优化前缀缓存机制和UI，并引入一些新工具.preview

llm_server/start_image_gen.sh

@@ -4,6 +4,9 @@ set -euo pipefail
 # ---------- 配置 ----------
 HOST_PORT="${HOST_PORT:-8904}"
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# 鉴权与模型名校验：SERVER_API_KEY / SERVER_MODELNAME 为空时服务端不校验（任意值放行）
+SERVER_API_KEY="${SERVER_API_KEY:-}"
+SERVER_MODELNAME="${SERVER_MODELNAME:-Z-Image-Turbo}"
 
 # ---------- Python 虚拟环境 ----------
 if [ -f "${SCRIPT_DIR}/../myenv/bin/activate" ]; then
@@ -20,5 +23,9 @@ if [ -n "${CUDA_DEV}" ]; then
 fi
 
 echo "[Z-Image] 启动图像生成服务 (端口=${HOST_PORT})"
+echo "[Z-Image] 服务模型名: ${SERVER_MODELNAME}"
+if [ -n "${SERVER_API_KEY}" ]; then
+  echo "[Z-Image] API Key 校验已启用"
+fi
 cd "${SCRIPT_DIR}"
-exec python image_gen_server.py --port "${HOST_PORT}"
+exec python image_gen_server.py --port "${HOST_PORT}" --model-name "${SERVER_MODELNAME}" --api-key "${SERVER_API_KEY}"