首页

主题

Qwen-Agent ​

项目地址:https://github.com/QwenLM/Qwen-Agent

Qwen-Agent 是一个基于 Qwen 的指令遵循、工具使用、规划和记忆能力的 LLM 应用程序开发框架。

它还附带了一些示例应用程序，如浏览器助手、代码解释器和自定义助手。现在 Qwen-Agent 作为 Qwen Chat 的后端。

以下示例基于官方修改 ​

python
import pprint
import urllib.parse
import json5
from qwen_agent.agents import Assistant
from qwen_agent.tools.base import BaseTool, register_tool
from qwen_agent.utils.output_beautify import typewriter_print


# Step 1 (Optional): Add a custom tool named `my_image_gen`.
@register_tool('my_image_gen')
class MyImageGen(BaseTool):
    # The `description` tells the agent the functionality of this tool.
    description = 'AI painting (image generation) service, input text description, and return the image URL drawn based on text information.'
    # The `parameters` tell the agent what input parameters the tool has.
    parameters = [{
        'name': 'prompt',
        'type': 'string',
        'description': 'Detailed description of the desired image content, in English',
        'required': True
    }]

    def call(self, params: str, **kwargs) -> str:
        # `params` are the arguments generated by the LLM agent.
        prompt = json5.loads(params)['prompt']
        prompt = urllib.parse.quote(prompt)
        return json5.dumps(
            {'image_url': f'https://image.pollinations.ai/prompt/{prompt}'},
            ensure_ascii=False)


# Step 2: Configure the LLM you are using.
llm_cfg = {
    # Use the model service provided by DashScope:
    'model': 'qwen-max-latest',
    'model_server': 'dashscope',
    # 'api_key': 'YOUR_DASHSCOPE_API_KEY',
    # It will use the `DASHSCOPE_API_KEY' environment variable if 'api_key' is not set here.

    # Use a model service compatible with the OpenAI API, such as vLLM or Ollama:
    # 'model': 'Qwen2.5-7B-Instruct',
    # 'model_server': 'http://localhost:8000/v1',  # base_url, also known as api_base
    # 'api_key': 'EMPTY',

    # (Optional) LLM hyperparameters for generation:
    'generate_cfg': {
        'top_p': 0.8
    }
}
# Step 3: Create an agent. Here we use the `Assistant` agent as an example, which is capable of using tools and reading files.
system_instruction = '''You are a helpful assistant.
After receiving the user's request, you should:
- first draw an image and obtain the image url,
- then run code `request.get(image_url)` to download the image,
- and finally select an image operation from the given document to process the image.
Please show the image using `plt.show()`.'''
tools = ['my_image_gen', 'code_interpreter']  # `code_interpreter` is a built-in tool for executing code.
files = ['doc.pdf']  # Give the bot a PDF file to read.
bot = Assistant(llm=llm_cfg,
                system_message=system_instruction,
                function_list=tools,
                files=files)

# Step 4: Run the agent as a chatbot.
messages = []  # This stores the chat history.
while True:
    # For example, enter the query "draw a dog and rotate it 90 degrees".
    query = input('\nuser query: ')
    # Append the user query to the chat history.
    messages.append({'role': 'user', 'content': query})
    response = []
    response_plain_text = ''
    print('bot response:')
    for response in bot.run(messages=messages):
        # Streaming output.
        response_plain_text = typewriter_print(response, response_plain_text)
    # Append the bot responses to the chat history.
    messages.extend(response)
image-20250403230028293

执行结果

user query:  一条链子拴住了一头猛兽
bot response:
2025-04-03 22:57:02,887 - memory.py - 116 - INFO - {"keywords_zh": ["链子", "拴住", "猛兽"], "keywords_en": ["chain", "tie up", "fierce beast", "animal"], "text": "一条链子拴住了一头猛兽"}
2025-04-03 22:57:03,479 - simple_doc_parser.py - 411 - INFO - Read parsed doc.pdf from cache.
2025-04-03 22:57:03,480 - doc_parser.py - 108 - INFO - Start chunking doc.pdf (doc.pdf)...
2025-04-03 22:57:03,480 - doc_parser.py - 126 - INFO - Finished chunking doc.pdf (doc.pdf). Time spent: 0.00013208389282226562 seconds.
2025-04-03 22:57:03,481 - base_search.py - 56 - INFO - all tokens: 783
2025-04-03 22:57:03,482 - base_search.py - 59 - INFO - use full ref
[TOOL_CALL] my_image_gen
{"prompt": "A fierce beast tethered by a chain"}
[TOOL_RESPONSE] my_image_gen
{image_url: "https://image.pollinations.ai/prompt/A%20fierce%20beast%20tethered%20by%20a%20chain"}
[TOOL_CALL] code_interpreter
{"code": "import requests\n\nimage_url = \"https://image.pollinations.ai/prompt/A%20fierce%20beast%20tethered%20by%20a%20chain\"\nresponse = requests.get(image_url)\n\nif response.status_code == 200:\n    with open('downloaded_image.jpg', 'wb') as file:\n        file.write(res
2025-04-03 22:57:09,145 - utils.py - 174 - INFO - Downloading doc.pdf to workspace/tools/code_interpreter/doc.pdf...
2025-04-03 22:57:09,146 - utils.py - 191 - INFO - Finished downloading doc.pdf to workspace/tools/code_interpreter/doc.pdf. Time spent: 0.000888824462890625 seconds.
2025-04-03 22:57:09,197 - code_interpreter.py - 173 - INFO - INFO: kernel process's PID = 60093
2025-04-03 22:57:25,322 - code_interpreter.py - 105 - INFO - stdout:

```
Exception reporting mode: Minimal

```

[TOOL_RESPONSE] code_interpreter
Finished execution.
[TOOL_CALL] code_interpreter
{"code": "from PIL import Image\nimport matplotlib.pyplot as plt\n\n# Open the image\ndef open_and_show_image(filename):\n    image = Image.open(filename)\n    \n    # Display the image using matplotlib\n    plt.imshow(image)\n    plt.axis('off')  # Hide axis\n    plt.show()\n\n# Example usage\nopen_and_show_image('downloaded_image.
2025-04-03 22:57:40,624 - utils.py - 174 - INFO - Downloading doc.pdf to workspace/tools/code_interpreter/doc.pdf...
2025-04-03 22:57:40,632 - utils.py - 191 - INFO - Finished downloading doc.pdf to workspace/tools/code_interpreter/doc.pdf. Time spent: 0.005845069885253906 seconds.

[TOOL_RESPONSE] code_interpreter
![fig-001](workspace/tools/code_interpreter/10a6b6b8-ab4f-4aeb-8a01-fe8de2fcc99e.png)
已根据您的描述生成了一张图片，展示了一头被链子拴住的猛兽。以下是该图片：

一头被链子拴住的猛兽](https://image.pollinations.ai/prompt/A%20fierce%20beast%20tethered%20by%20a%20chain)

我将展示下载后的图像。
[](workspace/tools/code_interpreter/10a6b6b8-ab4f-4aeb-8a01-fe8de2fcc99e.png)

希望对这张图片进行哪些操作呢？比如翻转或旋转？
A_iercebeasttetheredbyachain

Agent执行的结果

image-20250403230117430

shell
workspace
└── tools
    ├── code_interpreter
    │   ├── 10a6b6b8-ab4f-4aeb-8a01-fe8de2fcc99e.png
    │   ├── 4541f543-91d1-4050-98a3-7d83f62af592.png
    │   ├── dd207beb-e097-452f-886c-79047772ddef.png
    │   ├── doc.pdf
    │   ├── downloaded_image.jpg
    │   ├── kernel_connection_file_fc6f8461-90bc-4b35-a5fa-a910fed0b0b5_60065.json
    │   └── launch_kernel_fc6f8461-90bc-4b35-a5fa-a910fed0b0b5_60065.py
    ├── doc_parser
    │   └── 24cc42a6ca852ccdfaae715a9a9873dd03aa255ae8e213db8aeb720a8df2f947_without_chunking
    └── simple_doc_parser
        └── 24cc42a6ca852ccdfaae715a9a9873dd03aa255ae8e213db8aeb720a8df2f947_ori

WebUI ​

python
from qwen_agent.gui import WebUI
WebUI(bot).run()  # bot is the agent defined in the above code, we do not repeat the definition here for saving space.

image-20250403230230891

浏览器打开http://127.0.0.1:7860/

输出以下提示词

shell
日落余晖，多重树叶透光，胶片摄影，青橙色，情侣，隐约，光朦，高反差，高饱和，高逆光，高曝，多重曝光，氛围感拉满，层次感，梦幻，写意，奇妙意境，杰作，艺术英文信息解析点缀

image-20250403230859610

安装方法 ​

shell
git clone https://github.com/QwenLM/Qwen-Agent.git
cd Qwen-Agent
pip install -U "qwen-agent[gui,rag,code_interpreter,mcp]"