app.py

import streamlit as st
import logging
import requests
import json
from typing import Optional

# 配置日志
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

def generate_email_with_vveai(
    profile1: str, 
    profile2: str, 
    style: str = "friendly", 
    temperature: float = 0.7,
    api_key: Optional[str] = None
) -> str:
    """
    使用VVEAI API基于两个LinkedIn简介生成个性化咖啡聊天邀请邮件。
    
    参数:
        profile1 (str): 发件人的LinkedIn简介文本
        profile2 (str): 收件人的LinkedIn简介文本
        style (str): 邮件风格，可选项为 "formal", "friendly", "humorous"
        temperature (float): 生成文本的创造性程度 (0.0-1.0)
        api_key (str, optional): VVEAI API密钥
        
    返回:
        str: 生成的个性化咖啡聊天邮件
    """
    try:
        # 验证输入
        if not profile1 or not profile2:
            raise ValueError("两个简介都不能为空")
            
        valid_styles = ["formal", "friendly", "humorous"]
        if style not in valid_styles:
            logging.warning(f"不支持的风格 '{style}'，使用默认的 'friendly'")
            style = "friendly"
            
        if temperature < 0.0 or temperature > 1.0:
            logging.warning(f"温度 {temperature} 超出范围，调整至 0.7")
            temperature = 0.7
        
        # 验证API密钥
        if not api_key:
            raise ValueError("请提供VVEAI API密钥")
            
        # 构建提示词
        prompt = f"""
        I need to write a personalized coffee chat invitation email based on two LinkedIn profiles.

        MY PROFILE:
        {profile1}

        TARGET PERSON'S PROFILE:
        {profile2}

        TASK:
        1. Analyze both profiles to identify meaningful common points or connections (universities, companies, industries, skills, interests, etc.)
        2. Write a personalized coffee chat invitation email from me to the target person
        3. The email should:
           - Start with a brief introduction of myself
           - Mention 2-3 specific commonalities or interesting points from our profiles
           - Express genuine interest in their work/experience
           - Include a clear but polite request for a 15-30 minute coffee chat
           - Suggest flexibility for their schedule
           - End with a professional closing

        STYLE:
        The tone should be {style}. If formal: professional and respectful; if friendly: warm and conversational; if humorous: light and engaging but still professional.

        IMPORTANT GUIDELINES:
        - Keep the email concise (150-250 words)
        - Make it feel authentic and human, not templated
        - Focus on creating a genuine connection
        - Avoid generic flattery
        - Be specific about shared connections or interests
        - Don't be pushy or demanding

        Format the response as a complete, ready-to-send email without explanations.
        """
        
        # 准备API请求
        system_message = "You are an expert email writer who helps professionals create personalized networking messages."
        
        # VVEAI API端点
        url = "https://api.vveai.com/v1/chat/completions"
        
        # 构建请求
        headers = {
            "Content-Type": "application/json",
            "Authorization": f"Bearer {api_key}"
        }
        
        payload = {
            "model": "gpt-4",  # 可能需要根据VVEAI支持的模型调整
            "messages": [
                {"role": "system", "content": system_message},
                {"role": "user", "content": prompt}
            ],
            "temperature": temperature,
            "max_tokens": 1000
        }
        
        # 发送请求
        response = requests.post(url, headers=headers, data=json.dumps(payload), timeout=60)
        
        # 添加详细日志便于调试
        logging.info(f"API响应状态码: {response.status_code}")
        logging.info(f"API响应头: {response.headers}")
        
        # 处理响应
        if response.status_code == 200:
            result = response.json()
            # 记录完整响应便于调试
            logging.info(f"API响应内容: {result}")
            try:
                email_text = result["choices"][0]["message"]["content"].strip()
                return email_text
            except KeyError as e:
                logging.error(f"响应格式错误: {e}. 完整响应: {result}")
                return f"API响应格式错误: {e}。请联系VVEAI支持获取帮助。"
        else:
            error_message = f"API请求失败: {response.status_code} - {response.text}"
            logging.error(error_message)
            return f"API错误: {response.status_code}。请检查您的API密钥是否正确。详细信息: {response.text}"
        
    except ValueError as e:
        logging.error(f"输入错误: {str(e)}")
        return f"发生错误: {str(e)}"
    except requests.RequestException as e:
        logging.error(f"请求错误: {str(e)}")
        return f"网络请求错误: {str(e)}"
    except Exception as e:
        logging.error(f"意外错误: {str(e)}")
        return f"发生未预期的错误: {str(e)}"

# 页面配置
st.set_page_config(
    page_title="个性化咖啡聊天邮件生成器",
    page_icon="☕",
    layout="wide"
)

# 标题和说明
st.title("☕ 个性化咖啡聊天邮件生成器")
st.markdown("""
此应用可以帮助求职者生成个性化的咖啡聊天邀请邮件。
只需输入您的简介和目标联系人的简介，系统将分析共同点并生成一封真诚的邀请邮件。

**注意：此应用使用 VVEAI API 服务，请确保输入正确的 API 密钥。**
""")

# 侧边栏 - API设置
with st.sidebar:
    st.header("API设置")
    
    api_key = st.text_input("VVEAI API密钥", type="password", key="api_key", 
                           help="输入以 sk- 开头的 VVEAI API 密钥")
    
    # 添加API状态检测按钮
    if st.button("测试API连接"):
        if not api_key:
            st.error("请先输入API密钥")
        else:
            try:
                # 简单的API测试请求
                test_url = "https://api.vveai.com/v1/chat/completions"
                test_headers = {
                    "Content-Type": "application/json",
                    "Authorization": f"Bearer {api_key}"
                }
                test_payload = {
                    "model": "gpt-4",
                    "messages": [
                        {"role": "user", "content": "Hello"}
                    ],
                    "max_tokens": 5
                }
                
                test_response = requests.post(
                    test_url, 
                    headers=test_headers, 
                    data=json.dumps(test_payload),
                    timeout=10
                )
                
                if test_response.status_code == 200:
                    st.success("API连接成功！")
                else:
                    st.error(f"API连接失败: {test_response.status_code} - {test_response.text}")
            except Exception as e:
                st.error(f"测试API时出错: {str(e)}")
    
    st.subheader("邮件风格")
    style = st.radio(
        "选择风格:",
        ["formal", "friendly", "humorous"],
        index=1,
        format_func=lambda x: {
            "formal": "正式专业",
            "friendly": "友好温暖",
            "humorous": "幽默轻松"
        }[x]
    )
    
    temperature = st.slider("创造性程度", min_value=0.0, max_value=1.0, value=0.7, step=0.1,
                     help="较低的值使输出更可预测，较高的值使输出更创意")

# 主界面 - 两栏布局
col1, col2 = st.columns(2)

with col1:
    st.header("您的简介")
    profile1 = st.text_area(
        "粘贴您的LinkedIn简介或专业背景", 
        height=300,
        placeholder="例如：\n我是一名拥有5年经验的金融分析师，毕业于北京大学金融学专业。我在摩根士丹利工作了3年，专注于投资银行业务..."
    )
    
with col2:
    st.header("目标联系人的简介")
    profile2 = st.text_area(
        "粘贴目标联系人的LinkedIn简介或专业背景", 
        height=300,
        placeholder="例如：\n高级投资银行家，现任高盛副总裁。清华大学MBA毕业，擅长并购交易，在科技行业有丰富经验..."
    )

# 生成按钮和结果显示
submit_disabled = not api_key or not profile1 or not profile2
if submit_disabled:
    if not api_key:
        reason = "请先输入VVEAI API密钥"
    else:
        reason = "请填写两个简介"
    st.warning(reason)

if st.button("生成邮件", type="primary", disabled=submit_disabled):
    with st.spinner("正在分析简介并创建个性化邮件..."):
        email = generate_email_with_vveai(
            profile1=profile1,
            profile2=profile2,
            style=style,
            temperature=temperature,
            api_key=api_key
        )
        
    # 检查是否发生错误
    if email.startswith("发生错误") or email.startswith("API错误") or email.startswith("网络请求错误"):
        st.error(email)
    else:
        st.success("邮件已生成！")
        
        # 显示结果
        st.header("您的个性化咖啡聊天邮件")
        st.text_area("可复制的邮件内容", email, height=400)
        st.download_button(
            label="下载邮件",
            data=email,
            file_name="咖啡聊天邀请.txt",
            mime="text/plain"
        )

# 调试信息（可选，帮助解决API问题）
if st.checkbox("显示调试信息"):
    st.subheader("API信息")
    st.markdown("""
    - API端点: `https://api.vveai.com/v1/chat/completions`
    - 请求方法: POST
    - 内容类型: application/json
    - 授权方式: Bearer Token
    
    如果遇到持续的API问题，请联系VVEAI支持团队并提供以上信息。
    """)

# 使用提示
st.markdown("---")
st.subheader("💡 使用提示")
st.markdown("""
- **简介内容越详细**，生成的邮件就越个性化和有针对性
- **尝试不同风格**以适应不同的行业和联系人
- 生成后可以**微调邮件内容**，添加您认为合适的个人细节
- 适合用于**金融、咨询和科技行业**的专业人士
- 如果API连接测试失败，请确认您的**VVEAI API密钥**是否正确
""")