com.adobe.pdfservices.operation.pdfjobs.jobs.ExtractPDFJob Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of pdfservices-sdk Show documentation
Adobe PDF Services SDK allows you to access RESTful APIs to create, convert, and manipulate PDFs within your applications. Older versions can be found under groupId: com.adobe.documentservices, artifactId: pdftools-sdk
There is a newer version: 4.2.0
Show newest version
/*
 * Copyright 2024 Adobe
 * All Rights Reserved.
 *
 * NOTICE: Adobe permits you to use, modify, and distribute this file in
 * accordance with the terms of the Adobe license agreement accompanying
 * it. If you have received this file from a source other than Adobe,
 * then your use, modification, or distribution of it requires the prior
 * written permission of Adobe.
 */

package com.adobe.pdfservices.operation.pdfjobs.jobs;

import com.adobe.pdfservices.operation.PDFServicesJob;
import com.adobe.pdfservices.operation.config.notifier.NotifierConfig;
import com.adobe.pdfservices.operation.exception.ServiceApiException;
import com.adobe.pdfservices.operation.internal.ExecutionContext;
import com.adobe.pdfservices.operation.internal.PDFServicesHelper;
import com.adobe.pdfservices.operation.internal.constants.CustomErrorMessages;
import com.adobe.pdfservices.operation.internal.constants.OperationHeaderInfoEndpointMap;
import com.adobe.pdfservices.operation.internal.dto.request.PlatformApiRequest;
import com.adobe.pdfservices.operation.internal.dto.request.extractpdf.ExtractPDFExternalAssetRequest;
import com.adobe.pdfservices.operation.internal.dto.request.extractpdf.ExtractPDFInternalAssetRequest;
import com.adobe.pdfservices.operation.internal.http.DefaultRequestHeaders;
import com.adobe.pdfservices.operation.internal.http.HttpResponse;
import com.adobe.pdfservices.operation.internal.util.ObjectUtil;
import com.adobe.pdfservices.operation.io.Asset;
import com.adobe.pdfservices.operation.io.CloudAsset;
import com.adobe.pdfservices.operation.io.ExternalAsset;
import com.adobe.pdfservices.operation.pdfjobs.params.extractpdf.ExtractPDFParams;

import java.util.List;
import java.util.UUID;

/**
 * A job that extracts pdf elements such as text, images, tables in a structured format from a PDF.
 * Sample Usage:
 * {@code
 *             InputStream inputStream = new FileInputStream(new File("SOURCE_PATH"));
 *
 *             Credentials credentials = new ServicePrincipalCredentials(
 *                     System.getenv("PDF_SERVICES_CLIENT_ID"),
 *                     System.getenv("PDF_SERVICES_CLIENT_SECRET"));
 *
 *             PDFServices pdfServices = new PDFServices(credentials);
 *
 *             Asset asset = pdfServices.upload(inputStream, PDFServicesMediaType.PDF.getMediaType());
 *
 *             ExtractPDFParams extractPDFParams = ExtractPDFParams.extractPDFParamsBuilder()
 *                     .addElementsToExtract(Arrays.asList(ExtractElementType.TEXT))
 *                     .build();
 *
 *             ExtractPDFJob extractPDFJob = new ExtractPDFJob(asset).setParams(extractPDFParams);
 *
 *             String location = pdfServices.submit(extractPDFJob);
 *             PDFServicesResponse pdfServicesResponse = pdfServices.getJobResult(location, ExtractPDFResult.class);
 *
 *             Asset resultAsset = pdfServicesResponse.getResult().getResource();
 *             StreamAsset streamAsset = pdfServices.getContent(resultAsset);
 * }
 */
public class ExtractPDFJob extends PDFServicesJob {
    private Asset inputAsset;

    private Asset outputAsset;
    private ExtractPDFParams extractPDFParams;

    /**
     * Constructs a new {@code ExtractPDFJob} instance.
     *
     * @param asset {@link Asset} object containing the input file; can not be null.
     */
    public ExtractPDFJob(Asset asset) {
        ObjectUtil.requireNonNull(asset, String.format(CustomErrorMessages.GENERIC_CAN_NOT_BE_NULL, "Input Asset"));
        this.inputAsset = asset;
    }

    @Override
    protected String process(ExecutionContext executionContext) throws ServiceApiException {
        return this.process(executionContext, null);
    }

    @Override
    protected String process(ExecutionContext executionContext, List notifyConfigList) throws ServiceApiException {
        this.validate(executionContext);
        PlatformApiRequest extractPDFRequest = generatePlatformApiRequest(notifyConfigList);
        String xRequestId = UUID.randomUUID().toString();
        HttpResponse response = PDFServicesHelper.submitJob(executionContext, extractPDFRequest, xRequestId,
                OperationHeaderInfoEndpointMap.EXTRACT_PDF);
        return response.getHeaders().get(DefaultRequestHeaders.LOCATION_HEADER_NAME);
    }


    /**
     * Sets the parameters for the job.
     *
     * @param extractPDFParams {@link ExtractPDFParams} object containing the extract PDF parameters; can not be null.
     * @return {@code ExtractPDFJob} instance
     */
    public ExtractPDFJob setParams(ExtractPDFParams extractPDFParams) {
        ObjectUtil.requireNonNull(extractPDFParams, String.format(CustomErrorMessages.GENERIC_CAN_NOT_BE_NULL,
                "Extract PDF parameters"));
        this.extractPDFParams = extractPDFParams;
        return this;
    }

    /**
     * Sets the output asset for the job.
     * {@code @note} External assets can be set as output only when input is external asset as well
     *
     * @param asset {@link Asset} object representing the output asset; can not be null.
     * @return {@code ExtractPDFJob} instance
     */
    public ExtractPDFJob setOutput(Asset asset) {
        ObjectUtil.requireNonNull(asset, String.format(CustomErrorMessages.GENERIC_CAN_NOT_BE_NULL, "Output asset"));
        if (this.inputAsset instanceof CloudAsset) {
            throw new IllegalArgumentException(CustomErrorMessages.SET_OUTPUT_VALIDATE);
        }

        this.outputAsset = asset;
        return this;
    }

    private PlatformApiRequest generatePlatformApiRequest(List notifyConfigList) {
        PlatformApiRequest extractPDFRequest;
        if (this.inputAsset instanceof CloudAsset) {
            extractPDFRequest = new ExtractPDFInternalAssetRequest(( (CloudAsset) this.inputAsset ).getAssetId(),
                    this.extractPDFParams, notifyConfigList);
        } else {
            extractPDFRequest = new ExtractPDFExternalAssetRequest((ExternalAsset) this.inputAsset,
                    this.extractPDFParams, notifyConfigList).setOutput((ExternalAsset) outputAsset);
        }
        return extractPDFRequest;
    }
}