![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/06\/20250612072843-684a81ab6a8e7.png\")
<\/p>\n\u4e0d\u8bba\u662f\u7b80\u5355\u8fd8\u662f\u590d\u6742\u7684\u56fe，\u4e00\u4e2a\u660e\u663e\u7684\u9650\u5236\u662f：\u4e4b\u524d\u5185\u5bb9\u4e2d\u7684\u6240\u6709\u5b9e\u73b0\u90fd\u53ea\u80fd\u6267\u884c\u5355\u4e00\u4efb\u52a1\u3002\u4e5f\u5c31\u662f\u8bf4，\u4e00\u65e6\u56fe\u88ab\u7f16\u8bd1\u5e76\u6839\u636e\u7528\u6237\u7684\u8f93\u5165\u8fd0\u884c\u4ee5\u540e，\u5b83\u867d\u7136\u53ef\u4ee5\u6309\u7167\u65e2\u5b9a\u7684\u56fe\u6d41\u7a0b\u8f93\u51fa\u7ed3\u679c，\u4f46\u5728\u4e0b\u4e00\u6b21\u4ea4\u4e92\u65f6，\u8fd9\u4e2a\u56fe\u5c06\u65e0\u6cd5\u8bb0\u4f4f\u4e4b\u524d\u7684\u5bf9\u8bdd\u5185\u5bb9\u3002\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u4ee3\u7801\u8fdb\u884c\u6d4b\u8bd5，\u89c2\u5bdf\u8fd9\u4e00\u73b0\u8c61：<\/p>\n
mport getpass
\nimport os
\nfrom langchain_openai import ChatOpenAI
\nfrom typing import Annotated
\nfrom typing_extensions import TypedDict
\nfrom IPython.display import Image, display
\nfrom langgraph.graph import StateGraph, MessagesState, START, END
\nfrom langchain_core.messages import AnyMessage, SystemMessage, HumanMessage, ToolMessage
\nfrom langgraph.graph.message import add_messages<\/p>\n
from dotenv import load_dotenv
\nload_dotenv()<\/p>\n
# \u5b9a\u4e49\u5927\u6a21\u578b\u5b9e\u4f8b
\nllm = ChatOpenAI(model="deepseek-chat")<\/p>\n
# \u5b9a\u4e49\u72b6\u6001\u6a21\u5f0f
\nclass State(TypedDict):
\n messages: Annotated[list, add_messages]<\/p>\n
# \u5b9a\u4e49\u5927\u6a21\u578b\u4ea4\u4e92\u8282\u70b9
\ndef call_model(state: State):
\n response = llm.invoke(state["messages"])
\n return {"messages": response}<\/p>\n
# \u5b9a\u4e49\u7ffb\u8bd1\u8282\u70b9
\ndef translate_message(state: State):
\n system_prompt = """
\n Please translate the received text in any language into English as output
\n """
\n messages = state['messages'][-1]
\n messages = [SystemMessage(content=system_prompt)] + [HumanMessage(content=messages.content)]
\n response = llm.invoke(messages)
\n return {"messages": response}<\/p>\n
# \u6784\u5efa\u72b6\u6001\u56fe
\nbuilder = StateGraph(State)<\/p>\n
# \u5411\u56fe\u4e2d\u6dfb\u52a0\u8282\u70b9
\nbuilder.add_node("call_model", call_model)
\nbuilder.add_node("translate_message", translate_message)<\/p>\n
# \u6784\u5efa\u8fb9
\nbuilder.add_edge(START, "call_model")
\nbuilder.add_edge("call_model", "translate_message")
\nbuilder.add_edge("translate_message", END)<\/p>\n
async for chunk in simple_short_graph.astream(input={"messages": ["\u4f60\u597d，\u6211\u53ebbug"]}, stream_mode="values"):
\n message = chunk["messages"][-1].pretty_print()
\n# \u8fd0\u884c\u7ed3\u679c
\n================================ Human Message =================================<\/p>\n
\u4f60\u597d，\u6211\u53ebbug
\n================================== Ai Message ==================================<\/p>\n
\u4f60\u597d，Bug！\u8fd9\u4e2a\u540d\u5b57\u5f88\u6709\u8da3，\u662f\u6709\u4ec0\u4e48\u7279\u522b\u7684\u542b\u4e49\u5417？😄 <\/p>\n
\u5982\u679c\u662f\u7a0b\u5e8f\u5458\u670b\u53cb\u7684\u8bdd，\u53eb\u201cBug\u201d\u53ef\u80fd\u81ea\u5e26\u201c\u53cdbug\u5149\u73af\u201d\u54e6！\u9700\u8981\u5e2e\u5fd9debug\u8fd8\u662f\u5355\u7eaf\u60f3\u804a\u804a？\u968f\u65f6\u6b22\u8fce～ <\/p>\n
（P.S. \u5982\u679c\u662f\u771f\u540d\u7684\u8bdd\u4e5f\u5f88\u9177，\u50cf\u300a\u866b\u866b\u7279\u5de5\u961f\u300b\u91cc\u7684\u89d2\u8272\u2728）
\n================================== Ai Message ==================================<\/p>\n
Hello, Bug! That's a really interesting name \u2013 does it have any special meaning? 😄 <\/p>\n
If you're a programmer friend, being called "Bug" might come with built-in "anti-bug powers"! Do you need help debugging something, or just want to chat? You're always welcome here~ <\/p>\n
(P.S. If it's your real name, that's super cool too \u2013 like a character from *A Bug's Life* \u2728) <\/p>\n
*(Note: Kept the playful tone, translated cultural references naturally ("\u866b\u866b\u7279\u5de5\u961f"\u2192A Bug's Life), and maintained the emoji\/formatting style. Added "built-in" for smoother tech humor flow.)*<\/p>\n
\u7b2c\u4e8c\u6b21\u8be2\u95ee：<\/p>\n
async for chunk in simple_short_graph.astream(input={"messages": ["\u8bf7\u95ee，\u6211\u53eb\u4ec0\u4e48？"]}, stream_mode="values"):
\n message = chunk["messages"][-1].pretty_print()<\/p>\n
# \u8fd0\u884c\u7ed3\u679c
\n================================ Human Message =================================<\/p>\n
\u8bf7\u95ee，\u6211\u53eb\u4ec0\u4e48？
\n================================== Ai Message ==================================<\/p>\n
\u4f60\u8fd8\u6ca1\u6709\u544a\u8bc9\u6211\u4f60\u7684\u540d\u5b57\u5462！😊 \u4f60\u53ef\u4ee5\u544a\u8bc9\u6211\u4f60\u7684\u540d\u5b57，\u6216\u8005\u6211\u53ef\u4ee5\u79f0\u547c\u4f60\u4e3a\u201c\u670b\u53cb\u201d\u6216\u5176\u4ed6\u4f60\u559c\u6b22\u7684\u6635\u79f0～ \u4f60\u5e0c\u671b\u6211\u600e\u4e48\u79f0\u547c\u4f60\u5462？
\n================================== Ai Message ==================================<\/p>\n
You haven't told me your name yet! 😊 You can tell me your name, or I can call you "friend" or any other nickname you like~ How would you prefer me to address you? <\/p>\n
\u901a\u8fc7\u4e24\u6b21\u8be2\u95ee，\u6211\u4eec\u53ef\u4ee5\u786e\u5b9a\u5f53\u524d\u7684`graph`\u5b9e\u4f8b\u4e0d\u5177\u5907\u4efb\u4f55\u7684\u4e0a\u4e0b\u6587\u8bb0\u5fc6\u80fd\u529b\u3002<\/p>\n
`LangGraph`\u6846\u67b6\u4e2d\u7684`checkpointer`\u505a\u7684\u5c31\u662f\u8fd9\u6837\u7684\u4e8b\u3002\u5177\u4f53\u6765\u8bf4，\u5b83\u5c31\u662f\u901a\u8fc7\u4e00\u4e9b\u6570\u636e\u7ed3\u6784\u6765\u5b58\u50a8`State`\u72b6\u6001\u4e2d\u4ea7\u751f\u7684\u4fe1\u606f，\u5e76\u4e14\u5728\u6bcf\u4e2a`task`\u5f00\u59cb\u65f6\u53bb\u8bfb\u53d6\u5168\u5c40\u7684\u72b6\u6001\u3002\u4e3b\u8981\u901a\u8fc7\u4ee5\u4e0b\u56db\u79cd\u65b9\u5f0f\u6765\u5b9e\u73b0：<\/p>\n
– MemorySaver：\u7528\u4e8e\u5b9e\u9a8c\u6027\u8d28\u7684\u8bb0\u5fc6\u68c0\u67e5\u70b9\u3002 – SqliteSaver \/ AsyncSqliteSaver：\u4f7f\u7528 `SQLite` \u6570\u636e\u5e93 \u5b9e\u73b0\u7684\u8bb0\u5fc6\u68c0\u67e5\u70b9，\u9002\u5408\u5b9e\u9a8c\u6027\u8d28\u548c\u672c\u5730\u5de5\u4f5c\u6d41\u7a0b\u3002 – PostgresSaver \/ AsyncPostgresSaver：\u4f7f\u7528 `Postgres` \u6570\u636e\u5e93\u5b9e\u73b0\u7684\u9ad8\u7ea7\u68c0\u67e5\u70b9，\u9002\u5408\u5728\u751f\u4ea7\u7cfb\u7edf\u4e2d\u4f7f\u7528\u3002 – \u652f\u6301\u81ea\u5b9a\u4e49\u68c0\u67e5\u70b9\u3002<\/p>\n
`LangGraph` \u6846\u67b6\u6709\u4e00\u4e2a\u5185\u7f6e\u7684\u6301\u4e45\u5c42，\u901a\u8fc7`checkpointer`\u5b9e\u73b0\u3002\u5f53\u4f7f\u7528`checkpointer`\u7f16\u8bd1\u56fe\u65f6，\u68c0\u67e5\u70b9\u4f1a\u5728\u6bcf\u4e2a\u8d85\u7ea7\u6b65\u9aa4\u4e2d\u4fdd\u5b58\u56fe\u72b6\u6001\u7684`checkpoint`\u3002\u8fd9\u4e9b`checkpoint`\u88ab\u4fdd\u5b58\u5230\u4e00\u4e2a`thread`\u4e2d，\u53ef\u4ee5\u5728\u56fe\u6267\u884c\u540e\u8bbf\u95ee\u3002<\/p>\n
\u4f7f\u7528\u65b9\u5f0f\u5f88\u7b80\u5355：<\/p>\n
from langgraph.checkpoint.memory import MemorySaver
\nmemory = MemorySaver()
\ngraph_with_memory = builder.compile(checkpointer=memory) # \u5728\u7f16\u8bd1\u56fe\u7684\u65f6\u5019\u6dfb\u52a0\u68c0\u67e5\u70b9 <\/p>\n
\u793a\u4f8b\u4e2d\u4f7f\u7528：<\/p>\n
# \u5bfc\u5165\u68c0\u67e5\u70b9
\nfrom langgraph.checkpoint.memory import MemorySaver<\/p>\n
# if not os.environ.get("OPENAI_API_KEY"):
\n# os.environ["OPENAI_API_KEY"] = getpass.getpass("Enter your OpenAI API key: ")
\nfrom dotenv import load_dotenv
\nload_dotenv()
\nllm = ChatOpenAI(model="deepseek-chat")<\/p>\n
class State(TypedDict):
\n messages: Annotated[list, add_messages]<\/p>\n
def call_model(state: State):
\n response = llm.invoke(state["messages"])
\n return {"messages": response}<\/p>\n
def translate_message(state: State):
\n system_prompt = """
\n Please translate the received text in any language into English as output
\n """
\n messages = state['messages'][-1]
\n messages = [SystemMessage(content=system_prompt)] + [HumanMessage(content=messages.content)]
\n response = llm.invoke(messages)
\n return {"messages": response}<\/p>\n
builder = StateGraph(State)<\/p>\n
builder.add_node("call_model", call_model)
\nbuilder.add_node("translate_message", translate_message)<\/p>\n
builder.add_edge(START, "call_model")
\nbuilder.add_edge("call_model", "translate_message")
\nbuilder.add_edge("translate_message", END)<\/p>\n
memory = MemorySaver()
\n# \u5728\u7f16\u8bd1\u56fe\u7684\u65f6\u5019\u6dfb\u52a0\u68c0\u67e5\u70b9
\ngraph_with_memory = builder.compile(checkpointer=memory) <\/p>\n
# \u8fd9\u4e2a thread_id \u53ef\u4ee5\u53d6\u4efb\u610f\u6570\u503c
\nconfig = {"configurable": {"thread_id": "1"}}<\/p>\n
for chunk in graph_with_memory.stream({"messages": ["\u4f60\u597d，\u6211\u53ebbug"]}, config, stream_mode="values"):
\n chunk["messages"][-1].pretty_print()<\/p>\n
for chunk in graph_with_memory.stream({"messages": ["\u8bf7\u95ee\u6211\u53eb\u4ec0\u4e48？"]}, config, stream_mode="values"):
\n chunk["messages"][-1].pretty_print()<\/p>\n
\u00a0\u7ed3\u679c\u5982\u4e0b：<\/p>\n
<\/p>\n
\u00a0\u5982\u679c\u5c06thread_id\u66f4\u6539\u4e4b\u540e：<\/p>\n
config = {"configurable": {"thread_id": "3"}}<\/p>\n
for chunk in graph_with_memory.stream({"messages": ["\u8bf7\u95ee\u6211\u53eb\u4ec0\u4e48？"]}, config, stream_mode="debug"):
\n if chunk["type"] == "checkpoint":
\n # print(chunk)
\n print(f"Thread id:{chunk['payload']['config']['configurable']['thread_id']}")
\n print(f"CheckPoint id :{chunk['payload']['config']['configurable']['checkpoint_id']}")<\/p>\n
for message in chunk['payload']['values']['messages']:
\n print(f"Message id:{message.id}，Message content:{message.content}") <\/p>\n
\u8fd0\u884c\u7ed3\u679c\u5982\u4e0b：<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/06\/20250612072844-684a81ac362b2.png\")
<\/p>\n
\u5df2\u7ecf\u5b8c\u5168\u4e22\u5931\u4e86\u4e0a\u4e0b\u6587，\u65e0\u6cd5\u77e5\u9053\u4ee5\u524d\u7684\u6570\u636e\u3002\u00a0<\/p>\n
\u90a3\u4e48\u63a5\u4e0b\u6765\u8981\u8003\u8651\u7684\u662f： \u65e2\u7136\u6240\u5b9e\u9645\u8fdb\u884c\u5b58\u50a8\u7684\u662f `Checkpointer`， \u90a3\u4e48`Checkpointer`\u5982\u4f55\u53bb\u505a\u6301\u4e45\u5316\u7684\u5b58\u50a8\u5462？\u6b63\u5982\u6211\u4eec\u4e0a\u9762\u4f7f\u7528\u7684 `MemorySaver`， \u867d\u7136\u5728\u5f53\u524d\u7684\u4ee3\u7801\u8fd0\u884c\u73af\u5883\u4e0b\u53ef\u4ee5\u53bb\u6307\u5b9a\u7ebf\u7a0bID，\u83b7\u53d6\u5230\u5177\u4f53\u7684\u5386\u53f2\u4fe1\u606f，\u4f46\u662f，\u4e00\u65e6\u6211\u4eec\u91cd\u542f\u4ee3\u7801\u73af\u5883，\u5219\u6240\u6709\u7684\u6570\u636e\u90fd\u5c06\u88ab\u62b9\u9664\u3002\u90a3\u4e48\u4e00\u79cd\u6301\u4e45\u5316\u7684\u65b9\u6cd5\u5c31\u662f\u628a\u6bcf\u4e2a`checkpointer`\u5b58\u50a8\u5230\u672c\u5730\u7684\u6570\u636e\u5e93\u4e2d\u3002<\/p>\n
SqliteSaver`\u662f`checkponiter`\u7684\u7b2c\u4e8c\u79cd\u5b9e\u73b0\u5f62\u5f0f，\u4e0d\u540c\u4e8e`MemorySaver`\u4ec5\u901a\u8fc7\u5b57\u5178\u7684\u5f62\u5f0f\u5c06\u72b6\u6001\u4fe1\u606f\u5b58\u50a8\u5728\u5f53\u524d\u7684\u8fd0\u884c\u73af\u5883\u4e0b，`SqliteSaver`\u505a\u7684\u662f\u6301\u4e45\u5316\u5b58\u50a8，\u8fd9\u4e2a\u65b9\u6cd5\u4f1a\u628a`checkponiter`\u5b9e\u9645\u7684\u5b58\u50a8\u5728\u672c\u5730\u7684`SQLite` \u6570\u636e\u5e93\u4e2d，\u540c\u65f6\u63d0\u4f9b\u4e86\u5f02\u6b65\u73af\u5883\u4e0b\u7684\u5b9e\u73b0`AsyncSqliteSaver`，\u9002\u7528\u4e8e\u8f7b\u91cf\u7ea7\u7684\u5e94\u7528\u843d\u5730\u573a\u666f\u3002<\/p>\n
\u8fd9\u91cc\u4e3a\u4e86\u6f14\u793a`SqliteSaver`\u7684\u6267\u884c\u539f\u7406，\u6211\u4eec\u624b\u52a8\u6784\u5efa\u4e00\u4e2a\u6d4b\u8bd5\u7684`checkpointer`，\u5176\u9ed8\u8ba4\u5b9e\u73b0\u7684\u662f\u4ece`State`\u4e2d\u8fdb\u884c\u63d0\u53d6\u3002<\/p>\n
checkpoint_data = {
\n "thread_id": "bug123",
\n "thread_ts": "2025-06-11T07:23:38.656547+00:00",
\n "checkpoint": {
\n "id": "1ef968fe-1eb4-6049-bfff",
\n },
\n "metadata": {"timestamp": "2025-06-11T07:23:38.656547+00:00"}
\n} <\/p>\n
\u5728\u4e0b\u5217\u7a0b\u5e8f\u4f7f\u7528\u6211\u4eec\u7684\u6d4b\u8bd5\u6570\u636e，\u5e76\u6301\u4e45\u5316\u5230sqllite\u4e2d\u3002<\/p>\n
import sqlite3
\nfrom langgraph.checkpoint.sqlite import SqliteSaver <\/p>\n
with SqliteSaver.from_conn_string("checkpoints20250611.sqlite") as memory:
\n # \u4fdd\u5b58\u68c0\u67e5\u70b9，\u5305\u62ec\u65f6\u95f4\u6233
\n saved_config = memory.put(
\n config={"configurable": {"thread_id": checkpoint_data["thread_id"], "thread_ts": checkpoint_data["thread_ts"], "checkpoint_ns": ""}},
\n checkpoint=checkpoint_data["checkpoint"],
\n metadata=checkpoint_data["metadata"],
\n new_versions= {"writes": {"key": "value"}}
\n )<\/p>\n
# \u68c0\u7d22\u68c0\u67e5\u70b9\u7684\u6570\u636e
\n config = {"configurable": {"thread_id": checkpoint_data["thread_id"]}}<\/p>\n
# \u83b7\u53d6\u7ed9\u5b9a thread_id \u7684\u6240\u6709\u68c0\u67e5\u70b9
\n checkpoints = list(memory.list(config))
\n for checkpoint in checkpoints:
\n print(checkpoint)
\n#\u8fd0\u884c\u7ed3\u679c
\nCheckpointTuple(config={'configurable': {'thread_id': 'bug123', 'checkpoint_ns': '', 'checkpoint_id': '1ef968fe-1eb4-6049-bfff'}}, checkpoint={'id': '1ef968fe-1eb4-6049-bfff'}, metadata={'timestamp': '2025-06-11T07:23:38.656547+00:00', 'thread_id': 'bug123', 'thread_ts': '2025-06-11T07:23:38.656547+00:00'}, parent_config=None, pending_writes=[])<\/p>\n
\u4ecesqllit\u4e2d\u8fdb\u884c\u67e5\u8be2\u6570\u636e：<\/p>\n
# \u5efa\u7acb\u6570\u636e\u5e93\u8fde\u63a5
\nconn = sqlite3.connect("checkpoints20250611.sqlite")<\/p>\n
# \u521b\u5efa\u4e00\u4e2a\u6e38\u6807\u5bf9\u8c61\u6765\u6267\u884c\u4f60\u7684SQL\u67e5\u8be2
\ncursor = conn.cursor()<\/p>\n
# \u67e5\u8be2\u6570\u636e\u5e93\u4e2d\u6240\u6709\u8868\u7684\u540d\u79f0
\ncursor.execute("SELECT name FROM sqlite_master WHERE type='table';")<\/p>\n
# \u83b7\u53d6\u67e5\u8be2\u7ed3\u679c
\ntables = cursor.fetchall()
\n# \u6253\u5370\u6240\u6709\u8868\u540d
\nfor table in tables:
\n print(table)
\n# \u4ece\u68c0\u67e5\u70b9\u8868\u4e2d\u68c0\u7d22\u6240\u6709\u6570\u636e
\ncursor.execute(f"SELECT * FROM checkpoints;")
\nall_data = cursor.fetchall()<\/p>\n
# \u6253\u5370\u68c0\u67e5\u70b9\u8868\u4e2d\u7684\u6240\u6709\u6570\u636e
\nprint("Data in the 'checkpoints' table:")
\nfor row in all_data:
\n print(row)
\n# \u7ed3\u679c\u5982\u4e0b
\n('bug123', '', '1ef968fe-1eb4-6049-bfff', None, 'msgpack', b'\\\\x81\\\\xa2id\\\\xb71ef968fe-1eb4-6049-bfff', b'{"timestamp": "2025-06-11T07:23:38.656547+00:00", "thread_id": "bug123", "thread_ts": "2025-06-11T07:23:38.656547+00:00"}')<\/p>\n
\u5230\u6b64，\u6211\u4eec\u4e0d\u6a21\u62df\u6570\u636e，\u5c06\u771f\u5b9e\u8fd0\u884c\u4e2d\u7684\u6570\u636e\u4fdd\u5b58\u5728sqlLit\u6570\u636e\u5e93\u4e2d\u3002<\/p>\n
from langgraph.checkpoint.sqlite import SqliteSaver
\nfrom langgraph.prebuilt import create_react_agent<\/p>\n
with SqliteSaver.from_conn_string("checkpoints202506111958.sqlite") as checkpointer:
\n graph = create_react_agent(llm, "", checkpointer=checkpointer)
\n display(Image(graph.get_graph().draw_mermaid_png()))<\/p>\n
config = {"configurable": {"thread_id": "1"}}<\/p>\n
for chunk in graph.stream({"messages": ["\u4f60\u597d，\u6211\u53ebbug"]}, config, stream_mode="values"):
\n chunk["messages"][-1].pretty_print()<\/p>\n
\u6211\u4eec\u5728\u6b21\u4ecesqlite\u6570\u636e\u5e93\u54a8\u8be2，\u662f\u5426\u80fd\u8bfb\u53d6\u6570\u636e<\/p>\n
import asyncio
\nfrom contextlib import AsyncExitStack
\nfrom langgraph.checkpoint.sqlite.aio import AsyncSqliteSaver
\nfrom langgraph.prebuilt import create_react_agent<\/p>\n
stack = AsyncExitStack()
\nmemory = await stack.enter_async_context(AsyncSqliteSaver.from_conn_string("checkpoints202506111958.sqlite"))<\/p>\n
graph = create_react_agent(llm, "", checkpointer=memory)<\/p>\n
config = {"configurable": {"thread_id": "1"}}<\/p>\n
async for chunk in graph.astream({"messages": ["\u8bf7\u95ee\u6211\u521a\u624d\u95ee\u4e86\u4ec0\u4e48\u95ee\u9898？"]}, config, stream_mode="values"):
\n chunk["messages"][-1].pretty_print() <\/p>\n
\u7ed3\u679c\u5982\u4e0b：<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/06\/20250612072844-684a81ac68565.png\")
<\/p>\n
\u8bf4\u660e\u5df2\u7ecf\u4ece\u6587\u4ef6\u4e2d\u8bfb\u53d6\u5230\u5b58\u50a8\u7684\u4e0a\u4e0b\u6587\u4fe1\u606f\u3002\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"
\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb695\u6b21\uff0c\u70b9\u8d5e5\u6b21\uff0c\u6536\u85cf6\u6b21\u3002\u4e0d\u8bba\u662f\u7b80\u5355\u8fd8\u662f\u590d\u6742\u7684\u56fe\uff0c\u4e00\u4e2a\u660e\u663e\u7684\u9650\u5236\u662f\uff1a\u4e4b\u524d\u5185\u5bb9\u4e2d\u7684\u6240\u6709\u5b9e\u73b0\u90fd\u53ea\u80fd\u6267\u884c\u5355\u4e00\u4efb\u52a1\u3002\u4e5f\u5c31\u662f\u8bf4\uff0c\u4e00\u65e6\u56fe\u88ab\u7f16\u8bd1\u5e76\u6839\u636e\u7528\u6237\u7684\u8f93\u5165\u8fd0\u884c\u4ee5\u540e\uff0c\u5b83\u867d\u7136\u53ef\u4ee5\u6309\u7167\u65e2\u5b9a\u7684\u56fe\u6d41\u7a0b\u8f93\u51fa\u7ed3\u679c\uff0c\u4f46\u5728\u4e0b\u4e00\u6b21\u4ea4\u4e92\u65f6\uff0c\u8fd9\u4e2a\u56fe\u5c06\u65e0\u6cd5\u8bb0\u4f4f\u4e4b\u524d\u7684\u5bf9\u8bdd\u5185\u5bb9\u3002\u901a\u8fc7\u4e24\u6b21\u8be2\u95ee\uff0c\u6211\u4eec\u53ef\u4ee5\u786e\u5b9a\u5f53\u524d\u7684`graph`\u5b9e\u4f8b\u4e0d\u5177\u5907\u4efb\u4f55\u7684\u4e0a\u4e0b\u6587\u8bb0\u5fc6\u80fd\u529b\u3002<\/p>\n","protected":false},"author":2,"featured_media":43475,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[66,3503,81],"topic":[],"class_list":["post-43478","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-ai","tag-langgraph","tag-python"],"yoast_head":"\n