fix(files): preserve source drafts and image export fidelity

Author: waleedlatif1

apps/sim/app/api/files/export/[id]/route.test.ts

@@ -166,12 +166,26 @@ describe('markdown export bundling', () => {
   it('counts the document body against the export limit, not just its assets', async () => {
     // Assets alone sit under the cap; the body is what carries the bundle over it.
     embeds('a')
-    mockDownloadFile.mockResolvedValue(Buffer.alloc(250 * MB))
+    mockDownloadFile.mockResolvedValue(Buffer.alloc(2 * MB))
+    assetsResolveTo((id) => assetRecord(id, 249 * MB))
 
     const response = await GET(request(), context)
 
     expect(response.status).toBe(400)
     expect((await response.json()).error).toContain('document and its embedded files')
+    expect(mockDownloadFile).toHaveBeenCalledTimes(1)
+  })
+
+  it('downloads large Markdown verbatim without parsing it or querying assets', async () => {
+    const content = Buffer.alloc(11 * MB, 'a')
+    mockDownloadFile.mockResolvedValue(content)
+    const response = await GET(request(), context)
+    expect(response.status).toBe(200)
+    expect(Buffer.from(await response.arrayBuffer()).equals(content)).toBe(true)
+    expect(response.headers.get('content-type')).toBe('text/markdown; charset=utf-8')
+    expect(mockExtractEmbeddedFileRefs).not.toHaveBeenCalled()
+    expect(mockGetFileMetadataById).toHaveBeenCalledTimes(1)
+    expect(mockDownloadFile).toHaveBeenCalledTimes(1)
   })
 
   it('caps the document body read rather than loading it unbounded', async () => {

apps/sim/app/api/files/export/[id]/route.ts

@@ -16,6 +16,7 @@ import { downloadFile } from '@/lib/uploads/core/storage-service'
 import { extractEmbeddedFileRefs } from '@/lib/uploads/server/embedded-image-refs'
 import {
   createMarkdownExport,
+  MAX_EXPORT_MARKDOWN_PARSE_BYTES,
   MAX_EXPORT_TOTAL_BYTES,
   type MarkdownExportAsset,
   type MarkdownExportResult,
@@ -129,11 +130,12 @@ export const GET = withRouteHandler(
         { status: 400 }
       )
     }
-    const mdContent = mdBuffer.toString('utf-8')
-
     // Ids only: a serve-URL embed names a storage key, which the bundler has no id to rewrite the
     // markdown against, so those images stay pointed at their original URL.
-    const { ids: imageIds } = extractEmbeddedFileRefs(mdContent)
+    const imageIds =
+      mdBuffer.length <= MAX_EXPORT_MARKDOWN_PARSE_BYTES
+        ? extractEmbeddedFileRefs(mdBuffer.toString('utf-8')).ids
+        : []
 
     logger.info('Exporting markdown', { id, imageCount: imageIds.length })
 

apps/sim/app/workspace/[workspaceId]/files/components/file-viewer/text-editor-sync.test.tsx

@@ -1,9 +1,10 @@
 /**
  * @vitest-environment jsdom
  */
-import { act, type ComponentProps, Suspense } from 'react'
+import { act, type ComponentProps, createRef, Suspense } from 'react'
 import { createRoot, type Root } from 'react-dom/client'
 import { beforeEach, describe, expect, it, vi } from 'vitest'
+import type { FileDownloadSource } from '@/lib/uploads/client/download'
 import type { WorkspaceFileRecord } from '@/lib/uploads/contexts/workspace'
 import { SIM_PAGE_CONTENT_TYPE } from '@/lib/workspace-files/page-compile'
 import { TextEditor } from '@/app/workspace/[workspaceId]/files/components/file-viewer/text-editor'
@@ -18,6 +19,8 @@ interface MockMonacoProps {
 const state = vi.hoisted(() => ({
   content: 'initial',
   streaming: false,
+  loading: false,
+  error: false,
   editorProps: null as MockMonacoProps | null,
 }))
 
@@ -37,8 +40,8 @@ vi.mock(
         state.content = content
       },
       isStreamInteractionLocked: state.streaming,
-      isContentLoading: false,
-      hasContentError: false,
+      isContentLoading: state.loading,
+      hasContentError: state.error,
       saveImmediately: vi.fn(),
     }),
   })
@@ -161,10 +164,49 @@ describe('TextEditor content synchronization', () => {
     ;(globalThis as { IS_REACT_ACT_ENVIRONMENT?: boolean }).IS_REACT_ACT_ENVIRONMENT = true
     state.content = 'initial'
     state.streaming = false
+    state.loading = false
+    state.error = false
     state.editorProps = null
     useFileViewerStore.getState().reset()
   })
 
+  it('exports the current source draft synchronously without saving and clears it on unmount', () => {
+    const downloadSourceRef = createRef<FileDownloadSource | null>()
+    const root = createRoot(document.createElement('div'))
+    act(() => root.render(<TextEditor {...props} downloadSourceRef={downloadSourceRef} />))
+    expect(downloadSourceRef.current).toMatchObject({
+      fileId: file.id,
+      workspaceId: file.workspaceId,
+    })
+    expect(downloadSourceRef.current?.getContent()).toBe('initial')
+    const draft = '# Source ![image](image.png)\n\n| a | b |\n| - | - |\n'
+    act(() => state.editorProps?.onChange?.(draft))
+    expect(downloadSourceRef.current?.getContent()).toBe(draft)
+    act(() => state.editorProps?.onChange?.(''))
+    expect(downloadSourceRef.current?.getContent()).toBe('')
+    act(() => root.unmount())
+    expect(downloadSourceRef.current).toBeNull()
+  })
+
+  it.each(['loading', 'error'] as const)(
+    'does not expose unavailable source content: %s',
+    (condition) => {
+      state[condition] = true
+      const downloadSourceRef = createRef<FileDownloadSource | null>()
+      const root = createRoot(document.createElement('div'))
+      act(() => root.render(<TextEditor {...props} downloadSourceRef={downloadSourceRef} />))
+      expect(downloadSourceRef.current).toBeNull()
+      state[condition] = false
+      act(() =>
+        root.render(
+          <TextEditor {...props} file={{ ...file }} downloadSourceRef={downloadSourceRef} />
+        )
+      )
+      expect(downloadSourceRef.current?.getContent()).toBe('initial')
+      act(() => root.unmount())
+    }
+  )
+
   it('shares page recognition with already mounted viewers and keeps it across remounts', () => {
     const pageFile = { ...file, id: 'page-a', name: 'page.html', type: 'text/html' }
     const container = document.createElement('div')

apps/sim/app/workspace/[workspaceId]/files/components/file-viewer/text-editor.tsx

@@ -7,6 +7,7 @@ import {
   useCallback,
   useEffect,
   useId,
+  useImperativeHandle,
   useLayoutEffect,
   useMemo,
   useRef,
@@ -21,6 +22,7 @@ import {
   buildFileSelectionLabel,
   truncateSelectionText,
 } from '@/lib/copilot/chat/selection-context'
+import type { FileDownloadSource } from '@/lib/uploads/client/download'
 import type { WorkspaceFileRecord } from '@/lib/uploads/contexts/workspace'
 import { getFileExtension } from '@/lib/uploads/utils/file-utils'
 import { isSimPageSource, SIM_PAGE_CONTENT_TYPE } from '@/lib/workspace-files/page-compile'
@@ -403,6 +405,7 @@ interface TextEditorProps {
     retry?: () => Promise<void>
   ) => void
   saveRef?: React.MutableRefObject<(() => Promise<void>) | null>
+  downloadSourceRef?: React.MutableRefObject<FileDownloadSource | null>
   discardRef?: React.MutableRefObject<(() => void) | null>
   streamingContent?: string
   isAgentEditing?: boolean
@@ -419,6 +422,7 @@ export const TextEditor = memo(function TextEditor({
   onDirtyChange,
   onSaveStatusChange,
   saveRef,
+  downloadSourceRef,
   discardRef,
   streamingContent,
   isAgentEditing,
@@ -504,6 +508,19 @@ export const TextEditor = memo(function TextEditor({
     contentRef.current = content
   }, [content])
 
+  useImperativeHandle<FileDownloadSource | null, FileDownloadSource | null>(
+    downloadSourceRef,
+    () =>
+      isContentLoading || hasContentError
+        ? null
+        : {
+            fileId: file.id,
+            workspaceId,
+            getContent: () => contentRef.current,
+          },
+    [file.id, workspaceId, isContentLoading, hasContentError]
+  )
+
   // Enable once content has loaded — the container (and Monaco) only mount after
   // the `isContentLoading` early return below, so the bridge must (re-)attach then.
   useSelectionCopyBridge(containerRef, buildSelectionContext, workspaceId, !isContentLoading)

apps/sim/lib/uploads/client/download.test.ts

@@ -1,4 +1,6 @@
 /** @vitest-environment jsdom */
+
+import { PASTE_LIMITS } from '@sim/utils/paste'
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest'
 import { requestRaw } from '@/lib/api/client/request'
 import { exportWorkspaceFileSnapshotContract } from '@/lib/api/contracts/workspace-files'
@@ -20,7 +22,7 @@ const file: WorkspaceFileRecord = {
 }
 
 const fetchMock = vi.fn<typeof fetch>()
-const createObjectURL = vi.fn(() => 'blob:download')
+const createObjectURL = vi.fn((_blob: Blob) => 'blob:download')
 const click = vi.fn()
 let downloadedName = ''
 
@@ -56,6 +58,22 @@ function source(content = 'latest visible content'): FileDownloadSource {
 }
 
 describe('file download snapshots', () => {
+  it('downloads an oversized source draft directly instead of rejecting it or using stale storage', async () => {
+    const content = '😀'.repeat(Math.floor(PASTE_LIMITS.RICH_MARKDOWN_BYTES / 4) + 1)
+    await triggerFileDownload(file, source(content))
+    expect(requestRaw).not.toHaveBeenCalled()
+    expect(fetchMock).not.toHaveBeenCalled()
+    const blob = createObjectURL.mock.calls[0]![0]
+    const reader = new FileReader()
+    const read = new Promise<string>((resolve) => {
+      reader.onload = () => resolve(reader.result as string)
+    })
+    reader.readAsText(blob)
+    await vi.runAllTimersAsync()
+    expect(await read).toBe(content)
+    expect(downloadedName).toBe(file.name)
+    expect(click).toHaveBeenCalledOnce()
+  })
   it.each(['latest local and peer text', ''])(
     'captures the mounted content immediately: %j',
     async (content) => {

apps/sim/lib/uploads/client/download.ts

@@ -1,3 +1,4 @@
+import { PASTE_LIMITS, utf8ByteLength } from '@sim/utils/paste'
 import { requestRaw } from '@/lib/api/client/request'
 import { downloadWorkspaceFileItemsContract } from '@/lib/api/contracts/workspace-file-folders'
 import { exportWorkspaceFileSnapshotContract } from '@/lib/api/contracts/workspace-files'
@@ -56,6 +57,13 @@ export async function triggerFileDownload(
       : null
 
   if (content !== null) {
+    /** Source editing accepts larger drafts than the bounded image-bundling endpoint. */
+    if (
+      utf8ByteLength(content, PASTE_LIMITS.RICH_MARKDOWN_BYTES) > PASTE_LIMITS.RICH_MARKDOWN_BYTES
+    ) {
+      saveBlob(new Blob([content], { type: 'text/markdown; charset=utf-8' }), record.name)
+      return
+    }
     const response = await requestRaw(
       exportWorkspaceFileSnapshotContract,
       {