src/streamlit_app.py

import streamlit as st
from openai import OpenAI
import time
from datasets import load_dataset # 导入Hugging Face datasets库

# =======================================================================
# 1. 导入您的类和数据
# =======================================================================

# 请将 'your_agent_file' 替换为包含 MsPatient 类的实际文件名
from ms_patient import MsPatient 

# 使用Streamlit的缓存功能来加载和缓存数据集
@st.cache_data
def load_hf_dataset():
    """
    从Hugging Face Hub加载并缓存数据集。
    这将返回一个字典列表，每个字典代表一个病人数据。
    """
    try:
        # 加载'train'分割部分的数据
        dataset = load_dataset("sci-m-wang/Anna-CPsyCounD", split='train')
        # 转换为pandas DataFrame再转为字典列表，方便处理
        return dataset.to_pandas().to_dict('records')
    except Exception as e:
        st.error(f"从Hugging Face加载数据集失败: {e}")
        return []

# 加载数据
ALL_PATIENTS = load_hf_dataset()


# =======================================================================
# 2. Streamlit 应用界面
# =======================================================================

# --- 页面配置 ---
st.set_page_config(
    page_title="与Anna对话",
    page_icon="�",
    layout="wide"
)

# --- 自定义CSS样式 ---
st.markdown("""
<style>
    /* 主聊天容器 */
    .st-emotion-cache-1y4p8pa { 
        padding-top: 2rem;
    }
    /* 聊天消息 */
    .st-chat-message {
        border-radius: 0.8rem;
        padding: 0.9rem 1.2rem;
        box-shadow: 0 2px 5px rgba(0,0,0,0.05);
        background-color: #ffffff;
    }
    .st-chat-message[data-testid="chat-message-container-user"] {
        background-color: #dcf8c6;
    }
    /* 侧边栏 */
    .st-sidebar {
        background-color: #f8f9fa;
        border-right: 1px solid #e9ecef;
    }
    .st-sidebar h2 {
        color: #343a40;
    }
    .st-expanderHeader {
        font-size: 1.1rem;
        font-weight: 600;
    }
</style>
""", unsafe_allow_html=True)


# --- 初始化 Session State ---
if "patient_agent" not in st.session_state:
    st.session_state.patient_agent = None
if "messages" not in st.session_state:
    st.session_state.messages = []
if "selected_patient_id" not in st.session_state:
    st.session_state.selected_patient_id = None
if "openai_client" not in st.session_state:
    st.session_state.openai_client = None
if "model_name" not in st.session_state:
    st.session_state.model_name = "gpt-4o-mini" # 默认模型

# --- 侧边栏 ---
with st.sidebar:
    st.title("👩 AnnaAgent 设置")
    st.markdown("---")

    # API Key 输入
    with st.expander("🔑 API 设置", expanded=True):
        api_key = st.text_input("输入您的 OpenAI API Key", type="password", help="您的API Key将仅用于本次会话，不会被储存。")
        base_url = st.text_input("API Base URL (可选)", value="https://api.openai.com/v1")
        model_name = st.text_input("模型名称", value=st.session_state.model_name)

        if st.button("连接模型"):
            if api_key:
                try:
                    st.session_state.openai_client = OpenAI(api_key=api_key, base_url=base_url)
                    st.session_state.model_name = model_name
                    st.success("连接成功！")
                    if st.session_state.patient_agent:
                        st.session_state.patient_agent.client = st.session_state.openai_client
                except Exception as e:
                    st.error(f"连接失败: {e}")
            else:
                st.warning("请输入API Key。")

    st.markdown("---")

    # 病人选择
    if not ALL_PATIENTS:
        st.error("无法加载病人数据，请检查网络连接或数据集名称。")
    else:
        patient_options = {p["id"]: f"{p['portrait']['gender']}，{p['portrait']['age']}岁 - {p['portrait']['symptom']}" for p in ALL_PATIENTS}
        selected_id = st.selectbox(
            "选择一位病人进行对话",
            options=list(patient_options.keys()),
            format_func=lambda x: patient_options[x]
        )

        # 当选择的病人变化时，重置状态
        if st.session_state.selected_patient_id != selected_id:
            st.session_state.selected_patient_id = selected_id
            selected_patient_data = next((p for p in ALL_PATIENTS if p["id"] == selected_id), None)
            
            with st.spinner("正在生成病人角色..."):
                st.session_state.patient_agent = MsPatient(
                    portrait=selected_patient_data["portrait"],
                    report=selected_patient_data["report"],
                    previous_conversations=selected_patient_data["conversation"],
                    language="Chinese",
                    client=st.session_state.openai_client
                )
            st.session_state.messages = [{"role": "assistant", "content": "你好，医生..."}]
            st.rerun()

    # 显示病人信息
    if st.session_state.patient_agent:
        st.markdown("---")
        st.subheader("病人信息")
        agent = st.session_state.patient_agent
        st.info(f"""
        **基本情况**: {agent.portrait['gender']}, {agent.portrait['age']}岁, {agent.portrait['occupation']}, {agent.portrait['marital_status']}
        
        **近期状态**: {agent.status}
        """)

        with st.expander("查看完整系统提示 (System Prompt)"):
            st.code(agent.get_system_prompt(), language='markdown')


# --- 主聊天界面 ---
st.title("💬 与 Anna 对话")
st.caption("这是一个模拟心理咨询的AI Agent。由 `MsPatient` 类驱动。")

# 显示聊天记录
for message in st.session_state.messages:
    avatar = "👩" if message["role"] == "assistant" else "🧑‍⚕️"
    with st.chat_message(message["role"], avatar=avatar):
        st.markdown(message["content"])

# 获取用户输入
if prompt := st.chat_input("请输入您想说的话..."):
    st.session_state.messages.append({"role": "user", "content": prompt})
    with st.chat_message("user", avatar="🧑‍⚕️"):
        st.markdown(prompt)

    if st.session_state.patient_agent:
        with st.chat_message("assistant", avatar="👩"):
            with st.spinner("Anna正在思考..."):
                response = st.session_state.patient_agent.chat(prompt)
                st.markdown(response)
        
        st.session_state.messages.append({"role": "assistant", "content": response})
    else:
        st.warning("请先在左侧选择一位病人并配置API Key。")