Search in sources :

Example 6 with FindingLimits

use of com.google.privacy.dlp.v2.InspectConfig.FindingLimits in project java-docs-samples by GoogleCloudPlatform.

the class Inspect method inspectGcsFile.

// [END dlp_inspect_file]
// [START dlp_inspect_gcs]
/**
 * Inspect GCS file for Info types and wait on job completion using Google Cloud Pub/Sub
 * notification
 *
 * @param bucketName The name of the bucket where the file resides.
 * @param fileName The path to the file within the bucket to inspect (can include wildcards, eg.
 *     my-image.*)
 * @param minLikelihood The minimum likelihood required before returning a match
 * @param infoTypes The infoTypes of information to match
 * @param maxFindings The maximum number of findings to report (0 = server maximum)
 * @param topicId Google Cloud Pub/Sub topic Id to notify of job status
 * @param subscriptionId Google Cloud Subscription to above topic to listen for job status updates
 * @param projectId Google Cloud project ID
 */
private static void inspectGcsFile(String bucketName, String fileName, Likelihood minLikelihood, List<InfoType> infoTypes, int maxFindings, String topicId, String subscriptionId, String projectId) throws Exception {
    // Instantiates a client
    try (DlpServiceClient dlpServiceClient = DlpServiceClient.create()) {
        CloudStorageOptions cloudStorageOptions = CloudStorageOptions.newBuilder().setFileSet(CloudStorageOptions.FileSet.newBuilder().setUrl("gs://" + bucketName + "/" + fileName)).build();
        StorageConfig storageConfig = StorageConfig.newBuilder().setCloudStorageOptions(cloudStorageOptions).build();
        FindingLimits findingLimits = FindingLimits.newBuilder().setMaxFindingsPerRequest(maxFindings).build();
        InspectConfig inspectConfig = InspectConfig.newBuilder().addAllInfoTypes(infoTypes).setMinLikelihood(minLikelihood).setLimits(findingLimits).build();
        String pubSubTopic = String.format("projects/%s/topics/%s", projectId, topicId);
        Action.PublishToPubSub publishToPubSub = Action.PublishToPubSub.newBuilder().setTopic(pubSubTopic).build();
        Action action = Action.newBuilder().setPubSub(publishToPubSub).build();
        InspectJobConfig inspectJobConfig = InspectJobConfig.newBuilder().setStorageConfig(storageConfig).setInspectConfig(inspectConfig).addActions(action).build();
        // Semi-synchronously submit an inspect job, and wait on results
        CreateDlpJobRequest createDlpJobRequest = CreateDlpJobRequest.newBuilder().setParent(ProjectName.of(projectId).toString()).setInspectJob(inspectJobConfig).build();
        DlpJob dlpJob = dlpServiceClient.createDlpJob(createDlpJobRequest);
        System.out.println("Job created with ID:" + dlpJob.getName());
        final SettableApiFuture<Boolean> done = SettableApiFuture.create();
        // Set up a Pub/Sub subscriber to listen on the job completion status
        Subscriber subscriber = Subscriber.newBuilder(ProjectSubscriptionName.of(projectId, subscriptionId), (pubsubMessage, ackReplyConsumer) -> {
            if (pubsubMessage.getAttributesCount() > 0 && pubsubMessage.getAttributesMap().get("DlpJobName").equals(dlpJob.getName())) {
                // notify job completion
                done.set(true);
                ackReplyConsumer.ack();
            }
        }).build();
        subscriber.startAsync();
        // For long jobs, consider using a truly asynchronous execution model such as Cloud Functions
        try {
            done.get(1, TimeUnit.MINUTES);
            // Wait for the job to become available
            Thread.sleep(500);
        } catch (Exception e) {
            System.out.println("Unable to verify job completion.");
        }
        DlpJob completedJob = dlpServiceClient.getDlpJob(GetDlpJobRequest.newBuilder().setName(dlpJob.getName()).build());
        System.out.println("Job status: " + completedJob.getState());
        InspectDataSourceDetails inspectDataSourceDetails = completedJob.getInspectDetails();
        InspectDataSourceDetails.Result result = inspectDataSourceDetails.getResult();
        if (result.getInfoTypeStatsCount() > 0) {
            System.out.println("Findings: ");
            for (InfoTypeStats infoTypeStat : result.getInfoTypeStatsList()) {
                System.out.print("\tInfo type: " + infoTypeStat.getInfoType().getName());
                System.out.println("\tCount: " + infoTypeStat.getCount());
            }
        } else {
            System.out.println("No findings.");
        }
    }
}
Also used : ByteContentItem(com.google.privacy.dlp.v2.ByteContentItem) InspectResult(com.google.privacy.dlp.v2.InspectResult) KindExpression(com.google.privacy.dlp.v2.KindExpression) Options(org.apache.commons.cli.Options) Likelihood(com.google.privacy.dlp.v2.Likelihood) Subscriber(com.google.cloud.pubsub.v1.Subscriber) BigQueryOptions(com.google.privacy.dlp.v2.BigQueryOptions) HelpFormatter(org.apache.commons.cli.HelpFormatter) MimetypesFileTypeMap(javax.activation.MimetypesFileTypeMap) CloudStorageOptions(com.google.privacy.dlp.v2.CloudStorageOptions) ArrayList(java.util.ArrayList) DefaultParser(org.apache.commons.cli.DefaultParser) InspectDataSourceDetails(com.google.privacy.dlp.v2.InspectDataSourceDetails) FindingLimits(com.google.privacy.dlp.v2.InspectConfig.FindingLimits) InspectContentResponse(com.google.privacy.dlp.v2.InspectContentResponse) ServiceOptions(com.google.cloud.ServiceOptions) URLConnection(java.net.URLConnection) StorageConfig(com.google.privacy.dlp.v2.StorageConfig) CommandLine(org.apache.commons.cli.CommandLine) PartitionId(com.google.privacy.dlp.v2.PartitionId) Action(com.google.privacy.dlp.v2.Action) DatastoreOptions(com.google.privacy.dlp.v2.DatastoreOptions) ProjectTopicName(com.google.pubsub.v1.ProjectTopicName) Option(org.apache.commons.cli.Option) InspectJobConfig(com.google.privacy.dlp.v2.InspectJobConfig) DlpServiceClient(com.google.cloud.dlp.v2.DlpServiceClient) Finding(com.google.privacy.dlp.v2.Finding) CreateDlpJobRequest(com.google.privacy.dlp.v2.CreateDlpJobRequest) Files(java.nio.file.Files) CommandLineParser(org.apache.commons.cli.CommandLineParser) ContentItem(com.google.privacy.dlp.v2.ContentItem) InfoType(com.google.privacy.dlp.v2.InfoType) InfoTypeStats(com.google.privacy.dlp.v2.InfoTypeStats) SettableApiFuture(com.google.api.core.SettableApiFuture) ByteString(com.google.protobuf.ByteString) TimeUnit(java.util.concurrent.TimeUnit) InspectConfig(com.google.privacy.dlp.v2.InspectConfig) List(java.util.List) ProjectName(com.google.privacy.dlp.v2.ProjectName) GetDlpJobRequest(com.google.privacy.dlp.v2.GetDlpJobRequest) Paths(java.nio.file.Paths) ParseException(org.apache.commons.cli.ParseException) BigQueryTable(com.google.privacy.dlp.v2.BigQueryTable) ProjectSubscriptionName(com.google.pubsub.v1.ProjectSubscriptionName) OptionGroup(org.apache.commons.cli.OptionGroup) DlpJob(com.google.privacy.dlp.v2.DlpJob) InspectContentRequest(com.google.privacy.dlp.v2.InspectContentRequest) Collections(java.util.Collections) Action(com.google.privacy.dlp.v2.Action) FindingLimits(com.google.privacy.dlp.v2.InspectConfig.FindingLimits) StorageConfig(com.google.privacy.dlp.v2.StorageConfig) InspectDataSourceDetails(com.google.privacy.dlp.v2.InspectDataSourceDetails) ByteString(com.google.protobuf.ByteString) InspectConfig(com.google.privacy.dlp.v2.InspectConfig) CreateDlpJobRequest(com.google.privacy.dlp.v2.CreateDlpJobRequest) ParseException(org.apache.commons.cli.ParseException) InfoTypeStats(com.google.privacy.dlp.v2.InfoTypeStats) Subscriber(com.google.cloud.pubsub.v1.Subscriber) DlpServiceClient(com.google.cloud.dlp.v2.DlpServiceClient) CloudStorageOptions(com.google.privacy.dlp.v2.CloudStorageOptions) DlpJob(com.google.privacy.dlp.v2.DlpJob) InspectJobConfig(com.google.privacy.dlp.v2.InspectJobConfig)

Example 7 with FindingLimits

use of com.google.privacy.dlp.v2.InspectConfig.FindingLimits in project java-docs-samples by GoogleCloudPlatform.

the class Inspect method inspectFile.

// [END dlp_inspect_string]
// [START dlp_inspect_file]
/**
 * Inspect a local file
 *
 * @param filePath The path to a local file to inspect. Can be a text, JPG, or PNG file.
 * @param minLikelihood The minimum likelihood required before returning a match
 * @param maxFindings The maximum number of findings to report (0 = server maximum)
 * @param infoTypes The infoTypes of information to match
 * @param includeQuote Whether to include the matching string
 * @param projectId Google Cloud project ID
 */
private static void inspectFile(String filePath, Likelihood minLikelihood, int maxFindings, List<InfoType> infoTypes, boolean includeQuote, String projectId) {
    // Instantiates a client
    try (DlpServiceClient dlpServiceClient = DlpServiceClient.create()) {
        // detect file mime type, default to application/octet-stream
        String mimeType = URLConnection.guessContentTypeFromName(filePath);
        if (mimeType == null) {
            mimeType = MimetypesFileTypeMap.getDefaultFileTypeMap().getContentType(filePath);
        }
        ByteContentItem.BytesType bytesType;
        switch(mimeType) {
            case "image/jpeg":
                bytesType = ByteContentItem.BytesType.IMAGE_JPEG;
                break;
            case "image/bmp":
                bytesType = ByteContentItem.BytesType.IMAGE_BMP;
                break;
            case "image/png":
                bytesType = ByteContentItem.BytesType.IMAGE_PNG;
                break;
            case "image/svg":
                bytesType = ByteContentItem.BytesType.IMAGE_SVG;
                break;
            default:
                bytesType = ByteContentItem.BytesType.BYTES_TYPE_UNSPECIFIED;
                break;
        }
        byte[] data = Files.readAllBytes(Paths.get(filePath));
        ByteContentItem byteContentItem = ByteContentItem.newBuilder().setType(bytesType).setData(ByteString.copyFrom(data)).build();
        ContentItem contentItem = ContentItem.newBuilder().setByteItem(byteContentItem).build();
        FindingLimits findingLimits = FindingLimits.newBuilder().setMaxFindingsPerRequest(maxFindings).build();
        InspectConfig inspectConfig = InspectConfig.newBuilder().addAllInfoTypes(infoTypes).setMinLikelihood(minLikelihood).setLimits(findingLimits).setIncludeQuote(includeQuote).build();
        InspectContentRequest request = InspectContentRequest.newBuilder().setParent(ProjectName.of(projectId).toString()).setInspectConfig(inspectConfig).setItem(contentItem).build();
        InspectContentResponse response = dlpServiceClient.inspectContent(request);
        InspectResult result = response.getResult();
        if (result.getFindingsCount() > 0) {
            System.out.println("Findings: ");
            for (Finding finding : result.getFindingsList()) {
                if (includeQuote) {
                    System.out.print("\tQuote: " + finding.getQuote());
                }
                System.out.print("\tInfo type: " + finding.getInfoType().getName());
                System.out.println("\tLikelihood: " + finding.getLikelihood());
            }
        } else {
            System.out.println("No findings.");
        }
    } catch (Exception e) {
        System.out.println("Error in inspectFile: " + e.getMessage());
    }
}
Also used : FindingLimits(com.google.privacy.dlp.v2.InspectConfig.FindingLimits) ByteContentItem(com.google.privacy.dlp.v2.ByteContentItem) ByteString(com.google.protobuf.ByteString) InspectContentRequest(com.google.privacy.dlp.v2.InspectContentRequest) InspectConfig(com.google.privacy.dlp.v2.InspectConfig) ParseException(org.apache.commons.cli.ParseException) DlpServiceClient(com.google.cloud.dlp.v2.DlpServiceClient) InspectContentResponse(com.google.privacy.dlp.v2.InspectContentResponse) Finding(com.google.privacy.dlp.v2.Finding) InspectResult(com.google.privacy.dlp.v2.InspectResult) ByteContentItem(com.google.privacy.dlp.v2.ByteContentItem) ContentItem(com.google.privacy.dlp.v2.ContentItem)

Example 8 with FindingLimits

use of com.google.privacy.dlp.v2.InspectConfig.FindingLimits in project java-docs-samples by GoogleCloudPlatform.

the class Triggers method createTrigger.

// [START dlp_create_trigger]
/**
 * Schedule a DLP inspection trigger for a GCS location.
 *
 * @param triggerId (Optional) name of the trigger to be created
 * @param displayName (Optional) display name for the trigger to be created
 * @param description (Optional) description for the trigger to be created
 * @param scanPeriod How often to wait between scans, in days (minimum = 1 day)
 * @param infoTypes infoTypes of information to match eg. InfoType.PHONE_NUMBER,
 *     InfoType.EMAIL_ADDRESS
 * @param minLikelihood minimum likelihood required before returning a match
 * @param maxFindings maximum number of findings to report per request (0 = server maximum)
 * @param projectId The project ID to run the API call under
 */
private static void createTrigger(String triggerId, String displayName, String description, String bucketName, String fileName, int scanPeriod, List<InfoType> infoTypes, Likelihood minLikelihood, int maxFindings, String projectId) throws Exception {
    // instantiate a client
    DlpServiceClient dlpServiceClient = DlpServiceClient.create();
    try {
        CloudStorageOptions cloudStorageOptions = CloudStorageOptions.newBuilder().setFileSet(CloudStorageOptions.FileSet.newBuilder().setUrl("gs://" + bucketName + "/" + fileName)).build();
        StorageConfig storageConfig = StorageConfig.newBuilder().setCloudStorageOptions(cloudStorageOptions).build();
        InspectConfig.FindingLimits findingLimits = InspectConfig.FindingLimits.newBuilder().setMaxFindingsPerRequest(maxFindings).build();
        InspectConfig inspectConfig = InspectConfig.newBuilder().addAllInfoTypes(infoTypes).setMinLikelihood(minLikelihood).setLimits(findingLimits).build();
        InspectJobConfig inspectJobConfig = InspectJobConfig.newBuilder().setInspectConfig(inspectConfig).setStorageConfig(storageConfig).build();
        // Schedule scan of GCS bucket every scanPeriod number of days (minimum = 1 day)
        Duration duration = Duration.newBuilder().setSeconds(scanPeriod * 24 * 3600).build();
        Schedule schedule = Schedule.newBuilder().setRecurrencePeriodDuration(duration).build();
        JobTrigger.Trigger trigger = JobTrigger.Trigger.newBuilder().setSchedule(schedule).build();
        JobTrigger jobTrigger = JobTrigger.newBuilder().setInspectJob(inspectJobConfig).setName(triggerId).setDisplayName(displayName).setDescription(description).setStatus(JobTrigger.Status.HEALTHY).addTriggers(trigger).build();
        // Create scan request
        CreateJobTriggerRequest createJobTriggerRequest = CreateJobTriggerRequest.newBuilder().setParent(ProjectName.of(projectId).toString()).setJobTrigger(jobTrigger).build();
        JobTrigger createdJobTrigger = dlpServiceClient.createJobTrigger(createJobTriggerRequest);
        System.out.println("Created Trigger: " + createdJobTrigger.getName());
    } catch (Exception e) {
        System.out.println("Error creating trigger: " + e.getMessage());
    }
}
Also used : StorageConfig(com.google.privacy.dlp.v2.StorageConfig) DlpServiceClient(com.google.cloud.dlp.v2.DlpServiceClient) Schedule(com.google.privacy.dlp.v2.Schedule) CloudStorageOptions(com.google.privacy.dlp.v2.CloudStorageOptions) Duration(com.google.protobuf.Duration) InspectConfig(com.google.privacy.dlp.v2.InspectConfig) JobTrigger(com.google.privacy.dlp.v2.JobTrigger) ParseException(org.apache.commons.cli.ParseException) InspectJobConfig(com.google.privacy.dlp.v2.InspectJobConfig) CreateJobTriggerRequest(com.google.privacy.dlp.v2.CreateJobTriggerRequest)

Aggregations

DlpServiceClient (com.google.cloud.dlp.v2.DlpServiceClient)8 InspectConfig (com.google.privacy.dlp.v2.InspectConfig)8 ParseException (org.apache.commons.cli.ParseException)7 ByteContentItem (com.google.privacy.dlp.v2.ByteContentItem)6 ContentItem (com.google.privacy.dlp.v2.ContentItem)6 Finding (com.google.privacy.dlp.v2.Finding)6 FindingLimits (com.google.privacy.dlp.v2.InspectConfig.FindingLimits)6 InspectContentRequest (com.google.privacy.dlp.v2.InspectContentRequest)6 InspectContentResponse (com.google.privacy.dlp.v2.InspectContentResponse)6 InspectResult (com.google.privacy.dlp.v2.InspectResult)5 ByteString (com.google.protobuf.ByteString)5 CloudStorageOptions (com.google.privacy.dlp.v2.CloudStorageOptions)4 InfoType (com.google.privacy.dlp.v2.InfoType)4 InspectJobConfig (com.google.privacy.dlp.v2.InspectJobConfig)4 Likelihood (com.google.privacy.dlp.v2.Likelihood)4 StorageConfig (com.google.privacy.dlp.v2.StorageConfig)4 SettableApiFuture (com.google.api.core.SettableApiFuture)3 ServiceOptions (com.google.cloud.ServiceOptions)3 Subscriber (com.google.cloud.pubsub.v1.Subscriber)3 Action (com.google.privacy.dlp.v2.Action)3