一文彻底搞懂智能体Agent基于ReAct的工具调用

Answer the following questions as best you can. You have access to the following tools:{tool_strings}
The way you use the tools is by specifying a json blob.Specifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).
The only values that should be in the "action" field are: {tool_names}
The $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:
```{{{{"action": $TOOL_NAME,"action_input": $INPUT}}}}```
ALWAYS use the following format:
Question: the input question you must answerThought: you should always think about what to doAction:```$JSON_BLOB```Observation: the result of the action... (this Thought/Action/Observation can repeat N times)Thought: I now know the final answerFinal Answer: the final answer to the original input question
Begin! Reminder to always use the exact characters `Final Answer` when responding.

#!/bin/bashexport OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"export OPENAI_API_KEY="sk-xxx" # 替换为你的key
curl ${OPENAI_BASE_URL}/chat/completions \-H "Content-Type: application/json" \-H "Authorization: Bearer $OPENAI_API_KEY" \-d '{  "model": "deepseek-v3",  "messages": [    {      "role": "system",      "content": "\nAnswer the following questions as best you can. You have access to the following tools:\n{\"name\": \"get_weather\", \"description\": \"Get weather\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\", \"description\": \"the name of the location\"}}, \"required\": [\"location\"]}}\n\n\nThe way you use the tools is by specifying a json blob.\nSpecifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).\n\nThe only values that should be in the \"action\" field are: get_weather\n\nThe $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:\n\n```\n{{\n\"action\": $TOOL_NAME,\n\"action_input\": $INPUT\n}}\n```\n\nALWAYS use the following format:\n\nQuestion: the input question you must answer\nThought: you should always think about what to do\nAction:\n```\n$JSON_BLOB\n```\nObservation: the result of the action\n... (this Thought/Action/Observation can repeat N times)\nThought: I now know the final answer\nFinal Answer: the final answer to the original input question\n\n\nBegin! Reminder to always use the exact characters `Final Answer` when responding. \n"    },    {      "role": "user",      "content": "Question: 北京和广州天气怎么样\n\n"    }  ],  "stop": "Observation:"}'

Thought: 我需要获取北京和广州的天气信息。首先，我将获取北京的天气。Action:```{  "action": "get_weather",  "action_input": {    "location": "北京"  }}```

{  "choices": [    {      "message": {        "content": "Thought: 我需要获取北京和广州的天气信息。首先，我将获取北京的天气。\n\nAction:\n```\n{\n  \"action\": \"get_weather\",\n  \"action_input\": {\n    \"location\": \"北京\"\n  }\n}\n```",        "role": "assistant"      },      "finish_reason": "stop",      "index": 0,      "logprobs": null    }  ],  "object": "chat.completion",  "usage": {    "prompt_tokens": 305,    "completion_tokens": 49,    "total_tokens": 354  },  "created": 1745651748,  "system_fingerprint": null,  "model": "deepseek-v3",  "id": "chatcmpl-697b0627-4fca-975b-954c-7304386ac224"}

{  "msg": "success",  "code": 0,  "data": {    "location": {      "id": "54511",      "name": "北京",      "path": "中国, 北京, 北京"    },    "now": {      "precipitation": 0.0,      "temperature": 23.4,      "pressure": 1005.0,      "humidity": 43.0,      "windDirection": "西南风",      "windDirectionDegree": 216.0,      "windSpeed": 2.7,      "windScale": "微风",      "feelst": 23.1    },    "alarm": [],    "jieQi": "",    "lastUpdate": "2025/04/26 15:00"  }}

#!/bin/bashexport OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"export OPENAI_API_KEY="sk-xxx" # 替换为你的key
curl ${OPENAI_BASE_URL}/chat/completions \-H "Content-Type: application/json" \-H "Authorization: Bearer $OPENAI_API_KEY" \-d '{  "model": "deepseek-v3",  "messages": [    {      "role": "system",      "content": "\nAnswer the following questions as best you can. You have access to the following tools:\n{\"name\": \"get_weather\", \"description\": \"Get weather\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\", \"description\": \"the name of the location\"}}, \"required\": [\"location\"]}}\n\n\nThe way you use the tools is by specifying a json blob.\nSpecifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).\n\nThe only values that should be in the \"action\" field are: get_weather\n\nThe $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:\n\n```\n{{\n\"action\": $TOOL_NAME,\n\"action_input\": $INPUT\n}}\n```\n\nALWAYS use the following format:\n\nQuestion: the input question you must answer\nThought: you should always think about what to do\nAction:\n```\n$JSON_BLOB\n```\nObservation: the result of the action\n... (this Thought/Action/Observation can repeat N times)\nThought: I now know the final answer\nFinal Answer: the final answer to the original input question\n\n\nBegin! Reminder to always use the exact characters `Final Answer` when responding. \n"    },    {      "role": "user",      "content": "Question: 北京和广州天气怎么样\n\nThought: 我需要获取北京和广州的天气信息。首先，我将获取北京的天气。\n\nAction:\n```\n{\n\"action\": \"get_weather\",\n\"action_input\": {\n\"location\": \"北京\"\n}\n}\n```\nObservation: {\"msg\":\"success\",\"code\":0,\"data\":{\"location\":{\"id\":\"54511\",\"name\":\"北京\",\"path\":\"中国, 北京, 北京\"},\"now\":{\"precipitation\":0.0,\"temperature\":23.4,\"pressure\":1005.0,\"humidity\":43.0,\"windDirection\":\"西南风\",\"windDirectionDegree\":216.0,\"windSpeed\":2.7,\"windScale\":\"微风\",\"feelst\":23.1},\"alarm\":[],\"jieQi\":\"\",\"lastUpdate\":\"2025/04/26 15:00\"}}\n"    }  ],  "stop": "Observation:"}'

Thought: 我已经获取了北京的天气信息。接下来，我将获取广州的天气信息。Action:```{  "action": "get_weather",  "action_input": {    "location": "广州"  }}```

{  "choices": [    {      "message": {        "content": "Thought: 我已经获取了北京的天气信息。接下来，我将获取广州的天气信息。\n\nAction:\n```\n{\n\"action\": \"get_weather\",\n\"action_input\": {\n\"location\": \"广州\"\n}\n}\n```\nObservation",        "role": "assistant"      },      "finish_reason": "stop",      "index": 0,      "logprobs": null    }  ],  "object": "chat.completion",  "usage": {    "prompt_tokens": 472,    "completion_tokens": 46,    "total_tokens": 518  },  "created": 1745651861,  "system_fingerprint": null,  "model": "deepseek-v3",  "id": "chatcmpl-a822b8d7-9105-9dc2-8e98-4327afb50b3a"}

{  "msg": "success",  "code": 0,  "data": {    "location": {      "id": "59287",      "name": "广州",      "path": "中国, 广东, 广州"    },    "now": {      "precipitation": 0.0,      "temperature": 24.2,      "pressure": 1005.0,      "humidity": 79.0,      "windDirection": "东北风",      "windDirectionDegree": 31.0,      "windSpeed": 1.3,      "windScale": "微风",      "feelst": 27.1    },    "alarm": [],    "jieQi": "",    "lastUpdate": "2025/04/26 15:00"  }}

#!/bin/bashexport OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"export OPENAI_API_KEY="sk-xxx" # 替换为你的key
curl ${OPENAI_BASE_URL}/chat/completions \-H "Content-Type: application/json" \-H "Authorization: Bearer $OPENAI_API_KEY" \-d '{  "model": "deepseek-v3",  "messages": [    {      "role": "system",      "content": "\nAnswer the following questions as best you can. You have access to the following tools:\n{\"name\": \"get_weather\", \"description\": \"Get weather\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\", \"description\": \"the name of the location\"}}, \"required\": [\"location\"]}}\n\n\nThe way you use the tools is by specifying a json blob.\nSpecifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).\n\nThe only values that should be in the \"action\" field are: get_weather\n\nThe $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:\n\n```\n{{\n\"action\": $TOOL_NAME,\n\"action_input\": $INPUT\n}}\n```\n\nALWAYS use the following format:\n\nQuestion: the input question you must answer\nThought: you should always think about what to do\nAction:\n```\n$JSON_BLOB\n```\nObservation: the result of the action\n... (this Thought/Action/Observation can repeat N times)\nThought: I now know the final answer\nFinal Answer: the final answer to the original input question\n\n\nBegin! Reminder to always use the exact characters `Final Answer` when responding. \n"    },    {      "role": "user",      "content": "Question: 北京和广州天气怎么样\n\nThought: 我需要获取北京和广州的天气信息。首先，我将获取北京的天气。\n\nAction:\n```\n{\n\"action\": \"get_weather\",\n\"action_input\": {\n\"location\": \"北京\"\n}\n}\n```\nObservation: {\"msg\":\"success\",\"code\":0,\"data\":{\"location\":{\"id\":\"54511\",\"name\":\"北京\",\"path\":\"中国, 北京, 北京\"},\"now\":{\"precipitation\":0.0,\"temperature\":23.4,\"pressure\":1005.0,\"humidity\":43.0,\"windDirection\":\"西南风\",\"windDirectionDegree\":216.0,\"windSpeed\":2.7,\"windScale\":\"微风\",\"feelst\":23.1},\"alarm\":[],\"jieQi\":\"\",\"lastUpdate\":\"2025/04/26 15:00\"}}\nThought: 现在我已经获取了北京的天气信息，接下来我将获取广州的天气信息。\n\nAction:\n```\n{\n\"action\": \"get_weather\",\n\"action_input\": {\n\"location\": \"广州\"\n}\n}\n```\nObservation\nObservation: {\"msg\":\"success\",\"code\":0,\"data\":{\"location\":{\"id\":\"59287\",\"name\":\"广州\",\"path\":\"中国, 广东, 广州\"},\"now\":{\"precipitation\":0.0,\"temperature\":24.2,\"pressure\":1005.0,\"humidity\":79.0,\"windDirection\":\"东北风\",\"windDirectionDegree\":31.0,\"windSpeed\":1.3,\"windScale\":\"微风\",\"feelst\":27.1},\"alarm\":[],\"jieQi\":\"\",\"lastUpdate\":\"2025/04/26 15:00\"}}\n"    }  ],  "stop": "Observation:"}'

Thought: 我已经获取了北京和广州的天气信息，现在可以回答用户的问题了。
Final Answer: 北京的天气温度为23.4°C，湿度为43%，风向为西南风，风速为2.7米/秒。广州 的天气温度为24.2°C，湿度为79%，风向为东北风，风速为1.3米/秒。

{  "choices": [    {      "message": {        "content": "Thought: 我已经获取了北京和广州的天气信息，现在可以回答用户的问题了。\n\nFinal Answer: 北京的天气温度为23.4°C，湿度为43%，风向为西南风，风速为2.7米/秒。广州 的天气温度为24.2°C，湿度为79%，风向为东北风，风速为1.3米/秒。",        "role": "assistant"      },      "finish_reason": "stop",      "index": 0,      "logprobs": null    }  ],  "object": "chat.completion",  "usage": {    "prompt_tokens": 641,    "completion_tokens": 79,    "total_tokens": 720  },  "created": 1745652025,  "system_fingerprint": null,  "model": "deepseek-v3",  "id": "chatcmpl-d9b85f31-589e-9c6f-8694-cf813344e464"}

maxIter = 5 # 最大迭代次数agent_scratchpad = "" # agent思考过程(Thought/Action/Observation)for iterSeq in range(1, maxIter+1):    构造chat completion请求        messages有2条            第1条为系统提示词消息(含工具定义)            第2条为用户消息：Question + agent思考过程(Thought/Action/Observation)        stop参数设置为"Observation:"    获取chat completion结果    如果chat completion结果带有"Final Answer:"        返回最终答案    如果chat completion结果带有Action        解析并调用相应函数        更新agent思考过程：把本次LLM的输出(Though/Action)和工具调用结果(Observation)添加到agent_scratchpad        继续迭代

import jsonimport reimport requestsimport urllib.parsefrom typing import Iterablefrom openai import OpenAIfrom openai.types.chat.chat_completion_message_param import ChatCompletionMessageParamfrom openai.types.chat.chat_completion_user_message_param import (    ChatCompletionUserMessageParam,)from openai.types.chat.chat_completion_system_message_param import (    ChatCompletionSystemMessageParam,)
# 加载环境变量from dotenv import load_dotenvload_dotenv()
client = OpenAI()model = "deepseek-v3"
# 工具定义tools = [    {        "type": "function",        "function": {            "name": "get_weather",            "description": "Get weather",            "parameters": {                "type": "object",                "properties": {                    "location": {"type": "string", "description": "location"}                },                "required": ["location"],            },        },    }]
# 系统提示词def get_system_prompt():    tool_strings = "\n".join([json.dumps(tool["function"]) for tool in tools])    tool_names = ", ".join([tool["function"]["name"] for tool in tools])    systemPromptFormat = """Answer the following questions as best you can. You have access to the following tools:{tool_strings}

The way you use the tools is by specifying a json blob.Specifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).
The only values that should be in the "action" field are: {tool_names}
The $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:
```{{{{"action": $TOOL_NAME,"action_input": $INPUT}}}}```
ALWAYS use the following format:
Question: the input question you must answerThought: you should always think about what to doAction:```$JSON_BLOB```Observation: the result of the action... (this Thought/Action/Observation can repeat N times)Thought: I now know the final answerFinal Answer: the final answer to the original input question

Begin! Reminder to always use the exact characters `Final Answer` when responding. """    return systemPromptFormat.format(tool_strings=tool_strings, tool_names=tool_names)
# 实现获取天气def get_weather(location: str) -> str:    url = "http://weather.cma.cn/api/autocomplete?q=" + urllib.parse.quote(location)    response = requests.get(url)    data = response.json()    if data["code"] != 0:        return "没找到该位置的信息"    location_code = ""    for item in data["data"]:        str_array = item.split("|")        if (            str_array[1] == location            or str_array[1] + "市" == location            or str_array[2] == location        ):            location_code = str_array[0]            break    if location_code == "":        return "没找到该位置的信息"    url = f"http://weather.cma.cn/api/now/{location_code}"    return requests.get(url).text
# 实现工具调用def invoke_tool(toolName: str, toolParamaters) -> str:    result = ""    if toolName == "get_weather":        result = get_weather(toolParamaters["location"])    else:        result = f"函数{toolName}未定义"    return result
def main():    query = "北京和广州天气怎么样"    systemMsg = ChatCompletionSystemMessageParam(        role="system", content=get_system_prompt()    )    maxIter = 5  # 最大迭代次数    agent_scratchpad = ""  # agent思考过程    action_pattern = re.compile(r"\nAction:\n`{3}(?:json)?\n(.*?)`{3}.*?$", re.DOTALL)    for iterSeq in range(1, maxIter + 1):        messages: Iterable[ChatCompletionMessageParam] = list()        messages.append(systemMsg)        messages.append(            ChatCompletionUserMessageParam(                role="user", content=f"Question: {query}\n\n{agent_scratchpad}"            )        )        print(f">> iterSeq:{iterSeq}")        print(f">>> messages: {json.dumps(messages)}")        # 向LLM发起请求，注意需要设置stop参数        chat_completion = client.chat.completions.create(            messages=messages,            model=model,            stop="Observation:",        )        content = chat_completion.choices[0].message.content        print(f">>> content:\n{content}")        final_answer_match = re.search(r"\nFinal Answer:\s*(.*)", content)        if final_answer_match:            final_answer = final_answer_match.group(1)            print(f">>> 最终答案: {final_answer}")            return        action_match = action_pattern.search(content)        if action_match:            obj = json.loads(action_match.group(1))            toolName = obj["action"]            toolParameters = obj["action_input"]            print(f">>> tool name:{toolName}")            print(f">>> tool parameters:{toolParameters}")            result = invoke_tool(toolName, toolParameters)            print(f">>> tool result: {result}")            # 把本次LLM的输出(Though/Action)和工具调用结果(Observation)添加到agent_scratchpad            agent_scratchpad += content + f"\nObservation: {result}\n"        else:            print(">>> ERROR: detect invalid response")            return    print(">>> 迭代次数达到上限，我无法得到最终答案")
main()