents/chat_with_tools_agent/models.py

@@ -0,0 +1,36 @@
+from collections.abc import Sequence
+from typing import (
+    Annotated,
+    TypedDict,
+)
+
+from langchain_core.messages import (
+    BaseMessage,
+)
+from langgraph.graph.message import add_messages
+from pydantic import BaseModel, Field
+
+
+class Task(BaseModel):
+    title: str = Field(..., description="Title of the task")
+    description: str = Field(..., description="Description of the task")
+    priority: int = Field(..., description="Priority of the task (1-5)")
+    due_date: str | None = Field(None, description="Due date of the task (YYYY-MM-DD format)")
+    assigned_to: str | None = Field(None, description="Name of the agent assigned to the task")
+
+
+class TaskList(BaseModel):
+    tasks: Sequence[Task] = Field(..., description="List of tasks to be decomposed")
+
+
+class AgentInput(BaseModel):
+    request: str = Field(..., description="Request from the user")
+
+
+class AgentOutput(BaseModel):
+    response: str = Field(..., description="Response from the agent")
+
+
+class AgentState(TypedDict):
+    messages: Annotated[Sequence[BaseMessage], add_messages]
+    task_list: TaskList