文档首页/ AI开发平台ModelArts/ ModelArts用户指南（Studio）/ 通过Function Calling扩展大语言模型交互能力/ 通过Function Calling扩展大语言模型对外部环境的理解
更新时间：2024-12-16 GMT+08:00
查看PDF
分享

通过Function Calling扩展大语言模型对外部环境的理解

本示例将展示如何定义一个获取送货日期的函数，并通过LLM来调用外部API来获取外部信息。

操作步骤

  1. 设置Maas的api key和模型服务地址。 
    import requests
from openai import OpenAI
client = OpenAI(
    api_key="您的 APIKEY", # 从MaaS控制台鉴权管理处获取。
    base_url="https://infer-modelarts.cn-east-4.myhuaweicloud.com/v1/infers/xxxxxx/v1"  # MaaS模型服务的基础url，不包含尾部的chat/completions部分。
)
  2. 自定义一个获取送货日期的函数。 
    from datetime import datetime
def get_delivery_date(order_id: int) -> datetime:
    if order_id == 1:
        return datetime.strptime("2024-09-01 18:30", "%Y-%m-%d %H:%M")
    elif order_id == 2:
        return datetime.strptime("2024-10-20 12:00", "%Y-%m-%d %H:%M")
    else:
        return f"cannot find order_id {order_id}"

# Example usage
print(get_delivery_date(1))
  3. 编写函数的描述。 
    tools = [
    {
        "type": "function",
        "function": {
            "name": "get_delivery_date",
            "description": "Get the delivery date for a customer's order. Call this whenever you need to know the delivery date, for example when a customer asks 'Where is my package'",
            "parameters": {
                "type": "object",
                "properties": {
                    "order_id": {
                        "type": "string",
                        "description": "The customer's order ID.",
                    },
                },
                "required": ["order_id"],
                "additionalProperties": False,
            },
        }
    }
]
  4. 与LLM对话。 
    tools = [
    {
        "type": "function",
        "function": {
            "name": "get_delivery_date",
            "description": "Get the delivery date for a customer's order. Call this whenever you need to know the delivery date, for example when a customer asks 'Where is my package'",
            "parameters": {
                "type": "object",
                "properties": {
                    "order_id": {
                        "type": "string",
                        "description": "The customer's order ID.",
                    },
                },
                "required": ["order_id"],
                "additionalProperties": False,
            },
        }
    }
]
 
messages = [
    {"role": "system", "content": "You are a helpful customer support assistant. Use the supplied tools to assist the user."},
    {"role": "user", "content": "Hi, can you tell me the delivery date for my order?"}
]
messages.append({"role": "assistant", "content": "Hi there! I can help with that. Can you please provide your order ID?"})
messages.append({"role": "user", "content": "i think it is 1"})
 
response = client.chat.completions.create(
    model="Qwen2.5-72B-32K",
    messages=messages,
    tools=tools,
)
 
print(response)

相关文档