Inspired by tpyo/amazon-s3-php-class, this is a simple and configurable S3 PHP library. It was written to be as lightweight as possible, while still enabling access to all of the features of AWS (e.g. server-side encryption).

Additionally, curl_multi_exec is used (rather than curl_exec) for better performance when doing bulk operations.

$client = new S3($access_key, $secret_key [, $endpoint = null]);

• $client->useCurlOpts($curl_opts_array)

Provides the S3 class with any curl options to use in making requests.

The following options are passed by default in order to prevent 'hung' requests:

curl_opts = array(
    CURLOPT_CONNECTTIMEOUT => 30,
    CURLOPT_LOW_SPEED_LIMIT => 1,
    CURLOPT_LOW_SPEED_TIME => 30
);

Note: If you call this method, these defaults will not be used.

All of the available S3 operations take an optional $headers array that will be passed along to S3. These can include x-amz-meta-, x-amz-server-side-encryption, Content-Type, etc. Any Amazon headers specified will be properly included in the AWS signature as per AWS Signature v2.

Request headers that are common to all requests are located here.

All methods in the S3 class will return an instance of the S3Response class.

class S3Response {
    public $error;    // null if no error
    public $code;     // response code from AWS
    public $headers;  // response headers from AWS
    public $body;     // response body from AWS
}

If there is an error in curl or an error is returned from AWS, $response->error will be non-null and set to the following array:

$error = array(
    'code' => xxx, // error code from either curl or AWS
    'message' => xxx, // error string from either curl or AWS
    'resource' => [optional] // the S3 resource frmo the request
)

putObject($bucket, $path, $file [, $headers = array()])

• Uploads a file to the specified path and bucket. $file can either be the raw representation of a file (e.g. the result of file_get_contents()) or a valid stream resource.
• AWS Documentation

getObjectInfo($bucket, $path, [, $headers = array()])

• Retrieves metadata for the object.
• AWS Documentation

getObject($bucket, $path [, $resource = null [, $headers = array()]])

• Retrieves the contents of an object. If $resource is a valid stream resource, the contents will be written to the stream. Otherwise $response->body will contain the contents of the file.
• AWS Documentation

deleteObject($bucket, $path [, $headers = array()])

• Deletes an object from S3.
• AWS Documentation

getBucket($bucket [, $headers = array()])

• Returns a parsed response from S3 listing the contents of the specified bucket.
• AWS Documentation

Instantiating the S3 class:

$client = new S3(ACCESS_KEY, SECRET_KEY);

// [OPTIONAL] Specify different curl options
$client->useCurlOpts(array(
    CURLOPT_MAX_RECV_SPEED_LARGE => 1048576,
    CURLOPT_CONNECTTIMEOUT => 10
));

$response = $client->putObject(
    'bucket',
    'hello_world.txt',
    'hello world!',
    array(
        'Content-Type' => 'text/plain'
    )
);

print_r($response);

Output:

S3Response Object
(
    [error] => null
    [code] => 200
    [headers] => Array
        (
            [x-amz-id-2] => ...
            [x-amz-request-id] => ...
            [ETag] => "..."
            [Content-Length] => ...
            [Server] => ...
        )
    [body] => null
)

$resource = tmpfile();
$client->getObject('bucket', 'hello_world.txt', $resource);

print_r($response);
echo stream_get_contents($resource) . "\n";

Output:

S3Response Object
(
    [error] =>
    [code] => 200
    [headers] => Array
        (
            [x-amz-id-2] => ...
            [x-amz-request-id] => ...
            [ETag] => "..."
            [Accept-Ranges] => bytes
            [Content-Type] => text/plain
            [Content-Length] => 12
            [Server] => ...
        )

    [body] => Resource id #17
)

hello world!