AWS Glue ejemplos que utilizan SDK para PHP

Los siguientes ejemplos de código muestran cómo realizar acciones e implementar escenarios comunes mediante el uso del AWS SDK for PHP with AWS Glue.

Los conceptos básicos son ejemplos de código que muestran cómo realizar las operaciones esenciales dentro de un servicio.

Las acciones son extractos de código de programas más grandes y deben ejecutarse en contexto. Mientras las acciones muestran cómo llamar a las funciones de servicio individuales, es posible ver las acciones en contexto en los escenarios relacionados.

Cada ejemplo incluye un enlace al código fuente completo, donde puede encontrar instrucciones sobre cómo configurar y ejecutar el código en su contexto.

Conceptos básicos

En el siguiente ejemplo de código, se muestra cómo:

Cree un rastreador que rastree un bucket público de Amazon S3 y genere una base de datos de CSV metadatos con formato.
Enumere información sobre bases de datos y tablas en su. AWS Glue Data Catalog
Cree un trabajo para extraer CSV datos del depósito de S3, transformarlos y cargar la salida JSON con formato en otro depósito de S3.
Incluir información sobre las ejecuciones de trabajos, ver algunos de los datos transformados y limpiar los recursos.

Para obtener más información, consulte el tutorial: Primeros pasos con AWS Glue Studio.

SDK para PHP

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.


namespace Glue;

use Aws\Glue\GlueClient;
use Aws\S3\S3Client;
use AwsUtilities\AWSServiceClass;
use GuzzleHttp\Psr7\Stream;
use Iam\IAMService;

class GettingStartedWithGlue
{
    public function run()
    {
        echo("\n");
        echo("--------------------------------------\n");
        print("Welcome to the AWS Glue getting started demo using PHP!\n");
        echo("--------------------------------------\n");

        $clientArgs = [
            'region' => 'us-west-2',
            'version' => 'latest',
            'profile' => 'default',
        ];
        $uniqid = uniqid();

        $glueClient = new GlueClient($clientArgs);
        $glueService = new GlueService($glueClient);
        $iamService = new IAMService();
        $crawlerName = "example-crawler-test-" . $uniqid;

        AWSServiceClass::$waitTime = 5;
        AWSServiceClass::$maxWaitAttempts = 20;

        $role = $iamService->getRole("AWSGlueServiceRole-DocExample");

        $databaseName = "doc-example-database-$uniqid";
        $path = 's3://crawler-public-us-east-1/flight/2016/csv';
        $glueService->createCrawler($crawlerName, $role['Role']['Arn'], $databaseName, $path);
        $glueService->startCrawler($crawlerName);

        echo "Waiting for crawler";
        do {
            $crawler = $glueService->getCrawler($crawlerName);
            echo ".";
            sleep(10);
        } while ($crawler['Crawler']['State'] != "READY");
        echo "\n";

        $database = $glueService->getDatabase($databaseName);
        echo "Found a database named " . $database['Database']['Name'] . "\n";

        //Upload job script
        $s3client = new S3Client($clientArgs);
        $bucketName = "test-glue-bucket-" . $uniqid;
        $s3client->createBucket([
            'Bucket' => $bucketName,
            'CreateBucketConfiguration' => ['LocationConstraint' => 'us-west-2'],
        ]);

        $s3client->putObject([
            'Bucket' => $bucketName,
            'Key' => 'run_job.py',
            'SourceFile' => __DIR__ . '/flight_etl_job_script.py'
        ]);
        $s3client->putObject([
            'Bucket' => $bucketName,
            'Key' => 'setup_scenario_getting_started.yaml',
            'SourceFile' => __DIR__ . '/setup_scenario_getting_started.yaml'
        ]);

        $tables = $glueService->getTables($databaseName);

        $jobName = 'test-job-' . $uniqid;
        $scriptLocation = "s3://$bucketName/run_job.py";
        $job = $glueService->createJob($jobName, $role['Role']['Arn'], $scriptLocation);

        $outputBucketUrl = "s3://$bucketName";
        $runId = $glueService->startJobRun($jobName, $databaseName, $tables, $outputBucketUrl)['JobRunId'];

        echo "waiting for job";
        do {
            $jobRun = $glueService->getJobRun($jobName, $runId);
            echo ".";
            sleep(10);
        } while (!array_intersect([$jobRun['JobRun']['JobRunState']], ['SUCCEEDED', 'STOPPED', 'FAILED', 'TIMEOUT']));
        echo "\n";

        $jobRuns = $glueService->getJobRuns($jobName);

        $objects = $s3client->listObjects([
            'Bucket' => $bucketName,
        ])['Contents'];

        foreach ($objects as $object) {
            echo $object['Key'] . "\n";
        }

        echo "Downloading " . $objects[1]['Key'] . "\n";
        /** @var Stream $downloadObject */
        $downloadObject = $s3client->getObject([
            'Bucket' => $bucketName,
            'Key' => $objects[1]['Key'],
        ])['Body']->getContents();
        echo "Here is the first 1000 characters in the object.";
        echo substr($downloadObject, 0, 1000);

        $jobs = $glueService->listJobs();
        echo "Current jobs:\n";
        foreach ($jobs['JobNames'] as $jobsName) {
            echo "{$jobsName}\n";
        }

        echo "Delete the job.\n";
        $glueClient->deleteJob([
            'JobName' => $job['Name'],
        ]);

        echo "Delete the tables.\n";
        foreach ($tables['TableList'] as $table) {
            $glueService->deleteTable($table['Name'], $databaseName);
        }

        echo "Delete the databases.\n";
        $glueClient->deleteDatabase([
            'Name' => $databaseName,
        ]);

        echo "Delete the crawler.\n";
        $glueClient->deleteCrawler([
            'Name' => $crawlerName,
        ]);

        $deleteObjects = $s3client->listObjectsV2([
            'Bucket' => $bucketName,
        ]);
        echo "Delete all objects in the bucket.\n";
        $deleteObjects = $s3client->deleteObjects([
            'Bucket' => $bucketName,
            'Delete' => [
                'Objects' => $deleteObjects['Contents'],
            ]
        ]);
        echo "Delete the bucket.\n";
        $s3client->deleteBucket(['Bucket' => $bucketName]);

        echo "This job was brought to you by the number $uniqid\n";
    }
}

namespace Glue;

use Aws\Glue\GlueClient;
use Aws\Result;

use function PHPUnit\Framework\isEmpty;

class GlueService extends \AwsUtilities\AWSServiceClass
{
    protected GlueClient $glueClient;

    public function __construct($glueClient)
    {
        $this->glueClient = $glueClient;
    }

    public function getCrawler($crawlerName)
    {
        return $this->customWaiter(function () use ($crawlerName) {
            return $this->glueClient->getCrawler([
                'Name' => $crawlerName,
            ]);
        });
    }

    public function createCrawler($crawlerName, $role, $databaseName, $path): Result
    {
        return $this->customWaiter(function () use ($crawlerName, $role, $databaseName, $path) {
            return $this->glueClient->createCrawler([
                'Name' => $crawlerName,
                'Role' => $role,
                'DatabaseName' => $databaseName,
                'Targets' => [
                    'S3Targets' =>
                        [[
                            'Path' => $path,
                        ]]
                ],
            ]);
        });
    }

    public function startCrawler($crawlerName): Result
    {
        return $this->glueClient->startCrawler([
            'Name' => $crawlerName,
        ]);
    }

    public function getDatabase(string $databaseName): Result
    {
        return $this->customWaiter(function () use ($databaseName) {
            return $this->glueClient->getDatabase([
                'Name' => $databaseName,
            ]);
        });
    }

    public function getTables($databaseName): Result
    {
        return $this->glueClient->getTables([
            'DatabaseName' => $databaseName,
        ]);
    }

    public function createJob($jobName, $role, $scriptLocation, $pythonVersion = '3', $glueVersion = '3.0'): Result
    {
        return $this->glueClient->createJob([
            'Name' => $jobName,
            'Role' => $role,
            'Command' => [
                'Name' => 'glueetl',
                'ScriptLocation' => $scriptLocation,
                'PythonVersion' => $pythonVersion,
            ],
            'GlueVersion' => $glueVersion,
        ]);
    }

    public function startJobRun($jobName, $databaseName, $tables, $outputBucketUrl): Result
    {
        return $this->glueClient->startJobRun([
            'JobName' => $jobName,
            'Arguments' => [
                'input_database' => $databaseName,
                'input_table' => $tables['TableList'][0]['Name'],
                'output_bucket_url' => $outputBucketUrl,
                '--input_database' => $databaseName,
                '--input_table' => $tables['TableList'][0]['Name'],
                '--output_bucket_url' => $outputBucketUrl,
            ],
        ]);
    }

    public function listJobs($maxResults = null, $nextToken = null, $tags = []): Result
    {
        $arguments = [];
        if ($maxResults) {
            $arguments['MaxResults'] = $maxResults;
        }
        if ($nextToken) {
            $arguments['NextToken'] = $nextToken;
        }
        if (!empty($tags)) {
            $arguments['Tags'] = $tags;
        }
        return $this->glueClient->listJobs($arguments);
    }

    public function getJobRuns($jobName, $maxResults = 0, $nextToken = ''): Result
    {
        $arguments = ['JobName' => $jobName];
        if ($maxResults) {
            $arguments['MaxResults'] = $maxResults;
        }
        if ($nextToken) {
            $arguments['NextToken'] = $nextToken;
        }
        return $this->glueClient->getJobRuns($arguments);
    }

    public function getJobRun($jobName, $runId, $predecessorsIncluded = false): Result
    {
        return $this->glueClient->getJobRun([
            'JobName' => $jobName,
            'RunId' => $runId,
            'PredecessorsIncluded' => $predecessorsIncluded,
        ]);
    }

    public function deleteJob($jobName)
    {
        return $this->glueClient->deleteJob([
            'JobName' => $jobName,
        ]);
    }

    public function deleteTable($tableName, $databaseName)
    {
        return $this->glueClient->deleteTable([
            'DatabaseName' => $databaseName,
            'Name' => $tableName,
        ]);
    }

    public function deleteDatabase($databaseName)
    {
        return $this->glueClient->deleteDatabase([
            'Name' => $databaseName,
        ]);
    }

    public function deleteCrawler($crawlerName)
    {
        return $this->glueClient->deleteCrawler([
            'Name' => $crawlerName,
        ]);
    }
}

Para API obtener más información, consulte los siguientes temas en AWS SDK for PHP APIReference.