Update app.py

app.py

@@ -1,65 +1,52 @@
-import os
 import gradio as gr
 from huggingface_hub import InferenceClient
-from langchain_huggingface import HuggingFaceEmbeddings
-from langchain_community.vectorstores import FAISS
-from langchain_community.document_loaders import TextLoader
-from langchain_text_splitters import CharacterTextSplitter
 
-# 1. 初始化官方推理客户端 (直接绕过 LangChain 不兼容的 Endpoint)
-client = InferenceClient(
-    model="Qwen/Qwen2.5-7B-Instruct",
-    token=os.getenv("HF_TOKEN")
-)
+# 初始化客户端
+client = InferenceClient("your-model-id")
 
-# 2. 知识库加载与向量化
-def init_vector_db():
-    if not os.path.exists("knowledge.txt"):
-        with open("knowledge.txt", "w", encoding="utf-8") as f:
-            f.write("私有大脑已上线。")
-    
-    loader = TextLoader("knowledge.txt", encoding="utf-8")
-    # 按照语义切分，防止回答断章取义
-    text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=50)
-    docs = text_splitter.split_documents(loader.load())
-    
-    # 使用轻量级中文向量模型
-    embeddings = HuggingFaceEmbeddings(model_name="BAAI/bge-small-zh-v1.5")
-    return FAISS.from_documents(docs, embeddings)
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    messages = [{"role": "system", "content": system_message}]
+    for val in history:
+        if val[0]: messages.append({"role": "user", "content": val[0]})
+        if val[1]: messages.append({"role": "assistant", "content": val[1]})
 
-vector_db = init_vector_db()
+    messages.append({"role": "user", "content": message})
 
-# 3. 核心聊天逻辑
-def chat_fn(message, history):
-    try:
-        # 第一步：在知识库中寻找最相关的片段
-        docs = vector_db.similarity_search(message, k=3)
-        context = "\n".join([doc.page_content for doc in docs])
-        
-        # 第二步：构建提示词（Prompt）
-        prompt = f"你是全能私有大脑。请参考以下已知信息回答用户问题。\n\n已知信息：\n{context}\n\n问题：{message}\n回答："
-        
-        # 第三步：使用官方最新方法进行推理
-        response = ""
-        for token in client.chat_completion(
-            messages=[{"role": "user", "content": prompt}],
-            max_tokens=500,
-            stream=True
-        ):
-            token_str = token.choices[0].delta.content
-            if token_str:
-                response += token_str
-        return response
+    response = ""
+    # 修复了旧版 InferenceClient 可能存在的迭代问题，确保流式输出顺滑
+    for message in client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = message.choices[0].delta.content
+        response += token
+        yield response
+
+# 升级后的专业 UI 布局
+with gr.Blocks(theme=gr.themes.Soft(), css=".gradio-container {background-color: #0b0f19;}") as demo:
+    gr.Markdown("# 全能私有大脑 v4.0 (Ultra Stable)")
+    
+    with gr.Row():
+        # 左侧控制区
+        with gr.Column(scale=1):
+            system_input = gr.Textbox(value="你是一个资深安全审计专家...", label="系统指令")
+            with gr.Accordion("高级参数设置", open=False):
+                temp = gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature")
+                tokens = gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens")
+            
+            # 实时渲染预览区 (例如用于显示代码执行或漏洞拓扑)
+            render_box = gr.HTML("✨ 实时安全分析渲染就绪...")
 
-    except Exception as e:
-        return f"大脑响应异常，请尝试在 Settings 中 Factory Restart。错误详情: {str(e)}"
+        # 右侧聊天区
+        with gr.Column(scale=2):
+            chat = gr.Chatbot(height=600, show_copy_button=True, avatar_images=(None, "https://path-to-your-icon.png"))
+            msg = gr.Textbox(placeholder="输入指令进行分析...", container=False)
+            clear = gr.ClearButton([msg, chat])
 
-# 4. 界面设计
-demo = gr.ChatInterface(
-    chat_fn,
-    title="全能私有大脑 v3.0 (终极稳定版)",
-    description="已彻底解决 InferenceClient 兼容性问题。现在可以流畅调取私有知识库了。"
-)
+    msg.submit(respond, [msg, chat, system_input, tokens, temp], [chat])
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()