Update app.py

app.py

@@ -1,81 +1,71 @@
-import socket
-import subprocess
-import gradio as gr
-from openai import OpenAI
-
-def get_local_ip():
-    # 建立一個 UDP socket，連到外部伺服器（不會真的發送資料）
-    s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
-    try:
-        # 這裡用 Google 的公共 DNS IP 來確保路徑有效
-        s.connect(("8.8.8.8", 80))
-        ip = s.getsockname()[0]
-    except Exception:
-        ip = "127.0.0.1"
-    finally:
-        s.close()
-    return ip
-
-print("本機 IP:", get_local_ip())
-
-# ✅ 設定 base URL 連接本地 llama.cpp API
-client = OpenAI(
-    base_url="http://0.0.0.0:8000/v1",
-    api_key="sk-local",  # llama.cpp 不檢查內容，只要有就行
-    timeout=600
-)
-
-# ✅ 回應函式（改成 stream 模式）
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-    
-    for user, assistant in history:
-        if user:
-            messages.append({"role": "user", "content": user})
-        if assistant:
-            messages.append({"role": "assistant", "content": assistant})
-
-    messages.append({"role": "user", "content": message})
-
-    try:
-        # 🔹 修改 1: 開啟 stream 模式
-        stream = client.chat.completions.create(
-            model="qwen3",  # ⚠️ 替換成你 llama.cpp 載入的模型名稱
-            messages=messages,
-            max_tokens=max_tokens,
-            temperature=temperature,
-            top_p=top_p,
-            stream=True,
-        )
-
-        output = ""
-        # 🔹 修改 2: 逐步處理流式回應
-        for chunk in stream:
-            delta = chunk.choices[0].delta.content or ""
-            output += delta
-            yield output  # ✅ 即時回傳給 Gradio
-
-    except Exception as e:
-        print(f"[Error] {e}")
-        yield "⚠️ Llama.cpp server 沒有回應，請稍後再試。"
-
-# ✅ Gradio 介面（修改 3: 啟用 generator）
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly assistant.", label="System message"),
-        gr.Slider(minimum=1, maximum=4096, value=1024, 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()
+### app.py
+import socket
+import gradio as gr
+from openai import OpenAI
+
+def get_local_ip():
+    s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+    try:
+        s.connect(("8.8.8.8", 80))
+        ip = s.getsockname()[0]
+    except Exception:
+        ip = "127.0.0.1"
+    finally:
+        s.close()
+    return ip
+
+print("本機 IP:", get_local_ip())
+
+# ✅ 設定 base URL 連接本地 llama.cpp API
+client = OpenAI(
+    base_url="http://0.0.0.0:8000/v1",
+    api_key="sk-local",  # llama.cpp 不檢查內容，只要有就行
+    timeout=600
+)
+
+# ✅ 回應函式 (流式 generator)
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    # history 是 list of dict: [{"role": "user"/"assistant", "content": "..."}]
+    messages = [{"role": "system", "content": system_message}]
+    messages.extend(history)  # 直接加入舊對話
+    messages.append({"role": "user", "content": message})
+
+    try:
+        stream = client.chat.completions.create(
+            model="qwen3",   # ⚠️ 替換成你 llama.cpp 載入的模型 general.name
+            messages=messages,
+            max_tokens=max_tokens,
+            temperature=temperature,
+            top_p=top_p,
+            stream=True,
+        )
+
+        output = ""
+        for chunk in stream:
+            # 🔍 Debug log
+            # print("[DEBUG] chunk:", chunk)
+
+            if chunk.choices:
+                delta = chunk.choices[0].delta
+                if delta and delta.content:
+                    output += delta.content
+                    yield {"role": "assistant", "content": output}
+
+    except Exception as e:
+        print(f"[Error] {e}")
+        yield {"role": "assistant", "content": "⚠️ Llama.cpp server 沒有回應，請稍後再試。"}
+
+# ✅ Gradio 介面 (新版必須用 type="messages")
+demo = gr.ChatInterface(
+    respond,
+    type="messages",  # 🔑 使用 OpenAI 風格訊息格式
+    additional_inputs=[
+        gr.Textbox(value="You are a friendly assistant.", label="System message"),
+        gr.Slider(minimum=1, maximum=4096, value=1024, 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()