commit

chore: remove unused `template-lock` stuff

.env

@@ -0,0 +1,9 @@
+# （可选但推荐）设置您想要发送 traces 的项目名称，例如 'my-awesome-project'
+# To separate your traces from other application
+LANGSMITH_PROJECT=new-agent
+
+# Add API keys for connecting to LLM providers, data sources, and other integrations here
+LANGSMITH_API_KEY=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
+
+# 通常还需要设置 LangChain 的跟踪端点，默认就是以下地址，通常无需修改
+LANGCHAIN_TRACING_V2=true

.gitignore

@@ -162,3 +162,5 @@ cython_debug/
 #.idea/
 uv.lock
 .langgraph_api/
+
+.idea

README.md

@@ -15,14 +15,6 @@ You can extend this graph to orchestrate more complex agentic workflows that can
 
 ## Getting Started
 
-<!--
-Setup instruction auto-generated by `langgraph template lock`. DO NOT EDIT MANUALLY.
--->
-
-<!--
-End setup instructions
--->
-
 1. Install dependencies, along with the [LangGraph CLI](https://langchain-ai.github.io/langgraph/concepts/langgraph_cli/), which will be used to run the server.
 
 ```bash
@@ -53,7 +45,7 @@ For more information on getting started with LangGraph Server, [see here](https:
 
 ## How to customize
 
-1. **Define configurable parameters**: Modify the `Configuration` class in the `graph.py` file to expose the arguments you want to configure. For example, in a chatbot application you may want to define a dynamic system prompt or LLM to use. For more information on configurations in LangGraph, [see here](https://langchain-ai.github.io/langgraph/concepts/low_level/?h=configuration#configuration).
+1. **Define runtime context**: Modify the `Context` class in the `graph.py` file to expose the arguments you want to configure per assistant. For example, in a chatbot application you may want to define a dynamic system prompt or LLM to use. For more information on runtime context in LangGraph, [see here](https://langchain-ai.github.io/langgraph/agents/context/?h=context#static-runtime-context).
 
 2. **Extend the graph**: The core logic of the application is defined in [graph.py](./src/agent/graph.py). You can modify this file to add new nodes, edges, or change the flow of information.
 
@@ -67,14 +59,3 @@ For more advanced features and examples, refer to the [LangGraph documentation](
 
 LangGraph Studio also integrates with [LangSmith](https://smith.langchain.com/) for more in-depth tracing and collaboration with teammates, allowing you to analyze and optimize your chatbot's performance.
 
-<!--
-Configuration auto-generated by `langgraph template lock`. DO NOT EDIT MANUALLY.
-{
-  "config_schemas": {
-    "agent": {
-      "type": "object",
-      "properties": {}
-    }
-  }
-}
--->

langgraph.json

@@ -1,7 +1,8 @@
 {
   "dependencies": ["."],
   "graphs": {
-    "agent": "./src/agent/graph.py:graph"
+    "agent": "./src/agent/graph.py:graph",
+    "my_agent": "./src/agent/my_graph.py:make_graph"
   },
   "env": ".env",
   "image_distro": "wolfi"

requirements.txt

@@ -0,0 +1 @@
+langchain

src/agent/my_graph.py

@@ -0,0 +1,67 @@
+from typing import Annotated, Any
+
+from langchain.chat_models import init_chat_model
+from langchain_core.tools import tool
+from langchain_mcp_adapters.client import MultiServerMCPClient
+from langchain_tavily import TavilySearch
+from langchain_core.messages import BaseMessage
+from langgraph.graph.state import CompiledStateGraph
+from typing_extensions import TypedDict
+
+from langgraph.checkpoint.memory import InMemorySaver
+from langgraph.graph import StateGraph
+from langgraph.graph.message import add_messages
+from langgraph.prebuilt import ToolNode, tools_condition
+from langchain_community.chat_models.tongyi import ChatTongyi
+
+from src.mcp.mcp_tools import get_client
+
+llm = ChatTongyi(
+    model="qwen-max",   # 此处以qwen-max为例，您可按需更换模型名称。模型列表：https://help.aliyun.com/zh/model-studio/getting-started/models
+    streaming=True,
+    # other params...
+)
+
+
+@tool
+def get_wheather(location: str) -> str:
+    """输入城市查询天气"""
+    return f"The weather in {location} is 20 degrees Celsius."
+
+class State(TypedDict):
+    messages: Annotated[list, add_messages]
+
+
+
+async def make_graph() -> CompiledStateGraph[Any, Any, Any, Any]:
+    graph_builder = StateGraph(State)
+
+    # tool = TavilySearch(max_results=2)
+
+    client = get_client()
+
+    mcp_tools = await client.get_tools()
+
+    tools = [get_wheather]
+    tools.extend(mcp_tools)
+    llm_with_tools = llm.bind_tools(tools)
+
+    def chatbot(state: State):
+        return {"messages": [llm_with_tools.invoke(state["messages"])]}
+
+    graph_builder.add_node("chatbot", chatbot)
+
+    tool_node = ToolNode(tools=[tool])
+    graph_builder.add_node("tools", tool_node)
+
+    graph_builder.add_conditional_edges(
+        "chatbot",
+        tools_condition,
+    )
+    graph_builder.add_edge("tools", "chatbot")
+    graph_builder.set_entry_point("chatbot")
+    # memory = InMemorySaver()
+    # graph = graph_builder.compile(checkpointer=memory)
+    graph = graph_builder.compile()
+
+    return graph

src/mcp/mcp_servers.json

@@ -0,0 +1,9 @@
+{
+  "math": {
+    "command": "python",
+    "args": [
+      "/home/kexsh/src/knightutils/test/mcp/math_server.py"
+    ],
+    "transport": "stdio"
+  }
+}

src/mcp/mcp_tools.py

@@ -0,0 +1,9 @@
+from langchain_mcp_adapters.client import MultiServerMCPClient
+import json
+
+def get_client() -> MultiServerMCPClient:
+    mcp_tools = json.load(open("mcp_tools.json"))
+
+    client = MultiServerMCPClient(mcp_tools)
+
+    return client