app.py

import gradio as gr
from utils.embedding import ChatEmbedder
from utils.retrieval import ChatQABot
import json

class ChatAnalyzerApp:
    def __init__(self):
        self.embedder = ChatEmbedder()
        self.qa_bot = ChatQABot()
        self.setup_data()
    
    def setup_data(self):
        """初始化数据和索引"""
        try:
            count = self.embedder.load_chat_data('chat_log.json')
            index_size = self.embedder.build_index()
            print(f"✓ 加载了 {count} 条聊天记录，构建了 {index_size} 维索引")
            return True
        except Exception as e:
            print(f"× 数据加载失败: {e}")
            return False
    
    def query_chat(self, question: str, top_k: int = 3):
        """处理查询"""
        # 1. 检索相关文档
        results = self.embedder.search(question, top_k)
        
        if not results:
            return "未找到相关聊天记录。", ""
        
        # 2. 组合检索到的上下文
        context = "\n".join([f"{i+1}. {r['content']} (相关度: {r['score']:.3f})" 
                           for i, r in enumerate(results)])
        
        # 3. 生成答案
        answer = self.qa_bot.generate_answer(question, context)
        
        return answer, context

def main():
    app = ChatAnalyzerApp()
    
    if not app.setup_data():
        return
    
    # 创建Gradio界面
    # with gr.Blocks(title="群聊记录分析助手", theme=gr.themes.Soft()) as demo:
    with gr.Blocks(title="群聊记录分析助手") as demo:
        demo.theme = gr.themes.Soft()  # 单独设置theme属性
        gr.Markdown("# 📱 群聊记录分析与问答系统")
        gr.Markdown("基于RAG技术，智能分析群聊记录并回答问题")
        
        with gr.Row():
            with gr.Column(scale=2):
                question_input = gr.Textbox(
                    label="输入您的问题",
                    placeholder="例如：关于新版UI设计，最终的结论是什么？",
                    lines=2
                )
                
                top_k_slider = gr.Slider(
                    minimum=1,
                    maximum=5,
                    value=3,
                    step=1,
                    label="检索结果数量"
                )
                
                submit_btn = gr.Button("🔍 搜索", variant="primary")
            
            with gr.Column(scale=3):
                answer_output = gr.Textbox(
                    label="答案",
                    lines=4,
                    interactive=False
                )
                
                context_output = gr.Textbox(
                    label="检索到的相关聊天记录",
                    lines=6,
                    interactive=False
                )
        
        # 示例问题
        examples = gr.Examples(
            examples=[
                "关于新版UI设计，最终的结论是什么？",
                "谁负责联系市场部？",
                "市场部的联系方式是什么？",
                "讨论过哪些设计方案？"
            ],
            inputs=[question_input]
        )
        
        gr.Markdown("### 功能说明")
        gr.Markdown("""
        1. 系统基于模拟的飞书/企微群聊记录构建知识库
        2. 使用中文嵌入模型检索相关聊天片段
        3. 使用大语言模型生成准确答案
        4. 支持调整检索结果数量以优化答案质量
        """)
        
        # 绑定事件
        submit_btn.click(
            fn=app.query_chat,
            inputs=[question_input, top_k_slider],
            outputs=[answer_output, context_output]
        )
        
        question_input.submit(
            fn=app.query_chat,
            inputs=[question_input, top_k_slider],
            outputs=[answer_output, context_output]
        )
    
    demo.launch()

if __name__ == "__main__":
    main()