Update Dockerfile

Dockerfile

@@ -2,60 +2,32 @@ FROM python:3.11-slim
 
 WORKDIR /app
 
-# 安装必要库
+# 安装依赖
 RUN pip install fastapi uvicorn google-genai
 
-# 生成 main.py，包含动态模型列表 + Chat + Embeddings
+# 生成 main.py（支持自动获取模型列表）
 RUN cat > main.py << 'PYEOF'
 from fastapi import FastAPI, Request
-from fastapi.responses import JSONResponse, StreamingResponse
+from fastapi.responses import JSONResponse
 from google import genai
-import os, uvicorn, json, time
+import os, uvicorn
 
 app = FastAPI()
 GOOGLE_API_KEY = os.environ["GOOGLE_API_KEY"]
 client = genai.Client(api_key=GOOGLE_API_KEY)
 
-# ---------- 模型名称映射（短名 -> 完整 resources 名）----------
-MODEL_MAP = {
-    # 聊天模型
-    "gemini-2.5-flash": "models/gemini-2.5-flash",
-    "gemini-2.5-pro": "models/gemini-2.5-pro",
-    "gemini-1.5-flash": "models/gemini-1.5-flash",
-    "gemini-1.5-pro": "models/gemini-1.5-pro",
-    # 嵌入模型
-    "gemini-embedding-001": "models/gemini-embedding-001",
-    "gemini-embedding-2-preview": "models/gemini-embedding-2-preview",
-}
-
 # 静态模型列表：作为 API 调用失败时的后备方案
 STATIC_MODELS = [
-    {"id": "gemini-2.5-flash", "object": "model", "owned_by": "gemini"},
-    {"id": "gemini-2.5-pro", "object": "model", "owned_by": "gemini"},
-    {"id": "gemini-1.5-flash", "object": "model", "owned_by": "gemini"},
-    {"id": "gemini-1.5-pro", "object": "model", "owned_by": "gemini"},
     {"id": "gemini-embedding-001", "object": "model", "owned_by": "gemini"},
     {"id": "gemini-embedding-2-preview", "object": "model", "owned_by": "gemini"},
 ]
 
-# ---------- 辅助函数 ----------
-def get_full_model_name(short_name: str) -> str:
-    """将短模型名转换为完整的 resources 路径，若未找到则自动添加 models/ 前缀"""
-    return MODEL_MAP.get(short_name, f"models/{short_name}")
-
-def convert_openai_messages(messages: list) -> list:
-    """将 OpenAI 格式的 messages 转换为 Gemini 要求的 contents 格式"""
-    contents = []
-    for msg in messages:
-        role = msg.get("role", "user")
-        text = msg.get("content", "")
-        # Gemini 使用 "user" 和 "model" 角色，将 "assistant" 映射为 "model"
-        if role == "assistant":
-            role = "model"
-        elif role not in ("user", "model"):
-            role = "user"  # 其他角色一律视为 user
-        contents.append({"role": role, "parts": [{"text": text}]})
-    return contents
+# 模型名称映射（短名称 -> 完整路径），用于 /v1/embeddings 端点
+MODEL_MAP = {
+    "text-embedding-004": "models/text-embedding-004",
+    "gemini-embedding-001": "models/gemini-embedding-001",
+    "gemini-embedding-2-preview": "models/gemini-embedding-2-preview",
+}
 
 def fetch_embedding_models():
     """动态获取支持 embedContent 的 Gemini Embedding 模型列表"""
@@ -76,77 +48,19 @@ def fetch_embedding_models():
         print(f"动态获取模型列表失败，回退到静态列表: {e}")
         return STATIC_MODELS
 
-# ---------- API 端点 ----------
 @app.get("/v1/models")
 async def list_models():
-    """
-    返回可用模型列表。
-    优先动态获取嵌入模型，如果失败则返回静态列表。
-    """
     models = fetch_embedding_models()
-    # 确保基础聊天模型也出现在列表中（如果动态列表只返回嵌入模型）
-    chat_models = [m for m in STATIC_MODELS if "embedding" not in m["id"]]
-    for cm in chat_models:
-        if not any(m["id"] == cm["id"] for m in models):
-            models.append(cm)
     return JSONResponse({"object": "list", "data": models})
 
-@app.post("/v1/chat/completions")
-async def chat_completions(request: Request):
-    """OpenAI 兼容的聊天补全端点（非流式）"""
-    try:
-        body = await request.json()
-        model_short = body.get("model", "gemini-2.5-flash")
-        messages = body.get("messages", [])
-        temperature = body.get("temperature", 0.7)
-        max_tokens = body.get("max_tokens", 1000)
-        
-        model_full = get_full_model_name(model_short)
-        contents = convert_openai_messages(messages)
-        
-        # 调用 Gemini API
-        response = client.models.generate_content(
-            model=model_full,
-            contents=contents,
-            config={
-                "temperature": temperature,
-                "max_output_tokens": max_tokens,
-            }
-        )
-        
-        # 构造 OpenAI 兼容响应
-        choice = {
-            "index": 0,
-            "message": {
-                "role": "assistant",
-                "content": response.text
-            },
-            "finish_reason": "stop"
-        }
-        return JSONResponse({
-            "id": f"chatcmpl-{int(time.time())}",
-            "object": "chat.completion",
-            "created": int(time.time()),
-            "model": model_short,
-            "choices": [choice],
-            "usage": {
-                "prompt_tokens": response.usage_metadata.prompt_token_count,
-                "completion_tokens": response.usage_metadata.candidates_token_count,
-                "total_tokens": response.usage_metadata.total_token_count
-            }
-        })
-    except Exception as e:
-        return JSONResponse({"error": str(e)}, status_code=500)
-
 @app.post("/v1/embeddings")
 async def create_embeddings(request: Request):
-    """OpenAI 兼容的嵌入端点"""
     try:
         body = await request.json()
         model_short = body.get("model", "gemini-embedding-001")
+        # 自动补全模型名称为完整路径
+        model_full = MODEL_MAP.get(model_short, f"models/{model_short}")
         input_text = body.get("input", "")
-        model_full = get_full_model_name(model_short)
-        
         result = client.models.embed_content(
             model=model_full,
             contents=input_text