使用 Swarm 构建多智能体新闻助理，利用Swarm 框架和 Llama 3.2 来自动化新闻处理工作流

本文链接：https://blog.youkuaiyun.com/Code1994/article/details/143502080

本文将带你构建一个多智能体新闻助理，利用 OpenAI 的 Swarm 框架和 Llama 3.2 来自动化新闻处理工作流。在本地运行环境下，我们将实现一个多智能体系统，让不同的智能体各司其职，分步完成新闻搜索、信息综合与摘要生成等任务，而无需付费使用外部服务。

一、应用设计

二、工作流设计

整个流程将由三个智能体分工协作完成：

从 Searcher进行搜索到 Synthesizer 进行合成，最后由 Summarizer 进行总结。

三个智能体的职责描述如下：

Searcher：通过 DuckDuckGo 搜索与给定主题最相关且来源可靠的最新新闻，并以结构化格式返回结果。
Synthesizer：分析提供的原始新闻报道，确定关键主题和重要信息，综合多种来源的信息，编写一份全面而简明的综述，注重事实，保持新闻的客观性，并以清晰、专业的写作风格呈现。
Summarizer：以简洁明了的方式总结新闻，突出最重要的发展，包含关键利益相关者及其行动，添加相关数据，解释事件的重要性和直接影响，使用有力的动词和具体的描述，保持客观性，并在250-400字的段落中提供信息和吸引读者。

三、准备工作

下载 Llama3.2

通过 Ollama 下载 Llama3.2：

运行 Llama3.2 看看是否正常：

安装相关依赖项

创建如下 requirements.txt 文件：

git+https://github.com/openai/swarm.git  
streamlit   
duckduckgo-search

在命令行执行命令 pip install -r requirements.txt，安装相关依赖项。

四、代码实现

智能体实现

Searcher

search_agent = Agent(  
    name="News Searcher",  
    instructions="""  
    You are a news search specialist. Your task is to:  
    1. Search for the most relevant and recent news on the given topic  
    2. Ensure the results are from reputable sources  
    3. Return the raw search results in a structured format  
    """,  
    functions=[search_news],  
    model=MODEL  
)

创建新闻搜索智能体：

专门用于新闻搜索
专注于信誉良好的来源
返回格式化的结果

Synthesizer

synthesis_agent = Agent(  
    name="News Synthesizer",  
    instructions="""  
    You are a news synthesis expert. Your task is to:  
    1. Analyze the raw news articles provided  
    2. Identify the key themes and important information  
    3. Combine information from multiple sources  
    4. Create a comprehensive but concise synthesis  
    5. Focus on facts and maintain journalistic objectivity  
    6. Write in a clear, professional style  
    Provide a 2-3 paragraph synthesis of the main points.  
    """,  
    model=MODEL  
)

创建新闻合成智能体：

分析多种来源
确定关键主题
创建连贯的叙述

Summarizer

summary_agent = Agent(  
    name="News Summarizer",  
    instructions="""  
    You are an expert news summarizer combining AP and Reuters style clarity with digital-age brevity.  
  
    Your task:  
    1. Core Information:  
       - Lead with the most newsworthy development  
       - Include key stakeholders and their actions  
       - Add critical numbers/data if relevant  
       - Explain why this matters now  
       - Mention immediate implications  
  
    2. Style Guidelines:  
       - Use strong, active verbs  
       - Be specific, not general  
       - Maintain journalistic objectivity  
       - Make every word count  
       - Explain technical terms if necessary  
  
    Format: Create a single paragraph of 250-400 words that informs and engages.  
    Pattern: [Major News] + [Key Details/Data] + [Why It Matters/What's Next]  
  
    Focus on answering: What happened? Why is it significant? What's the impact?  
  
    IMPORTANT: Provide ONLY the summary paragraph. Do not include any introductory phrases,   
    labels, or meta-text like "Here's a summary" or "In AP/Reuters style."  
    Start directly with the news content.  
    """,  
    model=MODEL  
)

创建新闻摘要代理：

美联社/路透社风格写作
专业格式
简明摘要

全套的AI大模型学习资源已经整理打包，有需要的小伙伴可以微信扫描下方优快云官方认证二维码，免费领取【保证100%免费】

完整实现

import streamlit as st  
from duckduckgo_search import DDGS  
from swarm import Swarm, Agent  
from datetime import datetime  
from dotenv import load_dotenv  
  
# 加载环境变量  
load_dotenv()  
# 定义模型  
MODEL = "llama3.2:latest"  
# 初始化 Swarm 客户端  
client = Swarm()  
  
# 通过 Streamlit 创建用户界面，为页面和应用程序添加一个标题  
st.set_page_config(page_title="AI News Processor", page_icon="📰")  
st.title("📰 News Inshorts Agent")  
  
# 定义新闻搜索 Function，使用 DuckDuckGo 搜索 API，获取当月新闻，并返回结构化结果  
def search_news(topic):  
    """Search for news articles using DuckDuckGo"""  
    with DDGS() as ddg:  
        results = ddg.text(f"{topic} news {datetime.now().strftime('%Y-%m')}", max_results=3)  
        if results:  
            news_results = "\n\n".join([  
                f"Title: {result['title']}\nURL: {result['href']}\nSummary: {result['body']}"   
                for result in results  
            ])  
            return news_results  
        return f"No news found for {topic}."  
  
# 创建智能体  
search_agent = Agent(  
    name="News Searcher",  
    instructions="""  
    You are a news search specialist. Your task is to:  
    1. Search for the most relevant and recent news on the given topic  
    2. Ensure the results are from reputable sources  
    3. Return the raw search results in a structured format  
    """,  
    functions=[search_news],  
    model=MODEL  
)  
  
synthesis_agent = Agent(  
    name="News Synthesizer",  
    instructions="""  
    You are a news synthesis expert. Your task is to:  
    1. Analyze the raw news articles provided  
    2. Identify the key themes and important information  
    3. Combine information from multiple sources  
    4. Create a comprehensive but concise synthesis  
    5. Focus on facts and maintain journalistic objectivity  
    6. Write in a clear, professional style  
    Provide a 2-3 paragraph synthesis of the main points.  
    """,  
    model=MODEL  
)  
  
summary_agent = Agent(  
    name="News Summarizer",  
    instructions="""  
    You are an expert news summarizer combining AP and Reuters style clarity with digital-age brevity.  
  
    Your task:  
    1. Core Information:  
       - Lead with the most newsworthy development  
       - Include key stakeholders and their actions  
       - Add critical numbers/data if relevant  
       - Explain why this matters now  
       - Mention immediate implications  
  
    2. Style Guidelines:  
       - Use strong, active verbs  
       - Be specific, not general  
       - Maintain journalistic objectivity  
       - Make every word count  
       - Explain technical terms if necessary  
  
    Format: Create a single paragraph of 250-400 words that informs and engages.  
    Pattern: [Major News] + [Key Details/Data] + [Why It Matters/What's Next]  
  
    Focus on answering: What happened? Why is it significant? What's the impact?  
  
    IMPORTANT: Provide ONLY the summary paragraph. Do not include any introductory phrases,   
    labels, or meta-text like "Here's a summary" or "In AP/Reuters style."  
    Start directly with the news content.  
    """,  
    model=MODEL  
)  
  
# 实施新闻处理工作流程，按顺序处理，显示进度指标  
def process_news(topic):  
    """Run the news processing workflow"""  
    with st.status("Processing news...", expanded=True) as status:  
        # Search  
        status.write("🔍 Searching for news...")  
        search_response = client.run(  
            agent=search_agent,  
            messages=[{"role": "user", "content": f"Find recent news about {topic}"}]  
        )  
        raw_news = search_response.messages[-1]["content"]  
          
        # Synthesize  
        status.write("🔄 Synthesizing information...")  
        synthesis_response = client.run(  
            agent=synthesis_agent,  
            messages=[{"role": "user", "content": f"Synthesize these news articles:\n{raw_news}"}]  
        )  
        synthesized_news = synthesis_response.messages[-1]["content"]  
          
        # Summarize  
        status.write("📝 Creating summary...")  
        summary_response = client.run(  
            agent=summary_agent,  
            messages=[{"role": "user", "content": f"Summarize this synthesis:\n{synthesized_news}"}]  
        )  
        return raw_news, synthesized_news, summary_response.messages[-1]["content"]  
  
# 用户交互界面  
topic = st.text_input("Enter news topic:", value="artificial intelligence")  
if st.button("Process News", type="primary"):  
    if topic:  
        try:  
            raw_news, synthesized_news, final_summary = process_news(topic)  
            st.header(f"📝 News Summary: {topic}")  
            st.markdown(final_summary)  
        except Exception as e:  
            st.error(f"An error occurred: {str(e)}")  
    else:  
        st.error("Please enter a topic!")