app.py

### 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):
    messages = [{"role": "system", "content": system_message}]
    messages.extend(history)
    messages.append({"role": "user", "content": message})

    try:
        stream = client.chat.completions.create(
            model="qwen3",
            messages=messages,
            max_tokens=max_tokens,
            temperature=temperature,
            top_p=top_p,
            reasoning_effort="low",
            stream=True,
        )

        output = ""
        skip_tokens = ['<|channel|>', 'analysis']
        skipped = {token: False for token in skip_tokens}  # 追蹤每個 token 是否已忽略過

        for chunk in stream:
            print("[DEBUG] chunk:", chunk)
            delta = chunk.choices[0].delta
            if delta and delta.content:
                content = delta.content.strip()
                # 如果這個 token 是要跳過的，且還沒被跳過過
                if content in skip_tokens and not skipped[content]:
                    skipped[content] = True
                    continue
                output += delta.content  # 正常加入輸出

            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=16383, value=4096, 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()