Update app.py

app.py

@@ -1,64 +1,63 @@
+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")
+from langchain_community.vectorstores import FAISS
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from openai import OpenAI
 
+class CustomE5Embedding(HuggingFaceEmbeddings):
+    def embed_documents(self, texts):
+        texts = [f"passage: {t}" for t in texts]
+        return super().embed_documents(texts)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+    def embed_query(self, text):
+        return super().embed_query(f"query: {text}")
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+embedding_model = CustomE5Embedding(model_name="intfloat/multilingual-e5-small")
+db = FAISS.load_local("faiss_db", embedding_model, allow_dangerous_deserialization=True)
+retriever = db.as_retriever(search_kwargs={"k": 20})
 
-    messages.append({"role": "user", "content": message})
+client = OpenAI()
+model = "gpt-4o"
 
-    response = ""
+system_prompt = "你是清華大學校園資訊助理 AI，專門解答與課程、社團、場地與設施時段相關的問題。請根據資料內容，以台灣人熟悉的繁體中文提供清楚、簡潔且實用的回應。如無法在資料中找到答案，請誠實說明。"
 
-    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
+prompt_template = """以下是關於國立清華大學課程、社團與校內設施的資料片段：
+{retrieved_chunks}
 
-        response += token
-        yield response
+使用者的提問是：{question}
 
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
+請根據提供的內容回答問題，若提到課程請具體指出課程名稱、教師或時間地點；若提到社團或場地，也請回應具體活動內容或使用時段。
+如查無相關資訊，請明確告知查無資料。
 """
-demo = gr.ChatInterface(
-    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)",
-        ),
-    ],
-)
-
 
-if __name__ == "__main__":
-    demo.launch()
+def chat_with_rag(user_input):
+    docs = retriever.get_relevant_documents(user_input)
+    retrieved_chunks = "\n\n".join([doc.page_content for doc in docs])
+    final_prompt = prompt_template.format(retrieved_chunks=retrieved_chunks, question=user_input)
+
+    response = client.chat.completions.create(
+        model=model,
+        max_tokens=1000,
+        messages=[
+            {"role": "system", "content": system_prompt},
+            {"role": "assistant", "content": f"以下是相關資料片段：\n\n{retrieved_chunks}"},
+            {"role": "user", "content": f"問題：{user_input}\n\n請根據資料，詳細列出所有相關內容，包括課程名稱、授課教師、時間地點等，必要時請使用條列式說明。若查無資料，也請清楚說明。"}
+        ]
+    )
+    return response.choices[0].message.content
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🎓 校園資訊查詢小幫手")
+    gr.Markdown("請輸入你想詢問的清華大學的問題，包含資工系課程、校園設施、社團，我會根據資料幫你解答 💬")
+
+    chatbot = gr.Chatbot(label="📚 問答紀錄", height=600)
+    msg = gr.Textbox(placeholder="例如：大一有哪些必修課？", label="❓ 問題輸入")
+
+    def respond(message, chat_history_local):
+        response = chat_with_rag(message)
+        chat_history_local.append((message, response))
+        return "", chat_history_local
+
+    msg.submit(respond, [msg, chatbot], [msg, chatbot])
+
+demo.launch()