渲染错误: Mermaid 渲染失败: Parse error on line 10: ...t tools 绑定的工具列表 } Conversation | ----------------------^ Expecting 'ATTRIBUTE_WORD', got 'BLOCK_STOP'

pip install pyautogen python-dotenv yfinance matplotlib

OPENAI_API_KEY=你的API密钥

import os
from dotenv import load_dotenv
import autogen

# 加载环境变量
load_dotenv()

# 配置大模型参数
llm_config = {
    "model": "gpt-4o",
    "api_key": os.getenv("OPENAI_API_KEY"),
    "temperature": 0.1, # 低温度保证输出稳定
    "max_tokens": 2048,
    "timeout": 120
}

# 1. 定义助理Agent：数据分析师角色，负责生成代码和分析结果
data_analyst = autogen.AssistantAgent(
    name="数据分析师",
    llm_config=llm_config,
    system_message="""你是资深金融数据分析师，擅长Python数据分析。
    规则：
    1. 当需要执行操作时，只输出Python代码块，不需要额外说明
    2. 代码执行结果出来后，根据结果进行分析
    3. 当任务完全完成时，输出TERMINATE表示结束
    """
)

# 2. 定义用户代理Agent：代表用户执行代码，返回结果
user_proxy = autogen.UserProxyAgent(
    name="用户代理",
    human_input_mode="TERMINATE", # 只有收到TERMINATE时才需要用户确认
    max_consecutive_auto_reply=10, # 最多自动回复10次，避免无限循环
    code_execution_config={
        "work_dir": "./stock_analysis", # 代码执行的工作目录，自动创建
        "use_docker": False # 生产环境建议设置为True，用沙箱执行代码避免安全风险
    },
    system_message="你是用户的代理，负责执行代码并返回执行结果，不要修改代码内容。"
)

# 3. 发起对话
if __name__ == "__main__":
    user_proxy.initiate_chat(
        data_analyst,
        message="""
        请完成以下任务：
        1. 获取2024年1月1日到2024年5月31日的特斯拉（TSLA）股票日线数据
        2. 画出收盘价的折线图，包含5日均线和20日均线，保存为tsla_trend.png
        3. 分析这段时间的价格走势，给出3个核心结论
        """
    )

# 安装核心依赖
pip install pyautogen python-dotenv pandas numpy matplotlib seaborn scikit-learn pdfkit flask

# 安装wkhtmltopdf用于生成PDF报告（Ubuntu为例）
sudo apt install wkhtmltopdf
# Mac用户可以用 brew install wkhtmltopdf

渲染错误: Mermaid 渲染失败: Parsing failed: Lexer error on line 2, column 15: unexpected character: ->(<- at offset: 32, skipped 10 characters. Lexer error on line 3, column 16: unexpected character: ->(<- at offset: 58, skipped 1 characters. Lexer error on line 3, column 20: unexpected character: ->界<- at offset: 62, skipped 16 characters. Lexer error on line 4, column 16: unexpected character: ->(<- at offset: 102, skipped 1 characters. Lexer error on line 4, column 25: unexpected character: ->)<- at offset: 111, skipped 2 characters. Lexer error on line 4, column 30: unexpected character: ->网<- at offset: 116, skipped 3 characters. Lexer error on line 6, column 16: unexpected character: ->(<- at offset: 144, skipped 1 characters. Lexer error on line 6, column 22: unexpected character: ->层<- at offset: 150, skipped 10 characters. Lexer error on line 7, column 15: unexpected character: ->(<- at offset: 175, skipped 5 characters. Lexer error on line 7, column 25: unexpected character: ->)<- at offset: 185, skipped 7 characters. Lexer error on line 8, column 15: unexpected character: ->(<- at offset: 216, skipped 5 characters. Lexer error on line 8, column 25: unexpected character: ->)<- at offset: 226, skipped 8 characters. Lexer error on line 9, column 15: unexpected character: ->(<- at offset: 258, skipped 5 characters. Lexer error on line 9, column 25: unexpected character: ->)<- at offset: 268, skipped 8 characters. Lexer error on line 10, column 16: unexpected character: ->(<- at offset: 301, skipped 3 characters. Lexer error on line 10, column 24: unexpected character: ->)<- at offset: 309, skipped 8 characters. Lexer error on line 11, column 16: unexpected character: ->(<- at offset: 342, skipped 5 characters. Lexer error on line 11, column 26: unexpected character: ->)<- at offset: 352, skipped 7 characters. Lexer error on line 12, column 20: unexpected character: ->(<- at offset: 388, skipped 7 characters. Lexer error on line 14, column 15: unexpected character: ->(<- at offset: 438, skipped 11 characters. Lexer error on line 15, column 22: unexpected character: ->(<- at offset: 471, skipped 9 characters. Lexer error on line 15, column 37: unexpected character: ->/<- at offset: 486, skipped 1 characters. Lexer error on line 15, column 43: unexpected character: ->]<- at offset: 492, skipped 1 characters. Lexer error on line 16, column 21: unexpected character: ->(<- at offset: 522, skipped 8 characters. Lexer error on line 16, column 39: unexpected character: ->/<- at offset: 540, skipped 1 characters. Lexer error on line 16, column 47: unexpected character: ->]<- at offset: 548, skipped 1 characters. Lexer error on line 17, column 20: unexpected character: ->(<- at offset: 577, skipped 7 characters. Lexer error on line 17, column 39: unexpected character: ->/<- at offset: 596, skipped 1 characters. Lexer error on line 17, column 47: unexpected character: ->]<- at offset: 604, skipped 1 characters. Lexer error on line 18, column 24: unexpected character: ->(<- at offset: 637, skipped 9 characters. Lexer error on line 18, column 41: unexpected character: ->/<- at offset: 654, skipped 1 characters. Lexer error on line 18, column 48: unexpected character: ->]<- at offset: 661, skipped 1 characters. Lexer error on line 20, column 16: unexpected character: ->(<- at offset: 687, skipped 12 characters. Lexer error on line 21, column 17: unexpected character: ->(<- at offset: 716, skipped 1 characters. Lexer error on line 21, column 31: unexpected character: ->)<- at offset: 730, skipped 1 characters. Lexer error on line 22, column 19: unexpected character: ->(<- at offset: 759, skipped 1 characters. Lexer error on line 22, column 31: unexpected character: ->)<- at offset: 771, skipped 1 characters. Lexer error on line 23, column 17: unexpected character: ->(<- at offset: 798, skipped 5 characters. Lexer error on line 23, column 23: unexpected character: ->)<- at offset: 804, skipped 1 characters. Parse error on line 3, column 17: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Web' Parse error on line 3, column 37: Expecting token of type ':' but found `in`. Parse error on line 4, column 17: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'R' Parse error on line 4, column 22: Expecting token of type ':' but found `API`. Parse error on line 4, column 27: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'API' Parse error on line 4, column 34: Expecting token of type ':' but found `in`. Parse error on line 6, column 17: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 6, column 32: Expecting token of type ':' but found ` `. Parse error on line 7, column 20: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 7, column 33: Expecting token of type ':' but found `in`. Parse error on line 8, column 20: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 8, column 34: Expecting token of type ':' but found `in`. Parse error on line 9, column 20: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 9, column 34: Expecting token of type ':' but found `in`. Parse error on line 10, column 19: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 10, column 33: Expecting token of type ':' but found `in`. Parse error on line 11, column 21: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Agent' Parse error on line 11, column 34: Expecting token of type ':' but found `in`. Parse error on line 15, column 31: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Pandas' Parse error on line 15, column 38: Expecting token of type ':' but found `Numpy`. Parse error on line 15, column 45: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'in' Parse error on line 15, column 52: Expecting token of type ':' but found ` `. Parse error on line 16, column 29: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Matplotlib' Parse error on line 16, column 40: Expecting token of type ':' but found `Seaborn`. Parse error on line 16, column 49: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'in' Parse error on line 16, column 56: Expecting token of type ':' but found ` `. Parse error on line 17, column 27: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Scikit-learn' Parse error on line 17, column 40: Expecting token of type ':' but found `XGBoost`. Parse error on line 17, column 49: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'in' Parse error on line 17, column 56: Expecting token of type ':' but found ` `. Parse error on line 18, column 33: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'Markdown' Parse error on line 18, column 42: Expecting token of type ':' but found `PDFKit`. Parse error on line 18, column 50: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'in' Parse error on line 18, column 57: Expecting token of type ':' but found ` `. Parse error on line 21, column 18: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'OpenAI' Parse error on line 21, column 25: Expecting token of type ':' but found `GPT-4o`. Parse error on line 21, column 33: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'in' Parse error on line 21, column 41: Expecting token of type ':' but found ` `. Parse error on line 22, column 20: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: 'L' Parse error on line 22, column 26: Expecting token of type ':' but found `3`. Parse error on line 22, column 28: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: '70B' Parse error on line 22, column 33: Expecting token of type ':' but found `in`. Parse error on line 23, column 22: Expecting: one of these possible Token sequences: 1. [NEWLINE] 2. [EOF] but found: '4' Parse error on line 23, column 25: Expecting token of type ':' but found `in`. Parse error on line 25, column 9: Expecting token of type ':' but found `--`. Parse error on line 25, column 13: Expecting token of type 'ARROW_DIRECTION' but found `api`. Parse error on line 26, column 9: Expecting token of type ':' but found `--`. Parse error on line 26, column 13: Expecting token of type 'ARROW_DIRECTION' but found `manager`. Parse error on line 27, column 13: Expecting token of type ':' but found `--`. Parse error on line 27, column 17: Expecting token of type 'ARROW_DIRECTION' but found `pm`. Parse error on line 28, column 13: Expecting token of type ':' but found `--`. Parse error on line 28, column 17: Expecting token of type 'ARROW_DIRECTION' but found `de`. Parse error on line 29, column 13: Expecting token of type ':' but found `--`. Parse error on line 29, column 17: Expecting token of type 'ARROW_DIRECTION' but found `da`. Parse error on line 30, column 13: Expecting token of type ':' but found `--`. Parse error on line 30, column 17: Expecting token of type 'ARROW_DIRECTION' but found `mle`. Parse error on line 31, column 13: Expecting token of type ':' but found `--`. Parse error on line 31, column 17: Expecting token of type 'ARROW_DIRECTION' but found `rep`. Parse error on line 32, column 8: Expecting token of type ':' but found `--`. Parse error on line 32, column 12: Expecting token of type 'ARROW_DIRECTION' but found `data_tool`. Parse error on line 33, column 8: Expecting token of type ':' but found `--`. Parse error on line 33, column 12: Expecting token of type 'ARROW_DIRECTION' but found `data_tool`. Parse error on line 34, column 8: Expecting token of type ':' but found `--`. Parse error on line 34, column 12: Expecting token of type 'ARROW_DIRECTION' but found `vis_tool`. Parse error on line 35, column 9: Expecting token of type ':' but found `--`. Parse error on line 35, column 13: Expecting token of type 'ARROW_DIRECTION' but found `ml_tool`. Parse error on line 36, column 9: Expecting token of type ':' but found `--`. Parse error on line 36, column 13: Expecting token of type 'ARROW_DIRECTION' but found `report_tool`. Parse error on line 37, column 11: Expecting token of type ':' but found `--`. Parse error on line 37, column 15: Expecting token of type 'ARROW_DIRECTION' but found `model`.

import pandas as pd
import os
import pdfkit
from typing import Annotated

# 工具1：获取数据概要信息
def get_data_summary(file_path: Annotated[str, "CSV文件的路径"]) -> str:
    """获取CSV数据的基本统计信息，包括列名、数据类型、缺失值、样例数据"""
    try:
        df = pd.read_csv(file_path)
        summary = f"""
        数据总行数: {len(df)}
        数据总列数: {len(df.columns)}
        列名和数据类型: {dict(df.dtypes)}
        缺失值统计: {dict(df.isnull().sum())}
        前5行样例数据: {df.head().to_markdown()}
        数值列统计信息: {df.describe().to_markdown()}
        """
        return summary
    except Exception as e:
        return f"读取数据失败: {str(e)}"

# 工具2：生成PDF报告
def generate_pdf_report(markdown_content: Annotated[str, "Markdown格式的报告内容"], output_path: Annotated[str, "PDF文件保存路径"]) -> str:
    """把Markdown格式的报告转换为PDF文件"""
    try:
        # 先保存为临时md文件
        md_path = output_path.replace(".pdf", ".md")
        with open(md_path, "w", encoding="utf-8") as f:
            f.write(markdown_content)
        # 转换为PDF
        pdfkit.from_string(markdown_content, output_path)
        return f"报告生成成功，保存路径: {output_path}"
    except Exception as e:
        return f"生成报告失败: {str(e)}"

# 工具列表，供Agent调用
tools = [get_data_summary, generate_pdf_report]

import autogen
from dotenv import load_dotenv
import os

load_dotenv()

llm_config = {
    "model": "gpt-4o",
    "api_key": os.getenv("OPENAI_API_KEY"),
    "temperature": 0.1,
    "tools": tools # 绑定自定义工具
}

# 需求确认Agent：和用户确认需求，明确分析目标
pm_agent = autogen.AssistantAgent(
    name="需求确认专员",
    llm_config=llm_config,
    system_message="""你是需求确认专员，负责和用户确认数据分析的需求。
    规则：
    1. 首先调用get_data_summary工具获取上传数据的基本信息
    2. 根据数据信息和用户的初始需求，确认3个核心分析目标
    3. 需求确认后，把需求传递给数据工程师
    4. 不要写代码，只负责需求梳理
    """
)

# 数据清洗Agent：负责数据清洗、预处理
de_agent = autogen.AssistantAgent(
    name="数据工程师",
    llm_config=llm_config,
    system_message="""你是资深数据工程师，负责数据清洗和预处理。
    规则：
    1. 根据需求确认的目标，对数据进行清洗：处理缺失值、异常值、格式转换
    2. 输出清洗代码，由用户代理执行
    3. 清洗完成后，把清洗后的数据路径传递给数据分析师
    """
)

# 数据分析Agent：负责探索性分析和可视化
da_agent = autogen.AssistantAgent(
    name="数据分析师",
    llm_config=llm_config,
    system_message="""你是资深数据分析师，负责探索性分析和可视化。
    规则：
    1. 根据需求目标对清洗后的数据进行分析
    2. 生成可视化代码，保存图片到指定目录
    3. 输出分析结论，传递给算法工程师
    """
)

# 建模Agent：负责预测模型训练和评估
mle_agent = autogen.AssistantAgent(
    name="算法工程师",
    llm_config=llm_config,
    system_message="""你是资深算法工程师，负责预测模型的训练和评估。
    规则：
    1. 根据需求目标选择合适的模型进行训练
    2. 输出模型训练代码，评估模型效果
    3. 输出模型评估结论，传递给报告专员
    """
)

# 报告生成Agent：负责生成最终的PDF报告
rep_agent = autogen.AssistantAgent(
    name="报告专员",
    llm_config=llm_config,
    system_message="""你是报告专员，负责生成最终的数据分析报告。
    规则：
    1. 整合所有前面的分析结果、图表、模型结论，生成Markdown格式的报告
    2. 调用generate_pdf_report工具生成PDF报告
    3. 报告生成完成后，输出TERMINATE
    """
)

# 用户代理
user_proxy = autogen.UserProxyAgent(
    name="用户代理",
    human_input_mode="TERMINATE",
    max_consecutive_auto_reply=30,
    code_execution_config={
        "work_dir": "./data_analysis_workspace",
        "use_docker": True # 生产环境用Docker沙箱保证安全
    },
    function_map={ # 映射工具函数
        "get_data_summary": get_data_summary,
        "generate_pdf_report": generate_pdf_report
    }
)

# 自定义发言人选择规则：根据当前任务阶段选择合适的Agent
def custom_speaker_selection(last_speaker: autogen.Agent, groupchat: autogen.GroupChat) -> autogen.Agent:
    messages = groupchat.messages
    if len(messages) <= 2:
        return pm_agent # 前两轮是需求确认
    elif "数据清洗完成" in messages[-1]["content"]:
        return da_agent # 清洗完成后交给分析师
    elif "分析完成" in messages[-1]["content"]:
        return mle_agent # 分析完成后交给算法工程师
    elif "模型训练完成" in messages[-1]["content"]:
        return rep_agent # 建模完成后交给报告专员
    else:
        # 默认用大模型选择
        return groupchat.manager.select_speaker(last_speaker, groupchat)

# 初始化群聊
group_chat = autogen.GroupChat(
    agents=[pm_agent, de_agent, da_agent, mle_agent, rep_agent, user_proxy],
    messages=[],
    max_round=30,
    speaker_selection_method=custom_speaker_selection
)

# 群聊管理器
group_chat_manager = autogen.GroupChatManager(
    groupchat=group_chat,
    llm_config=llm_config
)

# 启动任务
if __name__ == "__main__":
    user_proxy.initiate_chat(
        group_chat_manager,
        message="""
        上传的文件路径是./data/sales_data.csv，是某电商平台2023年的销售数据。
        请分析销量的影响因素，预测2024年1月的销量，生成完整的分析报告。
        """
    )