Add image-to-text Describe on the bench via exclusive llama.cpp caption jobs.

server/api/generate/caption.post.ts

@@ -0,0 +1,50 @@
+import { validateCaptionBody } from '~/shared/caption.mjs'
+import { addStudioJob, kickStudioQueue, type StudioJobPayload } from '~/server/utils/studioQueue'
+import { getStill } from '~/server/utils/library'
+
+export default defineEventHandler(async (event) => {
+  const { owner } = assertLibraryOwner(event)
+  const raw = await readBody(event)
+  let body: ReturnType<typeof validateCaptionBody>
+  try {
+    body = validateCaptionBody(raw)
+  } catch (e: any) {
+    throw createError({ statusCode: e.statusCode || 400, statusMessage: e.message })
+  }
+  assertFolderAccess(event, body.folderId)
+  const still = getStill(owner, body.stillId)
+  assertFolderAccess(event, still.folderId)
+
+  const row = await addStudioJob({
+    ownerKey: owner,
+    kind: 'caption',
+    familyId: crypto.randomUUID(),
+    payload: {
+      prompt: `Describe · ${body.captionStyle}`,
+      name: `Describe · ${body.captionStyle}`,
+      folderId: body.folderId,
+      stillId: still.id,
+      captionStyle: body.captionStyle,
+      aspect: 'auto',
+      width: 0,
+      height: 0,
+      steps: 1,
+      turbo: true,
+      seed: 0,
+      cfg: 0,
+      fps: 24,
+      samplerName: 'euler',
+      scheduler: 'simple',
+      duration: 0,
+      sound: false,
+      workflow: 'minimax',
+      useIdentityRefs: false,
+      hideThumbnail: false,
+      referenceStillIds: [],
+      extensions: [],
+      queueAutoRun: true
+    } as StudioJobPayload
+  })
+  kickStudioQueue()
+  return { id: row.id, liveJobId: row.liveJobId, captionStyle: body.captionStyle, stillId: still.id }
+})