feat: file type validation (#191)

xgeekshq · Sep 26, 2023 · c5337af · c5337af
1 parent 3691ae3
commit c5337af
Show file tree

Hide file tree

Showing 4 changed files with 93 additions and 6 deletions.
diff --git a/packages/api/src/chats/chats.controller.ts b/packages/api/src/chats/chats.controller.ts
@@ -7,6 +7,7 @@ import { RemoveDocumentFromChatRequestDto } from '@/chats/dtos/remove-document-f
 import { ChatNotFoundExceptionSchema } from '@/chats/exceptions/chat-not-found.exception';
 import { DocumentNotFoundExceptionSchema } from '@/chats/exceptions/document-not-found.exception';
 import { DocumentPermissionsMismatchExceptionSchema } from '@/chats/exceptions/document-permissions-mismatch.exception';
+import { DocumentTypeMismatchExceptionSchema } from '@/chats/exceptions/document-type-mismatch.exception';
 import { MaxDocumentSizeLimitExceptionSchema } from '@/chats/exceptions/max-document-size-limit.exception';
 import { FindChatByRoomIdUsecase } from '@/chats/usecases/find-chat-by-room-id.usecase';
 import { FindChatMessageHistoryByRoomIdUsecase } from '@/chats/usecases/find-chat-message-history-by-room-id.usecase';
@@ -41,6 +42,7 @@ import {
   ApiConflictResponse,
   ApiConsumes,
   ApiNoContentResponse,
+  ApiNotAcceptableResponse,
   ApiNotFoundResponse,
   ApiOkResponse,
   ApiOperation,
@@ -146,6 +148,7 @@ export class ChatsController {
   @ApiNoContentResponse({ description: 'No content' })
   @ApiBadRequestResponse({ schema: MaxDocumentSizeLimitExceptionSchema })
   @ApiConflictResponse({ schema: DocumentPermissionsMismatchExceptionSchema })
+  @ApiNotAcceptableResponse({ schema: DocumentTypeMismatchExceptionSchema })
   @ApiNotFoundResponse({ schema: ChatNotFoundExceptionSchema })
   @ApiOperation({ description: 'Upload documents into a chat' })
   uploadDocumentsToChat(

diff --git a/packages/api/src/chats/exceptions/document-type-mismatch.exception.ts b/packages/api/src/chats/exceptions/document-type-mismatch.exception.ts
@@ -0,0 +1,36 @@
+import { HttpException, HttpStatus } from '@nestjs/common';
+
+export const DocumentTypeMismatchExceptionSchema = {
+  type: 'object',
+  properties: {
+    statusCode: {
+      type: 'number',
+      example: 406,
+    },
+    message: {
+      type: 'array',
+      items: {
+        type: 'string',
+        example: 'document: type_mismatch',
+      },
+    },
+    error: {
+      type: 'string',
+      example: "Document type extension doesn't match its contents",
+    },
+  },
+  required: ['statusCode', 'message', 'error'],
+};
+
+export class DocumentTypeMismatchException extends HttpException {
+  constructor() {
+    super(
+      {
+        statusCode: HttpStatus.NOT_ACCEPTABLE,
+        message: ['document: type_mismatch'],
+        error: "Document type extension doesn't match its contents",
+      },
+      HttpStatus.NOT_ACCEPTABLE
+    );
+  }
+}
diff --git a/packages/api/src/chats/usecases/upload-documents-to-chat.usecase.ts b/packages/api/src/chats/usecases/upload-documents-to-chat.usecase.ts
@@ -1,11 +1,14 @@
-import { AiService } from '@/ai/facades/ai.service';
-import { AppConfigService } from '@/app-config/app-config.service';
+import { readFileSync } from 'fs';
 import { ClerkAuthUserProvider } from '@/auth/providers/clerk/clerk-auth-user.provider';
 import { ChatsRepository } from '@/chats/chats.repository';
 import { ChatNotFoundException } from '@/chats/exceptions/chat-not-found.exception';
 import { DocumentPermissionsMismatchException } from '@/chats/exceptions/document-permissions-mismatch.exception';
+import { DocumentTypeMismatchException } from '@/chats/exceptions/document-type-mismatch.exception';
 import { MaxDocumentSizeLimitException } from '@/chats/exceptions/max-document-size-limit.exception';
-import { ACCEPTED_FILE_SIZE_LIMIT } from '@/common/constants/files';
+import {
+  ACCEPTED_FILE_SIZE_LIMIT,
+  fileBufferSignatureByMimetypeMap,
+} from '@/common/constants/files';
 import { CHAT_DOCUMENT_UPLOAD_QUEUE } from '@/common/constants/queues';
 import { createChatDocUploadJobFactory } from '@/common/jobs/chat-doc-upload.job';
 import { Chat } from '@/common/types/chat';
@@ -21,9 +24,7 @@ export class UploadDocumentsToChatUsecase implements Usecase {
     private readonly chatsRepository: ChatsRepository,
     private readonly clerkAuthUserProvider: ClerkAuthUserProvider,
     @InjectQueue(CHAT_DOCUMENT_UPLOAD_QUEUE)
-    private readonly chatDocUploadQueue: Queue,
-    private readonly appConfigService: AppConfigService,
-    private readonly aiService: AiService
+    private readonly chatDocUploadQueue: Queue
   ) {}
 
   async execute(
@@ -39,6 +40,8 @@ export class UploadDocumentsToChatUsecase implements Usecase {
 
     this.checkMaxDocumentsSizePerRoomInvariant(existingChat, files);
 
+    this.checkForValidDocumentTypesInvariant(files);
+
     await this.checkDocumentRolesMatchUserRolesInvariant(
       existingChat,
       fileRoles
@@ -79,6 +82,22 @@ export class UploadDocumentsToChatUsecase implements Usecase {
     }
   }
 
+  private checkForValidDocumentTypesInvariant(files: Express.Multer.File[]) {
+    for (const file of files) {
+      const buffer = readFileSync(`${file.destination}/${file.filename}`);
+      const fileSignatureVerifier =
+        fileBufferSignatureByMimetypeMap[file.mimetype];
+      const fileSignature = buffer.toString(
+        'hex',
+        0,
+        fileSignatureVerifier.relevantBytes
+      );
+      if (!fileSignatureVerifier.signature.includes(fileSignature)) {
+        throw new DocumentTypeMismatchException();
+      }
+    }
+  }
+
   private async checkDocumentRolesMatchUserRolesInvariant(
     chat: Chat,
     fileRoles: string[]

diff --git a/packages/api/src/common/constants/files.ts b/packages/api/src/common/constants/files.ts
@@ -16,6 +16,35 @@ export const acceptedFileMimetypes = [
   DOCX_MIMETYPE,
 ];
 
+export const PDF_BUFFER_SIGNATURE = '25504446'; // 4 relevant bytes
+export const TXT_UTF8_BUFFER_SIGNATURE = 'EFBBBF'; // 3 relevant bytes
+export const TXT_BUFFER_SIGNATURE = '546869'; // 3 relevant bytes
+export const DOCX_BUFFER_SIGNATURE = '504b0304'; // 4 relevant bytes
+
+export const acceptedFileBufferSignatures = [
+  PDF_BUFFER_SIGNATURE,
+  TXT_UTF8_BUFFER_SIGNATURE,
+  TXT_BUFFER_SIGNATURE,
+  DOCX_BUFFER_SIGNATURE,
+];
+
+export const fileBufferSignatureByMimetypeMap: {
+  [key: string]: { relevantBytes: number; signature: string | string[] };
+} = {
+  [PDF_MIMETYPE]: {
+    relevantBytes: 4,
+    signature: [PDF_BUFFER_SIGNATURE],
+  },
+  [TEXT_MIMETYPE]: {
+    relevantBytes: 3,
+    signature: [TXT_UTF8_BUFFER_SIGNATURE, TXT_BUFFER_SIGNATURE],
+  },
+  [DOCX_MIMETYPE]: {
+    relevantBytes: 4,
+    signature: [DOCX_BUFFER_SIGNATURE],
+  },
+};
+
 export function sanitizeFilename(filename: string): string {
   const roomIdLength = 24;
   const maxCollectionNameLength = 62;