address comments

Author: icecrasher321

apps/sim/app/api/files/parse/route.test.ts

@@ -69,6 +69,7 @@ vi.mock('@/app/api/files/authorization', () => ({
 vi.mock('@/lib/uploads', () => ({
   getStorageProvider: mockGetStorageProvider,
   isUsingCloudStorage: mockIsUsingCloudStorage,
+  StorageService: storageServiceMock,
 }))
 
 vi.mock('@/lib/file-parsers', () => ({
@@ -172,6 +173,7 @@ describe('File Parse API Route', () => {
 
     permissionsMockFns.mockGetUserEntityPermissions.mockResolvedValue({ canView: true })
     storageServiceMockFns.mockHasCloudStorage.mockReturnValue(true)
+    storageServiceMockFns.mockDownloadFile.mockResolvedValue(Buffer.from('test file content'))
     mockIsSupportedFileType.mockReturnValue(true)
     mockParseFile.mockResolvedValue({
       content: 'parsed content',
@@ -245,6 +247,48 @@ describe('File Parse API Route', () => {
     }
   })
 
+  it('should keep known binary extensions as binary even when the bytes are valid UTF-8', async () => {
+    setupFileApiMocks({
+      cloudEnabled: true,
+      storageProvider: 's3',
+      authenticated: true,
+    })
+    mockIsSupportedFileType.mockReturnValue(false)
+    storageServiceMockFns.mockDownloadFile.mockResolvedValue(Buffer.from('valid utf8 bytes'))
+
+    const req = createMockRequest('POST', {
+      filePath: '/api/files/serve/execution/workspace-1/workflow-1/execution-1/image.png',
+    })
+
+    const response = await POST(req)
+    const data = await response.json()
+
+    expect(response.status).toBe(200)
+    expect(data.success).toBe(true)
+    expect(data.output.content).toBe('[Binary PNG file - 16 bytes]')
+  })
+
+  it('should parse unknown extensions as text when the bytes look like UTF-8 text', async () => {
+    setupFileApiMocks({
+      cloudEnabled: true,
+      storageProvider: 's3',
+      authenticated: true,
+    })
+    mockIsSupportedFileType.mockReturnValue(false)
+    storageServiceMockFns.mockDownloadFile.mockResolvedValue(Buffer.from('plain text content'))
+
+    const req = createMockRequest('POST', {
+      filePath: '/api/files/serve/execution/workspace-1/workflow-1/execution-1/readme.customtext',
+    })
+
+    const response = await POST(req)
+    const data = await response.json()
+
+    expect(response.status).toBe(200)
+    expect(data.success).toBe(true)
+    expect(data.output.content).toBe('plain text content')
+  })
+
   it('should handle multiple files', async () => {
     setupFileApiMocks({
       cloudEnabled: false,

apps/sim/app/api/files/parse/route.ts

@@ -3,6 +3,7 @@ import { createHash } from 'crypto'
 import fsPromises, { readFile } from 'fs/promises'
 import path from 'path'
 import { createLogger } from '@sim/logger'
+import binaryExtensionsList from 'binary-extensions'
 import { type NextRequest, NextResponse } from 'next/server'
 import { fileParseContract } from '@/lib/api/contracts/storage-transfer'
 import { getValidationErrorMessage, parseRequest } from '@/lib/api/server'
@@ -38,6 +39,7 @@ const logger = createLogger('FilesParseAPI')
 
 const MAX_DOWNLOAD_SIZE_BYTES = 100 * 1024 * 1024 // 100 MB
 const DOWNLOAD_TIMEOUT_MS = 30000 // 30 seconds
+const BINARY_EXTENSIONS = new Set<string>(binaryExtensionsList)
 
 function isLikelyTextBuffer(fileBuffer: Buffer): boolean {
   return isUtf8(fileBuffer) && !fileBuffer.includes(0)
@@ -866,9 +868,11 @@ function handleGenericBuffer(
   extension: string,
   fileType?: string
 ): ParseResult {
-  const content = isLikelyTextBuffer(fileBuffer)
-    ? fileBuffer.toString('utf-8')
-    : `[Binary ${extension.toUpperCase()} file - ${fileBuffer.length} bytes]`
+  const normalizedExtension = extension.toLowerCase()
+  const content =
+    !BINARY_EXTENSIONS.has(normalizedExtension) && isLikelyTextBuffer(fileBuffer)
+      ? fileBuffer.toString('utf-8')
+      : `[Binary ${normalizedExtension.toUpperCase()} file - ${fileBuffer.length} bytes]`
 
   return {
     success: true,

apps/sim/package.json

@@ -108,6 +108,7 @@
     "ajv": "8.18.0",
     "better-auth": "1.3.12",
     "better-auth-harmony": "1.3.1",
+    "binary-extensions": "3.1.0",
     "browser-image-compression": "^2.0.2",
     "cheerio": "1.1.2",
     "class-variance-authority": "^0.7.1",