change model

prompt/llm.py

@@ -13,6 +13,9 @@ SYSTEM_PROMPT = """你是一个网页生成助手。根据用户的需求描述
 5. 在已有代码基础上修改时，返回完整的修改后代码，不要只返回片段
 6. 由于任何外部链接都被屏蔽，使用纯 HTML、CSS 和 JS 实现功能，不要依赖外部库"""
 
+# Models served by the ARK (Volcengine) endpoint
+ARK_MODELS = {"doubao-seed-2-0-mini-260215"}
+
 
 def build_messages(history: list[dict]) -> list[dict]:
     """Build the message list for the LLM API call."""
@@ -21,16 +24,35 @@ def build_messages(history: list[dict]) -> list[dict]:
     return messages
 
 
-async def stream_chat(history: list[dict]):
+def _get_client(model: str) -> tuple[AsyncOpenAI, str]:
+    """Return (client, model_id) for the given model name."""
+    if model in ARK_MODELS:
+        return (
+            AsyncOpenAI(
+                api_key=settings.ARK_API_KEY,
+                base_url=settings.ARK_BASE_URL,
+                timeout=120.0,
+            ),
+            model,
+        )
+    resolved_model = model or settings.LLM_MODEL
+    return (
+        AsyncOpenAI(
+            api_key=settings.LLM_API_KEY,
+            base_url=settings.LLM_BASE_URL,
+            timeout=120.0,
+        ),
+        resolved_model,
+    )
+
+
+async def stream_chat(history: list[dict], model: str = ""):
     """Stream chat completion from the LLM. Yields content chunks."""
     messages = build_messages(history)
-    async with AsyncOpenAI(
-        api_key=settings.LLM_API_KEY,
-        base_url=settings.LLM_BASE_URL,
-        timeout=120.0,
-    ) as client:
-        stream = await client.chat.completions.create(
-            model=settings.LLM_MODEL,
+    client, resolved_model = _get_client(model or settings.LLM_MODEL)
+    async with client as c:
+        stream = await c.chat.completions.create(
+            model=resolved_model,
             messages=messages,
             stream=True,
         )