import type { ExifTags } from "exifreader"; import { S3 } from "@webiny/aws-sdk/client-s3/index.js"; import { TaskDefinition, TaskHandler } from "@webiny/api-core/features/task/TaskDefinition/index.js"; import { GlobalKeyValueStore } from "@webiny/api-core/features/keyValueStore/index.js"; import { UpdateFileUseCase } from "@webiny/api-file-manager/features/file/UpdateFile/index.js"; import { MetadataReader } from "../WriteFileMetadata/MetadataReader.js"; export interface ExtractMetadataInput { fileId: string; } class ExtractMetadataTaskHandlerImpl implements TaskHandler.Interface { constructor( private keyValueStore: GlobalKeyValueStore.Interface, private updateFileUseCase: UpdateFileUseCase.Interface ) {} public async run({ input, controller }: TaskHandler.RunParams): Promise< TaskDefinition.Result > { if (controller.runtime.isAborted()) { return controller.response.aborted(); } // Load file metadata const metadataReader = new MetadataReader(this.keyValueStore); const fileMetadata = await metadataReader.read(input.fileId); if (!fileMetadata) { return controller.response.error({ message: `File metadata not found for file ID: ${input.fileId}` }); } // Check if it's an image file if (!fileMetadata.contentType.startsWith("image/")) { return controller.response.done(); } // Fetch the image from S3 const s3 = new S3(); const bucket = String(process.env.S3_BUCKET); try { const s3Object = await s3.getObject({ Bucket: bucket, Key: fileMetadata.bucketKey }); if (!s3Object.Body) { return controller.response.error({ message: `S3 object body is empty for file: ${input.fileId}` }); } const sharp = await import(/* webpackChunkName: "sharp" */ "sharp").then( m => m.default ); const ExifReader = await import(/* webpackChunkName: "exifreader" */ "exifreader").then( m => m.default ); // Convert S3 body to buffer const buffer = Buffer.from(await s3Object.Body.transformToByteArray()); // Extract metadata using Sharp for image dimensions const sharpInstance = sharp(buffer); const sharpMetadata = await sharpInstance.metadata(); // Use ExifReader to extract EXIF and IPTC data const tags = ExifReader.load(buffer, { expanded: true }); // Build metadata object const metadata: Record = { image: { width: sharpMetadata.width, height: sharpMetadata.height, format: sharpMetadata.format, orientation: sharpMetadata.orientation ?? 1 } }; // Extract EXIF data if available if (tags.exif) { metadata.exif = this.cleanValues(tags.exif); } // Extract IPTC data if available if (tags.iptc) { metadata.iptc = this.cleanValues(tags.iptc); } // Update the file with extracted metadata const updateResult = await this.updateFileUseCase.execute({ id: input.fileId, metadata }); if (updateResult.isFail()) { return controller.response.error({ message: `Failed to update file with metadata: ${updateResult.error.message}` }); } return controller.response.done(); } catch (error) { return controller.response.error({ message: `Failed to extract metadata: ${ error instanceof Error ? error.message : String(error) }` }); } } private cleanValues(tags: ExifTags) { const cleaned: Record = {}; for (const [key, tag] of Object.entries(tags)) { if (!tag || typeof tag !== "object") { continue; } // Use description if available, otherwise value if (tag.description !== undefined && tag.description !== null) { cleaned[key] = tag.description; } else if (Array.isArray(tag.value) && tag.value.length > 20) { // Skip large byte arrays } else if (tag.value !== undefined) { cleaned[key] = tag.value; } } return cleaned; } } const ExtractMetadataTaskHandler = TaskHandler.createImplementation({ implementation: ExtractMetadataTaskHandlerImpl, dependencies: [GlobalKeyValueStore, UpdateFileUseCase] }); class ExtractMetadataTask implements TaskDefinition.Interface { id = "fileManagerExtractMetadata"; title = "Extract image metadata (dimensions, EXIF, IPTC)"; description = "A task to extract metadata from uploaded image files"; maxIterations = 1; isPrivate = true; databaseLogs = false; selfCleanup = ["onSuccess" as const, "onAbort" as const]; handler = ExtractMetadataTaskHandler; } export const ExtractMetadataTaskDefinition = TaskDefinition.createImplementation({ implementation: ExtractMetadataTask, dependencies: [] });