refactor(config): 精简 image_gen 配置为 OpenAI 兼容同步 API

- 移除 yuntts 异步字段: aspect_ratio / x_channel / poll_max_wait / poll_interval
- 改用标准 OpenAI Images API 字段: width / height / quality
- 默认 base_url 改为 api.suchuang.vip/v1
- 模型默认值: gpt-image-2-token（此模型在 suchuang 上需换 gpt-image-1）

backend/internal/config/config.yml

@@ -21,11 +21,9 @@ llm:
   max_tokens: 2048
 
 image_gen:
-  base_url: "https://www.yuntts.com/api/v1"
+  base_url: "https://api.suchuang.vip/v1"
   api_key: ""
-  model: "gpt-image-2"
+  model: "gpt-image-2-token"
+  width: 1024
+  height: 1024
   quality: "low"
-  aspect_ratio: "1:1"
-  x_channel: "default"
-  poll_max_wait: 120
-  poll_interval: 3