feat(ai)：补充说明下，wan2.7-image 要注意，等 alibaba ai 兼容！！！

YunaiV
1 parent 814a9633
Showing 2 changed files with 17 additions and 5 deletions
yudao-module-ai/src/main/java/cn/iocoder/yudao/module/ai/service/image/AiImageServiceImpl.java
yudao-module-ai/src/test/java/cn/iocoder/yudao/module/ai/framework/ai/core/model/image/TongYiImagesModelTest.java
@@ -109,6 +109,7 @@ public class AiImageServiceImpl implements AiImageService {
     }
     @Async
+    @SuppressWarnings("ConstantValue")
     public void executeDrawImage(AiImageDO image, AiImageDrawReqVO reqVO, AiModelDO model) {
         try {
             // 1.1 构建请求
@@ -164,8 +165,8 @@ public class AiImageServiceImpl implements AiImageService {
                     .build();
         } else if (ObjUtil.equal(model.getPlatform(), AiPlatformEnum.TONG_YI.getPlatform())) {
             return DashScopeImageOptions.builder()
-                    .withModel(model.getModel()).withN(1)
-                    .withHeight(draw.getHeight()).withWidth(draw.getWidth())
+                    .model(model.getModel()).n(1)
+                    .height(draw.getHeight()).width(draw.getWidth())
                     .build();
         } else if (ObjUtil.equal(model.getPlatform(), AiPlatformEnum.YI_YAN.getPlatform())) {
             return QianFanImageOptions.builder()
@@ -12,23 +12,34 @@ import org.springframework.ai.image.ImageResponse;
 /**
  * {@link DashScopeImageModel} 集成测试类
  *
+ * TODO @芋艿：注：spring-ai-alibaba-dashscope（1.1.2.2）的 {@code DashScopeImageApi#resolveImagePath} 未给 {@code wan2.7-image} 加路由分支，
+ * 会落到默认的 {@code text2image/image-synthesis} 异步端点 + 旧版 {@code prompt} 入参，
+ * 而该模型实际要求 {@code multimodal-generation/generation} 同步端点 + {@code messages} 入参，
+ * 端点、异步头、入参结构全部对不上，所以走 SDK 直接调用必失败。
+ *
+ * 临时方案：改用 SDK 已支持的 {@code wan2.6-image}（异步）或 {@code qwen-image}（同步）；
+ * 或在项目内同包同名覆盖 {@code DashScopeImageApi}，把 {@code wan2.7*} 也路由到 {@code wan2.6-image} 那条 {@code image-generation/generation} 异步分支。
+ *
  * @author fansili
  */
 public class TongYiImagesModelTest {
     private final DashScopeImageModel imageModel = DashScopeImageModel.builder()
             .dashScopeApi(DashScopeImageApi.builder()
-                    .apiKey("sk-47aa124781be4bfb95244cc62f63f7d0")
+                    .apiKey("sk-cd9f39e99ea54840bd1888282325f55a") // https://bailian.console.aliyun.com/cn-beijing/?tab=model#/api-key 获取密钥
                     .build())
             .build();
+    // TODO @芋艿：
     @Test
     @Disabled
     public void imageCallTest() {
         // 准备参数
         ImageOptions options = DashScopeImageOptions.builder()
-                .withModel("wanx-v1")
-                .withHeight(256).withWidth(256)
+                .model("wan2.7-image")
+//                .withSize("2k")
+                .height(768).width(768)
+                .n(1)
                 .build();
         ImagePrompt prompt = new ImagePrompt("中国长城!", options);