pip install -U langchain langgraph langchain-community dashscope arxiv

from langchain_community.chat_models import ChatTongyi
from langchain_core.messages import HumanMessage
import os

# 初始化模型
llm = ChatTongyi(api_key=os.environ.get("DASHSCOPE_API_KEY"))

# 發(fā)送消息
response = llm.invoke([HumanMessage(content="你好，請用一句話介紹一下你自己。")])

print(response.content)

from langchain_community.chat_models import ChatTongyi
import os
llm = ChatTongyi(api_key=os.environ.get("DASHSCOPE_API_KEY"))

from langchain_community.agent_toolkits.load_tools import load_tools
tools = load_tools(["arxiv"])

tools = load_tools(["arxiv","llm-math", "wikipedia"], llm=llm)

tools = load_tools(["python_repl"], llm=llm, allow_dangerous_tools=True)

def calculate(what):
    return str(eval(what))

print(calculate("3 + 7 * 2"))   # 返回 17
print(calculate("10 / 4"))      # 返回 2.5

from langchain.tools import tool

@tool
def calculate(what):
    return str(eval(what))

from langchain.tools import tool

@tool
def calculate(what: str) -> str:
    """
    calculate:
    e.g. calculate: 4 * 7 / 3
    Runs a calculation and returns the number - uses Python so be sure to use floating point syntax if necessary
    """
    return str(eval(what))

from pydantic import BaseModel, Field

class CalcInput(BaseModel):
    """Input for math calculation"""
    what: str = Field(descriptinotallow="A mathematical expression, e.g., '4 * 7 / 3'")
    
from langchain.tools import tool

@tool(args_schema=CalcInput)
def calculate(what: str) -> str:
    """
    calculate:
    e.g. calculate: 4 * 7 / 3
    Runs a calculation and returns the number - uses Python so be sure to use floating point syntax if necessary
    """
    return str(eval(what))

functions = [
    {
        "name": "calculate",
        "description": "Runs a calculation and returns the number - uses Python so be sure to use floating point syntax if necessary",
        "parameters": {
            "type": "object",
            "properties": {
                "what": {
                    "type": "string",
                    "description": "A mathematical expression to evaluate, e.g. '4 * 7 / 3'"
                }
            },
            "required": ["what"]
        }
    }
]

tools = [calculate, average_dog_weight]

tools = load_tools(["arxiv"] + [calculate, average_dog_weight])

from langchain_community.agent_toolkits.load_tools import load_tools
tools = load_tools(["arxiv"])

from langgraph.checkpoint.memory import InMemorySaver 

memory = InMemorySaver()

{"messages": [...],
  "thread_model_call_count": 1,
  "run_model_call_count": 1}

[HumanMessage(cnotallow="hi! I'm bob", additional_kwargs={}, response_metadata={}, id='beb6e2ba-c211-4283-a5d6-78566f567c86'), 

AIMessage(cnotallow="Hello Bob! It's nice to meet you. How can I assist you today?", additional_kwargs={}, response_metadata={'model_name': 'qwen-max', 'finish_reason': 'stop', 'request_id': 'aa493e36-a2d8-40ac-abc6-18d7d31735f7', 'token_usage': {'input_tokens': 227, 'output_tokens': 17, 'prompt_tokens_details': {'cached_tokens': 0}, 'total_tokens': 244}}, id='lc_run--ca2c7e95-ff30-46a6-8019-293b186f0bf7-0')]

Assistant is a large language model trained by OpenAI.

Assistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.

Assistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.

Overall, Assistant is a powerful tool that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.

TOOLS:
------

Assistant has access to the following tools:

{tools}

To use a tool, please use the following format:
"""

Thought: Do I need to use a tool? Yes Action: the action to take, should be one of [{tool_names}] Action Input: the input to the action Observation: the result of the action

"""

When you have a response to say to the Human, or if you do not need to use a tool, you MUST use the format:
"""

Thought: Do I need to use a tool? No Final Answer: [your response here]

"""

Begin!

Previous conversation history:
{chat_history}

New input: {input}
{agent_scratchpad}

Thought: Do I need to use a tool? Yes
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action

action_match = re.search(r"Action: (.*)", text)
input_match = re.search(r"Action Input: (.*)", text)

{
  "role": "assistant",
  "content": "",
  "tool_calls": [
    {
      "name": "get_weather",
      "arguments": {"city": "廣州"},
      "id": "tool_call_1"
    }
  ]
}

{
  "role": "assistant",
  "content": "廣州的天氣是25度，祝你出行愉快",
  "tool_calls": []
}

from langchain.agents import create_agent

agent = create_agent(model=llm, 
                     tools=tools, 
                     system_prompt="You are a helpful assistant", 
                     checkpointer=memory)

{"messages": [{"role": "user", "content": "請使用 arxiv 工具查詢論文編號 1605.08386"}]}

cnotallow={"configurable": {"thread_id": "user_1"}}

result1 = agent.invoke({"messages": [{"role": "user", "content": "請使用 arxiv 工具查詢論文編號 1605.08386"}]}, cnotallow={"configurable": {"thread_id": "user_1"}})
print(result1)

{'messages': [HumanMessage(cnotallow='請使用 arxiv 工具查
詢論文編號 1605.08386', additional_kwargs={}, response_metadata={}, id='804b6b57-a615-4797-996e-70a6822d7eeb'), AIMessage(cnotallow='', additional_kwargs={'tool_calls': [{'function': {'arguments': '{"query": "1605.08386"}', 'name': 'arxiv'}, 'id': 'call_1257191a65e448d1840fed', 'index': 0, 'type': 'function'}]}, response_metadata={'model_name': 'qwen-turbo', 'finish_reason': 'tool_calls', 'request_id': '6274123b-3ef1-4e83-85e8-3e1be634cf6c', 'token_usage': {'input_tokens': 232, 'output_tokens': 28, 'prompt_tokens_details': {'cached_tokens': 0}, 'total_tokens': 260}}, id='lc_run--5eddaf4e-bc5a-4f8b-8802-b6f6ed45b1c1-0', tool_calls=[{'name': 'arxiv', 'args': {'query': '1605.08386'}, 'id': 'call_1257191a65e448d1840fed', 'type': 'tool_call'}]), ToolMessage(cnotallow='Published: 2016-05-26\nTitle: Heat-bath random walks with Markov bases\nAuthors: Caprice Stanley, Tobias Windisch\nSummary: Graphs on lattice points are studied whose edges come from a finite set of\nallowed moves of arbitrary length. We show that the diameter of these graphs on\nfibers of a fixed integer matrix can be bounded from above by a constant. We\nthen study the mixing behaviour of heat-bath random walks on these graphs. We\nalso state explicit conditions on the set of moves so that the heat-bath random\nwalk, a generalization of the Glauber dynamics, is an expander in fixed\ndimension.', name='arxiv', id='fa8a6ea5-0e38-42b7-b846-8e872b302f3f', tool_call_id='call_1257191a65e448d1840fed'), AIMessage(cnotallow='論文編號 1605.08386 的標(biāo)題是《Heat-bath random walks with Markov bases》。該論文于 2016 年 5 月 26 日發(fā)表，作者是 Caprice Stanley 和 Tobias Windisch。\n\n這篇論文研究了由有限的允許移動集生成的格點圖，這些移動可以是任意長度的。論文展示了在固定整數(shù)矩陣的纖維上，這些圖的直徑可以被常數(shù)從上面限制。然后研究了這些圖上的熱浴隨機游走的混合行為。此外，還給出了移動集的顯式條件，以確保熱浴隨機游走（Glauber 動力學(xué)的一種推廣）在固定維度下是一個擴展器。', additional_kwargs={}, response_metadata={'model_name': 'qwen-turbo', 'finish_reason': 'stop', 'request_id': 'f75c75a1-35ca-4bbb-bc71-ea9dd4233986', 'token_usage': {'input_tokens': 416, 'output_tokens': 156, 'prompt_tokens_details': {'cached_tokens': 0}, 'total_tokens': 572}}, id='lc_run--f0bf855c-4315-4559-b598-e4f4b74e5f05-0')]}

HumanMessage(cnotallow='請使用 arxiv 工具查詢論文編號 1605.08386')

AIMessage(
  cnotallow='',
  additional_kwargs={
    'tool_calls': [
      {
        'function': {'name': 'arxiv', 'arguments': '{"query": "1605.08386"}'},
        'id': 'call_1257191a65e448d1840fed',
        'type': 'function'
      }
    ]
  },
  response_metadata={'finish_reason': 'tool_calls', 'model_name': 'qwen-turbo', ...}
)

ToolMessage(
  name='arxiv',
  cnotallow='Published: 2016-05-26\nTitle: Heat-bath random walks with Markov bases\nAuthors: Caprice Stanley, Tobias Windisch\nSummary: ...',
  tool_call_id='call_1257191a65e448d1840fed'
)

AIMessage(cnotallow='論文編號 1605.08386 的標(biāo)題是《Heat-bath random walks with Markov bases》。該論文于 2016 年 5 月 26 日發(fā)表...')

print(result1["messages"][-1].content)

論文編號 1605.08386 的標(biāo)題是《Heat-bath random walks with Markov bases》。該論文于 2016 年 5 月 26 日發(fā)表，作者是 Caprice Stanley 和 Tobias Windisch。\n\n這篇論文研究了由有限的允許移動集生成的格點圖，這些移動可以是任意長度的。論文展示了在固定整數(shù)矩陣的纖維上，這些圖的直徑可以被常數(shù)從上面限制。然后研究了這些圖上的熱浴隨機游走的混合行為。此外，還給出了移動集的顯式條件，以確保熱浴隨機游走（Glauber 動力學(xué)的一種推廣）在固定維度下是一個擴展器。

from langchain_community.chat_models import ChatTongyi
import os
llm = ChatTongyi(api_key=os.environ.get("DASHSCOPE_API_KEY"))

from langchain_community.agent_toolkits.load_tools import load_tools
tools = load_tools(["arxiv"])

from langgraph.checkpoint.memory import InMemorySaver 

memory = InMemorySaver()

from langchain.agents import create_agent

agent = create_agent(model=llm, 
                     tools=tools, 
                     system_prompt="You are a helpful assistant", 
                     checkpointer=memory)

result1 = agent.invoke({"messages": [{"role": "user", "content": "請使用 arxiv 工具查詢論文編號 1605.08386"}]}, cnotallow={"configurable": {"thread_id": "user_1"}})
print(result1["messages"][-1].content)