fix(uploads): align register response schema with UserFile; skip presigned for KB large files

- registerWorkspaceFileResponseSchema now matches the UserFile shape the route actually returns; previous schema required workspace DB-row fields that were never populated, causing requestJson validation to reject successful uploads.
- KB batch presigned fetch now skips files >= LARGE_FILE_THRESHOLD since multipart bypasses the per-file presigned URL anyway.

Author: waleedlatif1

apps/sim/app/workspace/[workspaceId]/knowledge/hooks/use-knowledge-upload.ts

@@ -4,6 +4,7 @@ import { sleep } from '@sim/utils/helpers'
 import { useQueryClient } from '@tanstack/react-query'
 import {
   DirectUploadError,
+  LARGE_FILE_THRESHOLD,
   MULTIPART_MAX_RETRIES,
   MULTIPART_RETRY_BACKOFF,
   MULTIPART_RETRY_DELAY_MS,
@@ -98,43 +99,49 @@ interface BatchPresignedFile {
 }
 
 /**
- * Fetch presigned upload data for many files in one round trip.
- * Returns one PresignedUploadInfo per input file (in order).
+ * Fetch presigned upload data for the small files in `files`. Returns a sparse
+ * array aligned with the input: entries for files >= LARGE_FILE_THRESHOLD are
+ * `undefined` because those uploads use multipart and never consume a presigned
+ * single-PUT URL.
  */
-const fetchBatchPresignedData = async (files: File[]): Promise<PresignedUploadInfo[]> => {
-  const batches: File[][] = []
-  for (let start = 0; start < files.length; start += BATCH_REQUEST_SIZE) {
-    batches.push(files.slice(start, start + BATCH_REQUEST_SIZE))
+const fetchBatchPresignedData = async (
+  files: File[]
+): Promise<(PresignedUploadInfo | undefined)[]> => {
+  const result: (PresignedUploadInfo | undefined)[] = new Array(files.length).fill(undefined)
+  const smallFileIndices: number[] = []
+  for (let i = 0; i < files.length; i++) {
+    if (files[i].size < LARGE_FILE_THRESHOLD) smallFileIndices.push(i)
   }
+  if (smallFileIndices.length === 0) return result
+
+  for (let start = 0; start < smallFileIndices.length; start += BATCH_REQUEST_SIZE) {
+    const batchIndices = smallFileIndices.slice(start, start + BATCH_REQUEST_SIZE)
+    const batchFiles = batchIndices.map((i) => files[i])
+    const body: { files: BatchPresignedFile[] } = {
+      files: batchFiles.map((file) => ({
+        fileName: file.name,
+        contentType: getFileContentType(file),
+        fileSize: file.size,
+      })),
+    }
 
-  const batchResults = await Promise.all(
-    batches.map(async (batch, batchIndex) => {
-      const body: { files: BatchPresignedFile[] } = {
-        files: batch.map((file) => ({
-          fileName: file.name,
-          contentType: getFileContentType(file),
-          fileSize: file.size,
-        })),
-      }
-
-      const response = await fetch(KB_BATCH_PRESIGNED_ENDPOINT, {
-        method: 'POST',
-        headers: { 'Content-Type': 'application/json' },
-        body: JSON.stringify(body),
-      })
+    const response = await fetch(KB_BATCH_PRESIGNED_ENDPOINT, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify(body),
+    })
 
-      if (!response.ok) {
-        throw new Error(
-          `Batch ${batchIndex + 1} presigned URL generation failed: ${response.statusText}`
-        )
-      }
+    if (!response.ok) {
+      throw new Error(`Batch presigned URL generation failed: ${response.statusText}`)
+    }
 
-      const { files: presignedItems } = (await response.json()) as { files: unknown[] }
-      return batch.map((file, idx) => normalizePresignedData(presignedItems[idx], file.name))
+    const { files: presignedItems } = (await response.json()) as { files: unknown[] }
+    batchIndices.forEach((fileIdx, batchPos) => {
+      result[fileIdx] = normalizePresignedData(presignedItems[batchPos], batchFiles[batchPos].name)
     })
-  )
+  }
 
-  return batchResults.flat()
+  return result
 }
 
 /**
@@ -232,7 +239,7 @@ export function useKnowledgeUpload(options: UseKnowledgeUploadOptions = {}) {
   const uploadOneFile = async (
     file: File,
     fileIndex: number,
-    presigned: PresignedUploadInfo
+    presigned: PresignedUploadInfo | undefined
   ): Promise<UploadedFile> => {
     if (!options.workspaceId) {
       throw new KnowledgeUploadError('workspaceId is required for upload', 'MISSING_WORKSPACE_ID')

apps/sim/hooks/queries/workspace-files.ts

@@ -18,6 +18,7 @@ import {
   type UploadProgressEvent,
 } from '@/lib/uploads/client/direct-upload'
 import type { WorkspaceFileRecord } from '@/lib/uploads/contexts/workspace'
+import type { UserFile } from '@/executor/types'
 
 const logger = createLogger('WorkspaceFilesQuery')
 
@@ -207,7 +208,7 @@ interface UploadFileParams {
 
 interface UploadFileResponse {
   success: boolean
-  file: WorkspaceFileRecord
+  file: UserFile
 }
 
 async function uploadViaApiFallback(
@@ -232,7 +233,7 @@ async function parseUploadResponse(
   response: Response,
   fallbackMessage: string
 ): Promise<UploadFileResponse> {
-  let data: { success?: boolean; error?: string; file?: WorkspaceFileRecord } | null = null
+  let data: { success?: boolean; error?: string; file?: UserFile } | null = null
   try {
     data = await response.json()
   } catch {}

apps/sim/lib/api/contracts/workspace-files.ts

@@ -188,9 +188,19 @@ export const registerWorkspaceFileBodySchema = z.object({
 
 export type RegisterWorkspaceFileBody = z.input<typeof registerWorkspaceFileBodySchema>
 
+const registeredWorkspaceFileSchema = z.object({
+  id: z.string(),
+  name: z.string(),
+  url: z.string(),
+  size: z.number(),
+  type: z.string(),
+  key: z.string(),
+  context: z.string().optional(),
+})
+
 const registerWorkspaceFileResponseSchema = z.object({
   success: z.boolean(),
-  file: workspaceFileRecordSchema.optional(),
+  file: registeredWorkspaceFileSchema.optional(),
   error: z.string().optional(),
   isDuplicate: z.boolean().optional(),
 })

apps/sim/lib/uploads/client/direct-upload.ts

@@ -5,7 +5,7 @@ import { getFileContentType, isAbortError } from '@/lib/uploads/utils/file-utils
 const logger = createLogger('DirectUpload')
 
 const CHUNK_SIZE = 8 * 1024 * 1024
-const LARGE_FILE_THRESHOLD = 50 * 1024 * 1024
+export const LARGE_FILE_THRESHOLD = 50 * 1024 * 1024
 const BASE_TIMEOUT_MS = 2 * 60 * 1000
 const TIMEOUT_PER_MB_MS = 1500
 const MAX_TIMEOUT_MS = 10 * 60 * 1000