GuzzleHttp\Pool example: identifying responses to concurrent async requests

guzzle_pool_example.php

<?php

/*
*  Using a key => value pair with the yield keyword is
*  the cleanest method I could find to add identifiers or tags
*  to asynchronous concurrent requests in Guzzle,
*  so you can identify which response is from which request!
*/

$client = new GuzzleHttp\Client(['base_uri' => 'http://httpbin.org']);

$requestGenerator = function($searchTerms) use ($client) {
	foreach($searchTerms as $searchTerm) {
		// The magic happens here, with yield key => value
		yield $searchTerm => function() use ($client, $searchTerm) {
			// Our identifier does not have to be included in the request URI or headers
			return $client->getAsync('/get?q='.$searchTerm, ['headers' => ['X-Search-Term' => $searchTerm]]);
		};
	}
};

$searchTerms = ['apple','banana','pear','orange','melon','grape','raisin'];

$pool = new GuzzleHttp\Pool($client, $requestGenerator($searchTerms), [
	'concurrency' => 3,
	'fulfilled' => function(GuzzleHttp\Psr7\Response $response, $index) {
		// This callback is delivered each successful response
		// $index will be our special identifier we set when generating the request
		$json = json_decode((string)$response->getBody());

		// If these values don't match, something is very wrong
		echo "Requested search term: ", $index, "\n";
		echo "Parsed from response: ", $json->headers->{'X-Search-Term'}, "\n\n";
	},
	'rejected' => function(Exception $reason, $index) {
		// This callback is delivered each failed request
		echo "Requested search term: ", $index, "\n";
		echo $reason->getMessage(), "\n\n";
	},
]);

// Initiate the transfers and create a promise
$promise = $pool->promise();
// Force the pool of requests to complete
$promise->wait();