AWS Glue SDK for Python(Boto3)을 사용한 예제

import boto3
from botocore.exceptions import ClientError


def hello_glue():
    """
    Lists the job definitions in your AWS Glue account, using the AWS SDK for Python (Boto3).
    """
    try:
        # Create the Glue client
        glue = boto3.client("glue")

        # List the jobs, limiting the results to 10 per page
        paginator = glue.get_paginator("get_jobs")
        response_iterator = paginator.paginate(
            PaginationConfig={"MaxItems": 10, "PageSize": 10}
        )

        # Print the job names
        print("Here are the jobs in your account:")
        for page in response_iterator:
            for job in page["Jobs"]:
                print(f"\t{job['Name']}")

    except ClientError as e:
        print(f"Error: {e}")


if __name__ == "__main__":
    hello_glue()

시나리오에 사용되는 AWS Glue 함수를 래핑하는 클래스를 생성합니다.

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_crawler(self, name):
        """
        Gets information about a crawler.

        :param name: The name of the crawler to look up.
        :return: Data about the crawler.
        """
        crawler = None
        try:
            response = self.glue_client.get_crawler(Name=name)
            crawler = response["Crawler"]
        except ClientError as err:
            if err.response["Error"]["Code"] == "EntityNotFoundException":
                logger.info("Crawler %s doesn't exist.", name)
            else:
                logger.error(
                    "Couldn't get crawler %s. Here's why: %s: %s",
                    name,
                    err.response["Error"]["Code"],
                    err.response["Error"]["Message"],
                )
                raise
        return crawler


    def create_crawler(self, name, role_arn, db_name, db_prefix, s3_target):
        """
        Creates a crawler that can crawl the specified target and populate a
        database in your AWS Glue Data Catalog with metadata that describes the data
        in the target.

        :param name: The name of the crawler.
        :param role_arn: The Amazon Resource Name (ARN) of an AWS Identity and Access
                         Management (IAM) role that grants permission to let AWS Glue
                         access the resources it needs.
        :param db_name: The name to give the database that is created by the crawler.
        :param db_prefix: The prefix to give any database tables that are created by
                          the crawler.
        :param s3_target: The URL to an S3 bucket that contains data that is
                          the target of the crawler.
        """
        try:
            self.glue_client.create_crawler(
                Name=name,
                Role=role_arn,
                DatabaseName=db_name,
                TablePrefix=db_prefix,
                Targets={"S3Targets": [{"Path": s3_target}]},
            )
        except ClientError as err:
            logger.error(
                "Couldn't create crawler. Here's why: %s: %s",
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def start_crawler(self, name):
        """
        Starts a crawler. The crawler crawls its configured target and creates
        metadata that describes the data it finds in the target data source.

        :param name: The name of the crawler to start.
        """
        try:
            self.glue_client.start_crawler(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't start crawler %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def get_database(self, name):
        """
        Gets information about a database in your Data Catalog.

        :param name: The name of the database to look up.
        :return: Information about the database.
        """
        try:
            response = self.glue_client.get_database(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't get database %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["Database"]


    def get_tables(self, db_name):
        """
        Gets a list of tables in a Data Catalog database.

        :param db_name: The name of the database to query.
        :return: The list of tables in the database.
        """
        try:
            response = self.glue_client.get_tables(DatabaseName=db_name)
        except ClientError as err:
            logger.error(
                "Couldn't get tables %s. Here's why: %s: %s",
                db_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["TableList"]


    def create_job(self, name, description, role_arn, script_location):
        """
        Creates a job definition for an extract, transform, and load (ETL) job that can
        be run by AWS Glue.

        :param name: The name of the job definition.
        :param description: The description of the job definition.
        :param role_arn: The ARN of an IAM role that grants AWS Glue the permissions
                         it requires to run the job.
        :param script_location: The Amazon S3 URL of a Python ETL script that is run as
                                part of the job. The script defines how the data is
                                transformed.
        """
        try:
            self.glue_client.create_job(
                Name=name,
                Description=description,
                Role=role_arn,
                Command={
                    "Name": "glueetl",
                    "ScriptLocation": script_location,
                    "PythonVersion": "3",
                },
                GlueVersion="3.0",
            )
        except ClientError as err:
            logger.error(
                "Couldn't create job %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def start_job_run(self, name, input_database, input_table, output_bucket_name):
        """
        Starts a job run. A job run extracts data from the source, transforms it,
        and loads it to the output bucket.

        :param name: The name of the job definition.
        :param input_database: The name of the metadata database that contains tables
                               that describe the source data. This is typically created
                               by a crawler.
        :param input_table: The name of the table in the metadata database that
                            describes the source data.
        :param output_bucket_name: The S3 bucket where the output is written.
        :return: The ID of the job run.
        """
        try:
            # The custom Arguments that are passed to this function are used by the
            # Python ETL script to determine the location of input and output data.
            response = self.glue_client.start_job_run(
                JobName=name,
                Arguments={
                    "--input_database": input_database,
                    "--input_table": input_table,
                    "--output_bucket_url": f"s3://{output_bucket_name}/",
                },
            )
        except ClientError as err:
            logger.error(
                "Couldn't start job run %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRunId"]


    def list_jobs(self):
        """
        Lists the names of job definitions in your account.

        :return: The list of job definition names.
        """
        try:
            response = self.glue_client.list_jobs()
        except ClientError as err:
            logger.error(
                "Couldn't list jobs. Here's why: %s: %s",
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobNames"]


    def get_job_runs(self, job_name):
        """
        Gets information about runs that have been performed for a specific job
        definition.

        :param job_name: The name of the job definition to look up.
        :return: The list of job runs.
        """
        try:
            response = self.glue_client.get_job_runs(JobName=job_name)
        except ClientError as err:
            logger.error(
                "Couldn't get job runs for %s. Here's why: %s: %s",
                job_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRuns"]


    def get_job_run(self, name, run_id):
        """
        Gets information about a single job run.

        :param name: The name of the job definition for the run.
        :param run_id: The ID of the run.
        :return: Information about the run.
        """
        try:
            response = self.glue_client.get_job_run(JobName=name, RunId=run_id)
        except ClientError as err:
            logger.error(
                "Couldn't get job run %s/%s. Here's why: %s: %s",
                name,
                run_id,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRun"]


    def delete_job(self, job_name):
        """
        Deletes a job definition. This also deletes data about all runs that are
        associated with this job definition.

        :param job_name: The name of the job definition to delete.
        """
        try:
            self.glue_client.delete_job(JobName=job_name)
        except ClientError as err:
            logger.error(
                "Couldn't delete job %s. Here's why: %s: %s",
                job_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def delete_table(self, db_name, table_name):
        """
        Deletes a table from a metadata database.

        :param db_name: The name of the database that contains the table.
        :param table_name: The name of the table to delete.
        """
        try:
            self.glue_client.delete_table(DatabaseName=db_name, Name=table_name)
        except ClientError as err:
            logger.error(
                "Couldn't delete table %s. Here's why: %s: %s",
                table_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def delete_database(self, name):
        """
        Deletes a metadata database from your Data Catalog.

        :param name: The name of the database to delete.
        """
        try:
            self.glue_client.delete_database(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't delete database %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise


    def delete_crawler(self, name):
        """
        Deletes a crawler.

        :param name: The name of the crawler to delete.
        """
        try:
            self.glue_client.delete_crawler(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't delete crawler %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

시나리오를 실행하는 클래스를 생성합니다.

class GlueCrawlerJobScenario:
    """
    Encapsulates a scenario that shows how to create an AWS Glue crawler and job and use
    them to transform data from CSV to JSON format.
    """

    def __init__(self, glue_client, glue_service_role, glue_bucket):
        """
        :param glue_client: A Boto3 AWS Glue client.
        :param glue_service_role: An AWS Identity and Access Management (IAM) role
                                  that AWS Glue can assume to gain access to the
                                  resources it requires.
        :param glue_bucket: An S3 bucket that can hold a job script and output data
                            from AWS Glue job runs.
        """
        self.glue_client = glue_client
        self.glue_service_role = glue_service_role
        self.glue_bucket = glue_bucket

    @staticmethod
    def wait(seconds, tick=12):
        """
        Waits for a specified number of seconds, while also displaying an animated
        spinner.

        :param seconds: The number of seconds to wait.
        :param tick: The number of frames per second used to animate the spinner.
        """
        progress = "|/-\\"
        waited = 0
        while waited < seconds:
            for frame in range(tick):
                sys.stdout.write(f"\r{progress[frame % len(progress)]}")
                sys.stdout.flush()
                time.sleep(1 / tick)
            waited += 1

    def upload_job_script(self, job_script):
        """
        Uploads a Python ETL script to an S3 bucket. The script is used by the AWS Glue
        job to transform data.

        :param job_script: The relative path to the job script.
        """
        try:
            self.glue_bucket.upload_file(Filename=job_script, Key=job_script)
            print(f"Uploaded job script '{job_script}' to the example bucket.")
        except S3UploadFailedError as err:
            logger.error("Couldn't upload job script. Here's why: %s", err)
            raise

    def run(self, crawler_name, db_name, db_prefix, data_source, job_script, job_name):
        """
        Runs the scenario. This is an interactive experience that runs at a command
        prompt and asks you for input throughout.

        :param crawler_name: The name of the crawler used in the scenario. If the
                             crawler does not exist, it is created.
        :param db_name: The name to give the metadata database created by the crawler.
        :param db_prefix: The prefix to give tables added to the database by the
                          crawler.
        :param data_source: The location of the data source that is targeted by the
                            crawler and extracted during job runs.
        :param job_script: The job script that is used to transform data during job
                           runs.
        :param job_name: The name to give the job definition that is created during the
                         scenario.
        """
        wrapper = GlueWrapper(self.glue_client)
        print(f"Checking for crawler {crawler_name}.")
        crawler = wrapper.get_crawler(crawler_name)
        if crawler is None:
            print(f"Creating crawler {crawler_name}.")
            wrapper.create_crawler(
                crawler_name,
                self.glue_service_role.arn,
                db_name,
                db_prefix,
                data_source,
            )
            print(f"Created crawler {crawler_name}.")
            crawler = wrapper.get_crawler(crawler_name)
        pprint(crawler)
        print("-" * 88)

        print(
            f"When you run the crawler, it crawls data stored in {data_source} and "
            f"creates a metadata database in the AWS Glue Data Catalog that describes "
            f"the data in the data source."
        )
        print("In this example, the source data is in CSV format.")
        ready = False
        while not ready:
            ready = Question.ask_question(
                "Ready to start the crawler? (y/n) ", Question.is_yesno
            )
        wrapper.start_crawler(crawler_name)
        print("Let's wait for the crawler to run. This typically takes a few minutes.")
        crawler_state = None
        while crawler_state != "READY":
            self.wait(10)
            crawler = wrapper.get_crawler(crawler_name)
            crawler_state = crawler["State"]
            print(f"Crawler is {crawler['State']}.")
        print("-" * 88)

        database = wrapper.get_database(db_name)
        print(f"The crawler created database {db_name}:")
        pprint(database)
        print(f"The database contains these tables:")
        tables = wrapper.get_tables(db_name)
        for index, table in enumerate(tables):
            print(f"\t{index + 1}. {table['Name']}")
        table_index = Question.ask_question(
            f"Enter the number of a table to see more detail: ",
            Question.is_int,
            Question.in_range(1, len(tables)),
        )
        pprint(tables[table_index - 1])
        print("-" * 88)

        print(f"Creating job definition {job_name}.")
        wrapper.create_job(
            job_name,
            "Getting started example job.",
            self.glue_service_role.arn,
            f"s3://{self.glue_bucket.name}/{job_script}",
        )
        print("Created job definition.")
        print(
            f"When you run the job, it extracts data from {data_source}, transforms it "
            f"by using the {job_script} script, and loads the output into "
            f"S3 bucket {self.glue_bucket.name}."
        )
        print(
            "In this example, the data is transformed from CSV to JSON, and only a few "
            "fields are included in the output."
        )
        job_run_status = None
        if Question.ask_question(f"Ready to run? (y/n) ", Question.is_yesno):
            job_run_id = wrapper.start_job_run(
                job_name, db_name, tables[0]["Name"], self.glue_bucket.name
            )
            print(f"Job {job_name} started. Let's wait for it to run.")
            while job_run_status not in ["SUCCEEDED", "STOPPED", "FAILED", "TIMEOUT"]:
                self.wait(10)
                job_run = wrapper.get_job_run(job_name, job_run_id)
                job_run_status = job_run["JobRunState"]
                print(f"Job {job_name}/{job_run_id} is {job_run_status}.")
        print("-" * 88)

        if job_run_status == "SUCCEEDED":
            print(
                f"Data from your job run is stored in your S3 bucket '{self.glue_bucket.name}':"
            )
            try:
                keys = [
                    obj.key for obj in self.glue_bucket.objects.filter(Prefix="run-")
                ]
                for index, key in enumerate(keys):
                    print(f"\t{index + 1}: {key}")
                lines = 4
                key_index = Question.ask_question(
                    f"Enter the number of a block to download it and see the first {lines} "
                    f"lines of JSON output in the block: ",
                    Question.is_int,
                    Question.in_range(1, len(keys)),
                )
                job_data = io.BytesIO()
                self.glue_bucket.download_fileobj(keys[key_index - 1], job_data)
                job_data.seek(0)
                for _ in range(lines):
                    print(job_data.readline().decode("utf-8"))
            except ClientError as err:
                logger.error(
                    "Couldn't get job run data. Here's why: %s: %s",
                    err.response["Error"]["Code"],
                    err.response["Error"]["Message"],
                )
                raise
            print("-" * 88)

        job_names = wrapper.list_jobs()
        if job_names:
            print(f"Your account has {len(job_names)} jobs defined:")
            for index, job_name in enumerate(job_names):
                print(f"\t{index + 1}. {job_name}")
            job_index = Question.ask_question(
                f"Enter a number between 1 and {len(job_names)} to see the list of runs for "
                f"a job: ",
                Question.is_int,
                Question.in_range(1, len(job_names)),
            )
            job_runs = wrapper.get_job_runs(job_names[job_index - 1])
            if job_runs:
                print(f"Found {len(job_runs)} runs for job {job_names[job_index - 1]}:")
                for index, job_run in enumerate(job_runs):
                    print(
                        f"\t{index + 1}. {job_run['JobRunState']} on "
                        f"{job_run['CompletedOn']:%Y-%m-%d %H:%M:%S}"
                    )
                run_index = Question.ask_question(
                    f"Enter a number between 1 and {len(job_runs)} to see details for a run: ",
                    Question.is_int,
                    Question.in_range(1, len(job_runs)),
                )
                pprint(job_runs[run_index - 1])
            else:
                print(f"No runs found for job {job_names[job_index - 1]}")
        else:
            print("Your account doesn't have any jobs defined.")
        print("-" * 88)

        print(
            f"Let's clean up. During this example we created job definition '{job_name}'."
        )
        if Question.ask_question(
            "Do you want to delete the definition and all runs? (y/n) ",
            Question.is_yesno,
        ):
            wrapper.delete_job(job_name)
            print(f"Job definition '{job_name}' deleted.")
        tables = wrapper.get_tables(db_name)
        print(f"We also created database '{db_name}' that contains these tables:")
        for table in tables:
            print(f"\t{table['Name']}")
        if Question.ask_question(
            "Do you want to delete the tables and the database? (y/n) ",
            Question.is_yesno,
        ):
            for table in tables:
                wrapper.delete_table(db_name, table["Name"])
                print(f"Deleted table {table['Name']}.")
            wrapper.delete_database(db_name)
            print(f"Deleted database {db_name}.")
        print(f"We also created crawler '{crawler_name}'.")
        if Question.ask_question(
            "Do you want to delete the crawler? (y/n) ", Question.is_yesno
        ):
            wrapper.delete_crawler(crawler_name)
            print(f"Deleted crawler {crawler_name}.")
        print("-" * 88)


def parse_args(args):
    """
    Parse command line arguments.

    :param args: The command line arguments.
    :return: The parsed arguments.
    """
    parser = argparse.ArgumentParser(
        description="Runs the AWS Glue getting started with crawlers and jobs scenario. "
        "Before you run this scenario, set up scaffold resources by running "
        "'python scaffold.py deploy'."
    )
    parser.add_argument(
        "role_name",
        help="The name of an IAM role that AWS Glue can assume. This role must grant access "
        "to Amazon S3 and to the permissions granted by the AWSGlueServiceRole "
        "managed policy.",
    )
    parser.add_argument(
        "bucket_name",
        help="The name of an S3 bucket that AWS Glue can access to get the job script and "
        "put job results.",
    )
    parser.add_argument(
        "--job_script",
        default="flight_etl_job_script.py",
        help="The name of the job script file that is used in the scenario.",
    )
    return parser.parse_args(args)


def main():
    args = parse_args(sys.argv[1:])
    try:
        print("-" * 88)
        print(
            "Welcome to the AWS Glue getting started with crawlers and jobs scenario."
        )
        print("-" * 88)
        scenario = GlueCrawlerJobScenario(
            boto3.client("glue"),
            boto3.resource("iam").Role(args.role_name),
            boto3.resource("s3").Bucket(args.bucket_name),
        )
        scenario.upload_job_script(args.job_script)
        scenario.run(
            "doc-example-crawler",
            "doc-example-database",
            "doc-example-",
            "s3://crawler-public-us-east-1/flight/2016/csv",
            args.job_script,
            "doc-example-job",
        )
        print("-" * 88)
        print(
            "To destroy scaffold resources, including the IAM role and S3 bucket "
            "used in this scenario, run 'python scaffold.py destroy'."
        )
        print("\nThanks for watching!")
        print("-" * 88)
    except Exception:
        logging.exception("Something went wrong with the example.")

에서 작업 실행 중에 데이터를 추출, 변환 및 로드하는 AWS Glue 데 사용되는 ETL 스크립트를 생성합니다.

import sys
from awsglue.transforms import *
from awsglue.utils import getResolvedOptions
from pyspark.context import SparkContext
from awsglue.context import GlueContext
from awsglue.job import Job

"""
These custom arguments must be passed as Arguments to the StartJobRun request.
    --input_database    The name of a metadata database that is contained in your 
                        AWS Glue Data Catalog and that contains tables that describe 
                        the data to be processed.
    --input_table       The name of a table in the database that describes the data to
                        be processed.
    --output_bucket_url An S3 bucket that receives the transformed output data.  
"""
args = getResolvedOptions(
    sys.argv, ["JOB_NAME", "input_database", "input_table", "output_bucket_url"]
)
sc = SparkContext()
glueContext = GlueContext(sc)
spark = glueContext.spark_session
job = Job(glueContext)
job.init(args["JOB_NAME"], args)

# Script generated for node S3 Flight Data.
S3FlightData_node1 = glueContext.create_dynamic_frame.from_catalog(
    database=args["input_database"],
    table_name=args["input_table"],
    transformation_ctx="S3FlightData_node1",
)

# This mapping performs two main functions:
# 1. It simplifies the output by removing most of the fields from the data.
# 2. It renames some fields. For example, `fl_date` is renamed to `flight_date`.
ApplyMapping_node2 = ApplyMapping.apply(
    frame=S3FlightData_node1,
    mappings=[
        ("year", "long", "year", "long"),
        ("month", "long", "month", "tinyint"),
        ("day_of_month", "long", "day", "tinyint"),
        ("fl_date", "string", "flight_date", "string"),
        ("carrier", "string", "carrier", "string"),
        ("fl_num", "long", "flight_num", "long"),
        ("origin_city_name", "string", "origin_city_name", "string"),
        ("origin_state_abr", "string", "origin_state_abr", "string"),
        ("dest_city_name", "string", "dest_city_name", "string"),
        ("dest_state_abr", "string", "dest_state_abr", "string"),
        ("dep_time", "long", "departure_time", "long"),
        ("wheels_off", "long", "wheels_off", "long"),
        ("wheels_on", "long", "wheels_on", "long"),
        ("arr_time", "long", "arrival_time", "long"),
        ("mon", "string", "mon", "string"),
    ],
    transformation_ctx="ApplyMapping_node2",
)

# Script generated for node Revised Flight Data.
RevisedFlightData_node3 = glueContext.write_dynamic_frame.from_options(
    frame=ApplyMapping_node2,
    connection_type="s3",
    format="json",
    connection_options={"path": args["output_bucket_url"], "partitionKeys": []},
    transformation_ctx="RevisedFlightData_node3",
)

job.commit()

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def create_crawler(self, name, role_arn, db_name, db_prefix, s3_target):
        """
        Creates a crawler that can crawl the specified target and populate a
        database in your AWS Glue Data Catalog with metadata that describes the data
        in the target.

        :param name: The name of the crawler.
        :param role_arn: The Amazon Resource Name (ARN) of an AWS Identity and Access
                         Management (IAM) role that grants permission to let AWS Glue
                         access the resources it needs.
        :param db_name: The name to give the database that is created by the crawler.
        :param db_prefix: The prefix to give any database tables that are created by
                          the crawler.
        :param s3_target: The URL to an S3 bucket that contains data that is
                          the target of the crawler.
        """
        try:
            self.glue_client.create_crawler(
                Name=name,
                Role=role_arn,
                DatabaseName=db_name,
                TablePrefix=db_prefix,
                Targets={"S3Targets": [{"Path": s3_target}]},
            )
        except ClientError as err:
            logger.error(
                "Couldn't create crawler. Here's why: %s: %s",
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def create_job(self, name, description, role_arn, script_location):
        """
        Creates a job definition for an extract, transform, and load (ETL) job that can
        be run by AWS Glue.

        :param name: The name of the job definition.
        :param description: The description of the job definition.
        :param role_arn: The ARN of an IAM role that grants AWS Glue the permissions
                         it requires to run the job.
        :param script_location: The Amazon S3 URL of a Python ETL script that is run as
                                part of the job. The script defines how the data is
                                transformed.
        """
        try:
            self.glue_client.create_job(
                Name=name,
                Description=description,
                Role=role_arn,
                Command={
                    "Name": "glueetl",
                    "ScriptLocation": script_location,
                    "PythonVersion": "3",
                },
                GlueVersion="3.0",
            )
        except ClientError as err:
            logger.error(
                "Couldn't create job %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def delete_crawler(self, name):
        """
        Deletes a crawler.

        :param name: The name of the crawler to delete.
        """
        try:
            self.glue_client.delete_crawler(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't delete crawler %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def delete_database(self, name):
        """
        Deletes a metadata database from your Data Catalog.

        :param name: The name of the database to delete.
        """
        try:
            self.glue_client.delete_database(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't delete database %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def delete_job(self, job_name):
        """
        Deletes a job definition. This also deletes data about all runs that are
        associated with this job definition.

        :param job_name: The name of the job definition to delete.
        """
        try:
            self.glue_client.delete_job(JobName=job_name)
        except ClientError as err:
            logger.error(
                "Couldn't delete job %s. Here's why: %s: %s",
                job_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def delete_table(self, db_name, table_name):
        """
        Deletes a table from a metadata database.

        :param db_name: The name of the database that contains the table.
        :param table_name: The name of the table to delete.
        """
        try:
            self.glue_client.delete_table(DatabaseName=db_name, Name=table_name)
        except ClientError as err:
            logger.error(
                "Couldn't delete table %s. Here's why: %s: %s",
                table_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_crawler(self, name):
        """
        Gets information about a crawler.

        :param name: The name of the crawler to look up.
        :return: Data about the crawler.
        """
        crawler = None
        try:
            response = self.glue_client.get_crawler(Name=name)
            crawler = response["Crawler"]
        except ClientError as err:
            if err.response["Error"]["Code"] == "EntityNotFoundException":
                logger.info("Crawler %s doesn't exist.", name)
            else:
                logger.error(
                    "Couldn't get crawler %s. Here's why: %s: %s",
                    name,
                    err.response["Error"]["Code"],
                    err.response["Error"]["Message"],
                )
                raise
        return crawler

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_database(self, name):
        """
        Gets information about a database in your Data Catalog.

        :param name: The name of the database to look up.
        :return: Information about the database.
        """
        try:
            response = self.glue_client.get_database(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't get database %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["Database"]

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_job_run(self, name, run_id):
        """
        Gets information about a single job run.

        :param name: The name of the job definition for the run.
        :param run_id: The ID of the run.
        :return: Information about the run.
        """
        try:
            response = self.glue_client.get_job_run(JobName=name, RunId=run_id)
        except ClientError as err:
            logger.error(
                "Couldn't get job run %s/%s. Here's why: %s: %s",
                name,
                run_id,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRun"]

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_job_runs(self, job_name):
        """
        Gets information about runs that have been performed for a specific job
        definition.

        :param job_name: The name of the job definition to look up.
        :return: The list of job runs.
        """
        try:
            response = self.glue_client.get_job_runs(JobName=job_name)
        except ClientError as err:
            logger.error(
                "Couldn't get job runs for %s. Here's why: %s: %s",
                job_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRuns"]

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def get_tables(self, db_name):
        """
        Gets a list of tables in a Data Catalog database.

        :param db_name: The name of the database to query.
        :return: The list of tables in the database.
        """
        try:
            response = self.glue_client.get_tables(DatabaseName=db_name)
        except ClientError as err:
            logger.error(
                "Couldn't get tables %s. Here's why: %s: %s",
                db_name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["TableList"]

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def list_jobs(self):
        """
        Lists the names of job definitions in your account.

        :return: The list of job definition names.
        """
        try:
            response = self.glue_client.list_jobs()
        except ClientError as err:
            logger.error(
                "Couldn't list jobs. Here's why: %s: %s",
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobNames"]

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def start_crawler(self, name):
        """
        Starts a crawler. The crawler crawls its configured target and creates
        metadata that describes the data it finds in the target data source.

        :param name: The name of the crawler to start.
        """
        try:
            self.glue_client.start_crawler(Name=name)
        except ClientError as err:
            logger.error(
                "Couldn't start crawler %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def start_job_run(self, name, input_database, input_table, output_bucket_name):
        """
        Starts a job run. A job run extracts data from the source, transforms it,
        and loads it to the output bucket.

        :param name: The name of the job definition.
        :param input_database: The name of the metadata database that contains tables
                               that describe the source data. This is typically created
                               by a crawler.
        :param input_table: The name of the table in the metadata database that
                            describes the source data.
        :param output_bucket_name: The S3 bucket where the output is written.
        :return: The ID of the job run.
        """
        try:
            # The custom Arguments that are passed to this function are used by the
            # Python ETL script to determine the location of input and output data.
            response = self.glue_client.start_job_run(
                JobName=name,
                Arguments={
                    "--input_database": input_database,
                    "--input_table": input_table,
                    "--output_bucket_url": f"s3://{output_bucket_name}/",
                },
            )
        except ClientError as err:
            logger.error(
                "Couldn't start job run %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRunId"]