使用 LangChain 调用本地 DeepSeek API - AI创想

from langchain_community.llms import Ollama
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain
# 创建 Ollama 实例，指定模型名称和本地服务地址
llm = Ollama(model="deepseek-r1:32b", base_url="http://localhost:11434")# 定义提示模板
prompt = PromptTemplate(
input_variables=["question"],
template="请回答以下问题：{question}")# 创建 LLM 链
chain = LLMChain(llm=llm, prompt=prompt)# 定义问题
question ="1+3=？"# 调用链进行推理
result = chain.invoke({"question": question})ifisinstance(result,dict):
result = result.get("text","")# 打印结果print(result)

复制代码

from langchain_deepseek import ChatDeepSeek
llm = ChatOpenAI(
model="deepseek-v3",
openai_api_key="XXXX",
openai_api_base="XXXXX/v1/",#注意此处url层级到v1
max_tokens=1024,# 其他参数...)
messages =[("system","您是一个有用的助手，负责将英语翻译成中文。请翻译用户句子。",),("human","I love programming."),]
ai_msg = llm.invoke(messages)
ai_msg.content

复制代码

from langchain_core.prompts import ChatPromptTemplate
prompt = ChatPromptTemplate([("system","您是一个有用的助手，负责将 {input_language} 翻译成 {output_language}。",),("human","{input}"),])
chain = prompt | llm
chain.invoke({"input_language":"英语","output_language":"德语","input":"我喜欢编程。",})

复制代码

import datetime
from langchain.tools import tool
from langchain_openai import ChatOpenAI
# 构建OpenAI客户端
llm = ChatOpenAI(
model="deepseek-v3",
openai_api_key="your api key",
openai_api_base="api url")# 定义⼯具注意要添加注释@tooldefadd(a:int, b:int)->int:"""Adds a and b."""return a + b
print(add.invoke({"a":11,"b":21}))

复制代码

import datetime
from langchain.tools import tool
from langchain_openai import ChatOpenAI
from langchain.tools.render import render_text_description
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import JsonOutputParser
# 构建OpenAI客户端
llm = ChatOpenAI(
model="deepseek-v3",
openai_api_key="your api key",
openai_api_base="api url")# 定义⼯具注意要添加注释@tooldefadd(a:int, b:int)->int:"""Adds a and b."""return a + b
@tooldefget_current_date():"""获取今天⽇期"""return datetime.datetime.today().strftime("%Y-%m-%d")
tools =[add, get_current_date]# 构建工具条件和调用描述
rendered_tools = render_text_description(tools)print("rendered_tools:", rendered_tools)# 构建系统提示词
system_prompt =f"""
You are an assistant that has access to the following set of tools. Here are the names and descriptions for each tool:
{rendered_tools}
Given the user input, return the name and input of the tool to use. Return your response as a JSON blob with 'name' and 'arguments' keys."""# 构建输入提示词模板
prompt = ChatPromptTemplate.from_messages([("system", system_prompt),("user","{input}")])# 构建chain以及最后格式化输出
chain = prompt | llm
print(chain.invoke({"input":"what's 11 + 12"}))# 输出所有结果
tool_map ={tool.name: tool for tool in tools}deftools_call(model_output):
chosen_tool = tool_map[model_output["name"]]return chosen_tool.invoke(model_output["arguments"])# 在链条后面添加JsonOutputParser()和tools_call用来限制输出内容，只输出工具的结果# 链式组合
chain = prompt | llm| JsonOutputParser()| tools_call
print(chain.invoke({"input":"what's 11 + 33"}))print(chain.invoke({"input":"what's the date today?"}))

复制代码

from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI
# 构建OpenAI客户端
llm = ChatOpenAI(
model="deepseek-v3",
openai_api_key="your api key",
openai_api_base="api url")
agent = create_react_agent(
model=llm,
tools=[get_current_date],
prompt="你是一个有用的助手，负责根据用户提出的问题调用对应的工具获得结果后输出",)
result = agent.invoke({"messages":[{"role":"user","content":"今天是几月几日?"}]}print(result["messages"][-1].content)

复制代码