@shop-os/modelkit 0.2.9 → 0.2.10

package/dist/index.cjs

@@ -1569,6 +1569,7 @@ var VIDEO_MODELS = [
   },
   {
     // generateAudio defaults to false here (trigger-workflows); Space Builder defaults it to true.
+    // omni_reference_task_type edit/extend: adaptive ratio, edit sends duration -1.
     id: "seedance-2.5",
     aliases: ["seedance-2-5", "dreamina-seedance-2-5-260628"],
     label: "Seedance 2.5",
@@ -1583,7 +1584,7 @@ var VIDEO_MODELS = [
       durations: { min: 4, max: 30 },
       references: { images: 30, videos: 10, audios: 10 },
       ratioIgnoredWith: ["firstFrame", "sourceVideo"],
-      supports: ["generateAudio", "seed", "lastFrame", "watermark"]
+      supports: ["generateAudio", "seed", "lastFrame", "watermark", "taskType"]
     },
     defaults: {
       durationSeconds: 5,
@@ -1692,6 +1693,9 @@ var VIDEO_MODELS = [
   },
   {
     // mode standard/pro selects 720p/1080p; the endpoint takes no separate resolution.
+    // Replicate: "Max 7 [images] without video, 4 with video"; 4K does not support
+    // reference_video. No `videos` limit yet: the adapter sends no reference video,
+    // so `imagesWithVideo` stays dormant until one is wired through.
     id: "kling-v3-omni-video",
     aliases: ["kwaivgi/kling-v3-omni-video", "kling-v3-omni"],
     label: "Kling V3 Omni",
@@ -1704,7 +1708,7 @@ var VIDEO_MODELS = [
       aspectRatios: [...KLING_RATIOS],
       resolutions: ["720p", "1080p"],
       durations: { min: 3, max: 15 },
-      references: { images: 7 },
+      references: { images: 7, imagesWithVideo: 4 },
       supports: ["generateAudio", "lastFrame"]
     },
     defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: false, mode: "pro" },
@@ -1980,6 +1984,7 @@ var SupportedOptionSchema = import_zod.z.enum([
   "upscale",
   "watermark",
   "safetyChecker",
+  "taskType",
   "instrumental"
 ]);
 var UsageUnitSchema = import_zod.z.enum([
@@ -2002,6 +2007,7 @@ var ProviderOptionValueSchema = import_zod.z.lazy(
 var ProviderOptionsSchema = import_zod.z.record(import_zod.z.string(), import_zod.z.record(import_zod.z.string(), ProviderOptionValueSchema));
 var ReferenceLimitsSchema = import_zod.z.looseObject({
   images: import_zod.z.number().optional(),
+  imagesWithVideo: import_zod.z.number().optional(),
   videos: import_zod.z.number().optional(),
   audios: import_zod.z.number().optional(),
   total: import_zod.z.number().optional(),
@@ -2063,6 +2069,7 @@ var DefaultsSchema = import_zod.z.looseObject({
   promptExpansion: import_zod.z.enum(["fast", "balanced", "quality"]).optional(),
   watermark: import_zod.z.boolean().optional(),
   safetyChecker: import_zod.z.boolean().optional(),
+  taskType: import_zod.z.enum(["reference", "edit", "extend"]).optional(),
   mode: import_zod.z.enum(["standard", "pro"]).optional(),
   instrumental: import_zod.z.boolean().optional()
 });
@@ -2622,6 +2629,9 @@ function resolveVideoOptions(model, request, onUnsupported = "error") {
     model,
     warnings
   );
+  if (limits.imagesWithVideo !== void 0 && (limits.videos ?? 0) > 0 && videos.length > 0 && images.length > limits.imagesWithVideo) {
+    images = trim("referenceImages", images, limits.imagesWithVideo, model, warnings);
+  }
   if (limits.total !== void 0) {
     let over = images.length + videos.length + audios.length - limits.total;
     if (over > 0) {
@@ -2655,7 +2665,8 @@ function resolveVideoOptions(model, request, onUnsupported = "error") {
     "promptExpansion",
     "lastFrame",
     "watermark",
-    "safetyChecker"
+    "safetyChecker",
+    "taskType"
   ]);
   return { request: merged, warnings };
 }
@@ -4096,9 +4107,22 @@ function seedanceErrorMessage(error, status) {
   return `Seedance task ${status}`;
 }
 async function generateSeedanceVideo(req, ctx) {
-  const key = arkKey(ctx);
   const modelId = ctx.model.id;
-  const duration = req.durationSeconds ?? 5;
+  const taskType = req.taskType;
+  const hasVideo = Boolean(req.sourceVideo) || (req.referenceVideos?.length ?? 0) > 0;
+  if ((taskType === "edit" || taskType === "extend") && !hasVideo) {
+    throw new ModelKitError({
+      code: "missing_input",
+      message: `${ctx.model.id} ${taskType} task needs a source or reference video`,
+      provider: "ark",
+      model: ctx.model.id,
+      retryable: false
+    });
+  }
+  const key = arkKey(ctx);
+  const requestedDuration = req.durationSeconds ?? 5;
+  let duration = requestedDuration;
+  if (taskType === "edit") duration = -1;
   const resolution = req.resolution ?? "720p";
   const content = [{ type: "text", text: req.prompt }];
   if (req.firstFrame) {
@@ -4150,7 +4174,10 @@ async function generateSeedanceVideo(req, ctx) {
     duration,
     resolution
   };
-  if (!hasFrameAnchor && req.aspectRatio !== void 0) body.ratio = req.aspectRatio;
+  if (taskType === "edit" || taskType === "extend") body.ratio = "adaptive";
+  else if (!hasFrameAnchor && req.aspectRatio !== void 0) body.ratio = req.aspectRatio;
+  if (taskType !== void 0 && ctx.model.id === "seedance-2.5")
+    body.omni_reference_task_type = taskType;
   if (req.generateAudio !== void 0) body.generate_audio = req.generateAudio;
   if (req.seed !== void 0) body.seed = req.seed;
   if (req.watermark !== void 0) body.watermark = req.watermark;
@@ -4170,7 +4197,7 @@ async function generateSeedanceVideo(req, ctx) {
   }, retryOptsFor(ctx));
   if (!submit.id) throwArkError(void 0, "Seedance task create returned no id", modelId);
   const jobId = submit.id;
-  const usage2 = seedanceUsage(ctx, req, duration);
+  const usage2 = seedanceUsage(ctx, req, taskType === "edit" ? requestedDuration : duration);
   const fetchStatus = async () => {
     const res = await arkFetch(ctx, `${arkBaseUrl(ctx)}${SUBMIT_PATH}/${jobId}`, {
       headers: { Authorization: `Bearer ${key}` }
@@ -4217,6 +4244,10 @@ async function generateSeedanceVideo(req, ctx) {
               "Seedance task succeeded but returned no video_url",
               modelId
             );
+          if (polled.duration !== void 0 && polled.duration > 0) {
+            usage2.quantity = polled.duration;
+            usage2.durationSeconds = polled.duration;
+          }
           const tokens = polled.usage;
           if (tokens?.total_tokens !== void 0 && tokens.completion_tokens !== void 0) {
             usage2.inputTokens = tokens.total_tokens - tokens.completion_tokens;
@@ -5270,7 +5301,7 @@ async function generateOmniVideo(req, ctx, createGenAIClient) {
   const apiKey = googleKey(ctx);
   const modelId = ctx.model.id;
   const providerModelId = ctx.model.providerModelId;
-  const duration = req.durationSeconds ?? 8;
+  const duration = req.durationSeconds ?? 10;
   const images = req.referenceImages ?? [];
   const videos = req.referenceVideos ?? [];
   const task = req.sourceVideo ? "edit" : req.firstFrame ? "image_to_video" : images.length > 0 || videos.length > 0 ? "reference_to_video" : "text_to_video";