Create app.py

app.py

@@ -0,0 +1,181 @@
+import gradio as gr
+import os
+import json
+import httpx
+
+# ============================================================
+# API 配置
+# ============================================================
+STEPFUN_API_KEY = os.environ.get("STEPFUN_API_KEY", "")
+STEPFUN_BASE_URL = "https://api.stepfun.com/v1"
+MODEL_NAME = "step-3.5-flash"
+HF_CONFIG_URL = "https://huggingface.co/stepfun-ai/Step-3.5-Flash/raw/main/config.json"
+STEPFUN_LOGO = "https://huggingface.co/stepfun-ai/Step-3.5-Flash/resolve/main/stepfun.svg"
+
+cached_config = None
+
+
+def fetch_model_config():
+    global cached_config
+    try:
+        response = httpx.get(HF_CONFIG_URL, timeout=10.0)
+        if response.status_code == 200:
+            cached_config = response.json()
+            return cached_config
+    except Exception as e:
+        print(f"拉取 config.json 失败: {e}")
+    return cached_config
+
+
+def format_messages(history, system_prompt: str, user_message: str):
+    """将 chatbot history 转换为 API 消息格式"""
+    messages = []
+    if system_prompt.strip():
+        messages.append({"role": "system", "content": system_prompt})
+    for user_msg, bot_msg in history:
+        if user_msg:
+            messages.append({"role": "user", "content": user_msg})
+        if bot_msg:
+            messages.append({"role": "assistant", "content": bot_msg})
+    messages.append({"role": "user", "content": user_message})
+    return messages
+
+
+def chat_stream(message: str, history, system_prompt: str, max_tokens: int, temperature: float, top_p: float):
+    """流式聊天，返回 (reasoning, content) 生成器"""
+    fetch_model_config()
+    messages = format_messages(history, system_prompt, message)
+
+    reasoning = ""
+    content = ""
+
+    try:
+        headers = {
+            "Authorization": f"Bearer {STEPFUN_API_KEY}",
+            "Content-Type": "application/json",
+        }
+        payload = {
+            "model": MODEL_NAME,
+            "messages": messages,
+            "stream": True,
+            "max_tokens": max_tokens,
+            "temperature": temperature if temperature > 0 else 0.01,
+            "top_p": top_p,
+        }
+
+        with httpx.stream("POST", f"{STEPFUN_BASE_URL}/chat/completions", headers=headers, json=payload, timeout=120.0) as response:
+            response.raise_for_status()
+            for line in response.iter_lines():
+                if not line or not line.startswith("data: "):
+                    continue
+                data_str = line[6:]
+                if data_str == "[DONE]":
+                    break
+                try:
+                    chunk = json.loads(data_str)
+                    delta = chunk.get("choices", [{}])[0].get("delta", {})
+                    if delta.get("reasoning"):
+                        reasoning += delta["reasoning"]
+                        yield reasoning, content
+                    if delta.get("content"):
+                        content += delta["content"]
+                        yield reasoning, content
+                except json.JSONDecodeError:
+                    continue
+        yield reasoning, content
+
+    except httpx.HTTPStatusError as e:
+        yield reasoning, f"❌ API 错误: {e.response.status_code}"
+    except Exception as e:
+        yield reasoning, f"❌ 错误: {str(e)}"
+
+
+def create_demo():
+    examples = [
+        "请解释一下什么是机器学习？",
+        "帮我写一个 Python 快速排序算法",
+        "1000以内有多少个质数？",
+        "一个农夫需要把狼、羊和白菜都带过河，请问农夫该怎么办？",
+    ]
+
+    with gr.Blocks(title="Step-3.5-Flash", theme=gr.themes.Soft()) as demo:
+        gr.Markdown("# 🚀 Step-3.5-Flash")
+
+        with gr.Row():
+            # 左侧：思考过程 (1)
+            with gr.Column(scale=1, min_width=250):
+                gr.Markdown("### 💭 思考过程")
+                thinking_display = gr.Textbox(
+                    value="等待输入...",
+                    lines=20,
+                    max_lines=20,
+                    interactive=False,
+                    show_label=False,
+                )
+
+            # 右侧：对话 (4)
+            with gr.Column(scale=4):
+                gr.Markdown("### 💬 对话")
+                chatbot = gr.Chatbot(
+                    height=450,
+                    show_label=False,
+                    avatar_images=(None, STEPFUN_LOGO),
+                )
+
+                with gr.Row():
+                    msg = gr.Textbox(
+                        placeholder="输入消息...",
+                        show_label=False,
+                        scale=8,
+                        container=False,
+                    )
+                    submit_btn = gr.Button("发送", variant="primary", scale=1)
+                    clear_btn = gr.Button("🗑️", scale=0, min_width=50)
+
+        # 设置（折叠）
+        with gr.Accordion("⚙️ 设置", open=False):
+            with gr.Row():
+                system_prompt = gr.Textbox(label="系统提示词", value="你是一个有帮助的 AI 助手。", scale=2)
+                max_tokens = gr.Slider(256, 131072, value=4096, step=256, label="最大长度", scale=1)
+                temperature = gr.Slider(0.0, 1.5, value=0.7, step=0.1, label="Temperature", scale=1)
+                top_p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p", scale=1)
+
+        gr.Examples(examples, inputs=msg, label="💡 试试这些")
+
+        # 事件处理
+        def respond(message, history, system_prompt, max_tokens, temperature, top_p):
+            if not message.strip():
+                yield history, "", ""
+                return
+
+            # 添加用户消息
+            history = history + [[message, None]]
+            yield history, "", "思考中..."
+
+            reasoning = ""
+            content = ""
+
+            for r, c in chat_stream(message, history[:-1], system_prompt, max_tokens, temperature, top_p):
+                reasoning = r if r else ""
+                content = c if c else "▌"
+                history[-1][1] = content
+                yield history, "", reasoning
+
+            history[-1][1] = content
+            yield history, "", reasoning
+
+        def on_clear():
+            return [], "", "等待输入..."
+
+        msg.submit(respond, [msg, chatbot, system_prompt, max_tokens, temperature, top_p], [chatbot, msg, thinking_display])
+        submit_btn.click(respond, [msg, chatbot, system_prompt, max_tokens, temperature, top_p], [chatbot, msg, thinking_display])
+        clear_btn.click(on_clear, outputs=[chatbot, msg, thinking_display])
+        demo.load(fetch_model_config)
+
+    return demo
+
+
+if __name__ == "__main__":
+    demo = create_demo()
+    demo.queue()
+    demo.launch(server_name="0.0.0.0", server_port=7860)