Update app.py

app.py

@@ -1,64 +1,114 @@
+from openai import OpenAI
+from docx import Document
+import numpy as np
+import faiss
+from sentence_transformers import SentenceTransformer
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+# 配置参数
+WORD_DOC_PATH = "知识库.docx"  # Word文档路径
+VECTOR_INDEX_PATH = "faiss_index.index"  # 向量索引保存路径
+TEXT_DATA_PATH = "text_data.npy"  # 文本数据保存路径
+EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"  # 嵌入模型
 
+# 初始化模型和客户端
+client = OpenAI(
+    base_url='https://api-inference.modelscope.cn/v1/',
+    api_key='7ed44f86-e2c6-4b85-9c4a-26eacfc2e5ee',
+)
+embedder = SentenceTransformer(EMBEDDING_MODEL)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+def process_word_document():
+    """处理Word文档并分块（保持原有实现不变）"""
+    doc = Document(WORD_DOC_PATH)
+    chunks = []
+    current_chunk = []
+    chunk_size = 300
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    for para in doc.paragraphs:
+        text = para.text.strip()
+        if text:
+            words = text.split()
+            current_chunk.extend(words)
 
-    messages.append({"role": "user", "content": message})
+            while len(current_chunk) > chunk_size:
+                chunks.append(" ".join(current_chunk[:chunk_size]))
+                current_chunk = current_chunk[chunk_size:]
 
-    response = ""
+    if current_chunk:
+        chunks.append(" ".join(current_chunk))
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
+    return chunks
+
+def create_vector_store():
+    """创建并保存向量存储（保持原有实现不变）"""
+    if os.path.exists(VECTOR_INDEX_PATH):
+        return
+
+    chunks = process_word_document()
+    embeddings = embedder.encode(chunks, convert_to_tensor=False)
+    embeddings = np.array(embeddings).astype('float32')
+
+    dimension = embeddings.shape[1]
+    index = faiss.IndexFlatL2(dimension)
+    index.add(embeddings)
+
+    faiss.write_index(index, VECTOR_INDEX_PATH)
+    np.save(TEXT_DATA_PATH, np.array(chunks))
+
+def search_knowledge(query, top_k=3):
+    """知识检索（保持原有实现不变）"""
+    index = faiss.read_index(VECTOR_INDEX_PATH)
+    text_data = np.load(TEXT_DATA_PATH, allow_pickle=True)
+
+    query_embedding = embedder.encode([query], convert_to_tensor=False)
+    query_embedding = np.array(query_embedding).astype('float32')
+
+    distances, indices = index.search(query_embedding, top_k)
+    return "\n".join([text_data[i] for i in indices[0]])
+
+def respond(message, history, max_tokens, temperature, top_p):
+    """Gradio响应函数"""
+    # 检索相关知识
+    context = search_knowledge(message)
+    
+    # 构建对话消息
+    messages = [
+        {"role": "system", "content": f"基于以下知识回答问题，如果不知道就说不知道：\n{context}"},
+        {"role": "user", "content": message}
+    ]
+
+    # 流式生成响应
+    full_response = ""
+    response = client.chat.completions.create(
+        model='deepseek-ai/DeepSeek-R1',
+        messages=messages,
         stream=True,
+        max_tokens=max_tokens,
         temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+        top_p=top_p
+    )
 
-        response += token
-        yield response
+    for chunk in response:
+        content = chunk.choices[0].delta.content or ""
+        full_response += content
+        yield full_response
 
+# 初始化向量存储
+create_vector_store()
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# 创建Gradio界面
 demo = gr.ChatInterface(
-    respond,
+    fn=respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
+        gr.Slider(512, 2048, value=512, step=1, label="最大Token数"),
+        gr.Slider(0.1, 2.0, value=0.7, step=0.1, label="温度参数"),
+        gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top-p采样"),
     ],
+    title="制度文档问答系统",
+    description="输入关于广西警察学院制度的问题进行问答"
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()