UNIFIED DATA PROCESSING FRAMEWORK

{
    "name": "flow-php/examples",
    "description": "Flow PHP - Examples",
    "license": "MIT",
    "type": "library",
    "require": {
        "flow-php/etl": "1.x-dev",
        "symfony/dotenv": "^7.2",
        "flow-php/filesystem-azure-bridge": "1.x-dev",
        "flow-php/etl-adapter-parquet": "1.x-dev",
        "flow-php/filesystem": "1.x-dev",
        "flow-php/snappy": "1.x-dev",
        "flow-php/parquet": "1.x-dev",
        "nyholm/psr7": "^1.8",
        "php-http/curl-client": "^2.3"
    },
    "config": {
        "allow-plugins": {
            "php-http/discovery": false
        }
    }
}

<?php

declare(strict_types=1);

use function Flow\Azure\SDK\DSL\{azure_blob_service, azure_blob_service_config, azure_shared_key_authorization_factory};
use function Flow\ETL\Adapter\Parquet\{from_parquet, to_parquet};
use function Flow\ETL\DSL\{config_builder, data_frame, from_array, overwrite, to_stream};
use function Flow\Filesystem\Bridge\Azure\DSL\azure_filesystem;
use function Flow\Filesystem\DSL\path;
use Symfony\Component\Dotenv\Dotenv;

require __DIR__ . '/vendor/autoload.php';

if (!\file_exists(__DIR__ . '/.env')) {
    print 'Example skipped. Please create .env file with Azure Storage Account credentials.' . PHP_EOL;

    return;
}

$dotenv = new Dotenv();
$dotenv->load(__DIR__ . '/.env');

$config = config_builder()
    ->mount(
        azure_filesystem(
            azure_blob_service(
                azure_blob_service_config(
                    $_ENV['AZURE_ACCOUNT'],
                    $_ENV['AZURE_CONTAINER']
                ),
                azure_shared_key_authorization_factory(
                    $_ENV['AZURE_ACCOUNT'],
                    $_ENV['AZURE_ACCOUNT_KEY']
                ),
            )
        )
    );

data_frame($config)
    ->read(from_array([
        ['id' => 1, 'name' => 'test'],
        ['id' => 2, 'name' => 'test'],
        ['id' => 3, 'name' => 'test'],
        ['id' => 4, 'name' => 'test'],
    ]))
    ->saveMode(overwrite())
    ->write(to_parquet(path('azure-blob://test.parquet')))
    ->run();

data_frame($config)
    ->read(from_parquet(path('azure-blob://test.parquet')))
    ->write(to_stream(__DIR__ . '/output.txt', truncate: false))
    ->run();