Esempi per Amazon Textract con SDK per Python (Boto3)

class TextractWrapper:
    """Encapsulates Textract functions."""

    def __init__(self, textract_client, s3_resource, sqs_resource):
        """
        :param textract_client: A Boto3 Textract client.
        :param s3_resource: A Boto3 Amazon S3 resource.
        :param sqs_resource: A Boto3 Amazon SQS resource.
        """
        self.textract_client = textract_client
        self.s3_resource = s3_resource
        self.sqs_resource = sqs_resource


    def analyze_file(
        self, feature_types, *, document_file_name=None, document_bytes=None
    ):
        """
        Detects text and additional elements, such as forms or tables, in a local image
        file or from in-memory byte data.
        The image must be in PNG or JPG format.

        :param feature_types: The types of additional document features to detect.
        :param document_file_name: The name of a document image file.
        :param document_bytes: In-memory byte data of a document image.
        :return: The response from Amazon Textract, including a list of blocks
                 that describe elements detected in the image.
        """
        if document_file_name is not None:
            with open(document_file_name, "rb") as document_file:
                document_bytes = document_file.read()
        try:
            response = self.textract_client.analyze_document(
                Document={"Bytes": document_bytes}, FeatureTypes=feature_types
            )
            logger.info("Detected %s blocks.", len(response["Blocks"]))
        except ClientError:
            logger.exception("Couldn't detect text.")
            raise
        else:
            return response

class TextractWrapper:
    """Encapsulates Textract functions."""

    def __init__(self, textract_client, s3_resource, sqs_resource):
        """
        :param textract_client: A Boto3 Textract client.
        :param s3_resource: A Boto3 Amazon S3 resource.
        :param sqs_resource: A Boto3 Amazon SQS resource.
        """
        self.textract_client = textract_client
        self.s3_resource = s3_resource
        self.sqs_resource = sqs_resource


    def detect_file_text(self, *, document_file_name=None, document_bytes=None):
        """
        Detects text elements in a local image file or from in-memory byte data.
        The image must be in PNG or JPG format.

        :param document_file_name: The name of a document image file.
        :param document_bytes: In-memory byte data of a document image.
        :return: The response from Amazon Textract, including a list of blocks
                 that describe elements detected in the image.
        """
        if document_file_name is not None:
            with open(document_file_name, "rb") as document_file:
                document_bytes = document_file.read()
        try:
            response = self.textract_client.detect_document_text(
                Document={"Bytes": document_bytes}
            )
            logger.info("Detected %s blocks.", len(response["Blocks"]))
        except ClientError:
            logger.exception("Couldn't detect text.")
            raise
        else:
            return response

class TextractWrapper:
    """Encapsulates Textract functions."""

    def __init__(self, textract_client, s3_resource, sqs_resource):
        """
        :param textract_client: A Boto3 Textract client.
        :param s3_resource: A Boto3 Amazon S3 resource.
        :param sqs_resource: A Boto3 Amazon SQS resource.
        """
        self.textract_client = textract_client
        self.s3_resource = s3_resource
        self.sqs_resource = sqs_resource


    def get_analysis_job(self, job_id):
        """
        Gets data for a previously started detection job that includes additional
        elements.

        :param job_id: The ID of the job to retrieve.
        :return: The job data, including a list of blocks that describe elements
                 detected in the image.
        """
        try:
            response = self.textract_client.get_document_analysis(JobId=job_id)
            job_status = response["JobStatus"]
            logger.info("Job %s status is %s.", job_id, job_status)
        except ClientError:
            logger.exception("Couldn't get data for job %s.", job_id)
            raise
        else:
            return response

Per avviare un processo asincrono per analizzare un documento.

class TextractWrapper:
    """Encapsulates Textract functions."""

    def __init__(self, textract_client, s3_resource, sqs_resource):
        """
        :param textract_client: A Boto3 Textract client.
        :param s3_resource: A Boto3 Amazon S3 resource.
        :param sqs_resource: A Boto3 Amazon SQS resource.
        """
        self.textract_client = textract_client
        self.s3_resource = s3_resource
        self.sqs_resource = sqs_resource


    def start_analysis_job(
        self,
        bucket_name,
        document_file_name,
        feature_types,
        sns_topic_arn,
        sns_role_arn,
    ):
        """
        Starts an asynchronous job to detect text and additional elements, such as
        forms or tables, in an image stored in an Amazon S3 bucket. Textract publishes
        a notification to the specified Amazon SNS topic when the job completes.
        The image must be in PNG, JPG, or PDF format.

        :param bucket_name: The name of the Amazon S3 bucket that contains the image.
        :param document_file_name: The name of the document image stored in Amazon S3.
        :param feature_types: The types of additional document features to detect.
        :param sns_topic_arn: The Amazon Resource Name (ARN) of an Amazon SNS topic
                              where job completion notification is published.
        :param sns_role_arn: The ARN of an AWS Identity and Access Management (IAM)
                             role that can be assumed by Textract and grants permission
                             to publish to the Amazon SNS topic.
        :return: The ID of the job.
        """
        try:
            response = self.textract_client.start_document_analysis(
                DocumentLocation={
                    "S3Object": {"Bucket": bucket_name, "Name": document_file_name}
                },
                NotificationChannel={
                    "SNSTopicArn": sns_topic_arn,
                    "RoleArn": sns_role_arn,
                },
                FeatureTypes=feature_types,
            )
            job_id = response["JobId"]
            logger.info(
                "Started text analysis job %s on %s.", job_id, document_file_name
            )
        except ClientError:
            logger.exception("Couldn't analyze text in %s.", document_file_name)
            raise
        else:
            return job_id

Per avviare un processo asincrono per rilevare il testo in un documento.

class TextractWrapper:
    """Encapsulates Textract functions."""

    def __init__(self, textract_client, s3_resource, sqs_resource):
        """
        :param textract_client: A Boto3 Textract client.
        :param s3_resource: A Boto3 Amazon S3 resource.
        :param sqs_resource: A Boto3 Amazon SQS resource.
        """
        self.textract_client = textract_client
        self.s3_resource = s3_resource
        self.sqs_resource = sqs_resource


    def start_detection_job(
        self, bucket_name, document_file_name, sns_topic_arn, sns_role_arn
    ):
        """
        Starts an asynchronous job to detect text elements in an image stored in an
        Amazon S3 bucket. Textract publishes a notification to the specified Amazon SNS
        topic when the job completes.
        The image must be in PNG, JPG, or PDF format.

        :param bucket_name: The name of the Amazon S3 bucket that contains the image.
        :param document_file_name: The name of the document image stored in Amazon S3.
        :param sns_topic_arn: The Amazon Resource Name (ARN) of an Amazon SNS topic
                              where the job completion notification is published.
        :param sns_role_arn: The ARN of an AWS Identity and Access Management (IAM)
                             role that can be assumed by Textract and grants permission
                             to publish to the Amazon SNS topic.
        :return: The ID of the job.
        """
        try:
            response = self.textract_client.start_document_text_detection(
                DocumentLocation={
                    "S3Object": {"Bucket": bucket_name, "Name": document_file_name}
                },
                NotificationChannel={
                    "SNSTopicArn": sns_topic_arn,
                    "RoleArn": sns_role_arn,
                },
            )
            job_id = response["JobId"]
            logger.info(
                "Started text detection job %s on %s.", job_id, document_file_name
            )
        except ClientError:
            logger.exception("Couldn't detect text in %s.", document_file_name)
            raise
        else:
            return job_id

Mostra come usare AWS SDK per Python (Boto3) con Amazon Textract per rilevare elementi di testo, forme e tabelle nell’immagine di un documento. L’immagine di input e l’output di Amazon Textract sono mostrati in un’applicazione Tkinter che consente di esplorare gli elementi rilevati.

Per il codice sorgente completo e le istruzioni su come configurare ed eseguire, consulta l’esempio completo su GitHub.

Mostra come usare AWS SDK per Python (Boto3) in un notebook Jupyter per rilevare entità nel testo estratto da un’immagine. In questo esempio viene utilizzato Amazon Textract per estrarre il testo da un’immagine archiviata in Amazon Simple Storage Service (Amazon S3) e Amazon Comprehend per rilevare le entità nel testo estratto.

Questo esempio è un notebook Jupyter e deve essere eseguito in un ambiente in grado di ospitare notebook. Per istruzioni sull’esecuzione dell’esempio utilizzando Amazon SageMaker AI, consulta le indicazioni in TextractAndComprehendNotebook.ipynb.

Per il codice sorgente completo e le istruzioni su come configurare ed eseguire, consulta l’esempio completo su GitHub.