Non-blocking HTTP Client in PHP

Learn how to do standards-compliant HTTP resource traversal and RESTful web service consumption.

Suggest edit

AMPHP is a collection of event-driven libraries for PHP designed with fibers and concurrency in mind. This package provides an asynchronous HTTP client for PHP based on Revolt. Its API simplifies standards-compliant HTTP resource traversal and RESTful web service consumption without obscuring the underlying protocol. The library manually implements HTTP over TCP sockets; as such it has no dependency on ext/curl.

Features

Installation

This package can be installed as a Composer dependency.

composer require amphp/http-client

Additionally, you might want to install the nghttp2 library to take advantage of FFI to speed up and reduce the memory usage.

Usage

The main interaction point with this library is the HttpClient class. HttpClient instances can be built using HttpClientBuilder without knowing about the existing implementations.

HttpClientBuilder allows to register two kinds of interceptors, which allows customizing the HttpClient behavior in a composable fashion.

In its simplest form, the HTTP client takes a request with a URL as string and interprets that as a GET request to that resource without any custom headers. Standard headers like Accept, Connection or Host will automatically be added if not present.

use Amp\Http\Client\HttpClientBuilder;

$client = HttpClientBuilder::buildDefault();

$response = $client->request(new Request("https://httpbin.org/get"));

var_dump($response->getStatus());
var_dump($response->getHeaders());
var_dump($response->getBody()->buffer());

Request

The HttpClient requires a Request being passed as first argument to request(). The Request class can be used to specify further specifics of the request such as setting headers or changing the request method.

Request objects are mutable (instead of immutable as in amphp/artax / PSR-7).

Cloning Request objects will result in a deep clone, but doing so is usually only required if requests are retried or cloned for sub-requests.

Request URI

The constructor requires an absolute request URI. Request::setUri(string $uri) allows changing the request URI.

$request = new Request("https://httpbin.org/post", "POST");
$request->setBody("foobar");
$request->setUri("https://google.com/");

Request::getUri() exposes the request URI of the given Request object.

Request Method

The constructor accepts an optional request method, it defaults to GET. Request::setMethod(string $method) allows changing the request method.

$request = new Request("https://httpbin.org/post", "POST");
$request->setBody("foobar");
$request->setMethod("PUT");

Request::getMethod() exposes the request method of the given Request object.

Request Headers

Request::setHeader(string $field, string $value) allows changing the request headers. It will remove any previous values for that field. Request::addHeader(string $field, string $value) allows adding an additional header line without removing existing lines.

Request::setHeaders(array $headers) allows adding multiple headers at once with the array keys being the field names and the values being the header values. The header values can also be arrays of strings to set multiple header lines.

Request::hasHeader(string $field) checks whether at least one header line with the given name exists.

Request::getHeader(string $field) returns the first header line with the given name or null if no such header exists.

Request::getHeaderArray(string $field) returns an array of header lines with the given name. An empty array is returned if no header with the given name exists.

Request::getHeaders() returns an associative array with the keys being the header names and the values being arrays of header lines.

$request = new Request("https://httpbin.org/post", "POST");
$request->setHeader("X-Foobar", "Hello World");
$request->setBody("foobar");

Request Bodies

Request::setBody($body) allows changing the request body. Accepted types are string, null, and HttpContent. string and null are automatically converted to an instance of HttpContent.

HttpContent is basically a factory for request bodies. We cannot simply accept streams here, because a request body might have to be sent again on a redirect / retry.

$request = new Request("https://httpbin.org/post", "POST");
$request->setBody("foobar");

Request::getBody() exposes the request body of the given Request object and will always return a HttpContent.

Response

HttpClient::request() returns a Response as soon as the response headers are successfully received.

Response objects are mutable (instead of immutable as in Artax v3 / PSR-7)

Response Status

You can retrieve the response’s HTTP status using getStatus(). It returns the status as an integer. The optional (and possibly empty) reason associated with the status can be retrieved using getReason().

$response = $client->request($request);

var_dump($response->getStatus(), $response->getReason());

Response Protocol Version

You can retrieve the response’s HTTP protocol version using getProtocolVersion().

$response = $client->request($request);

var_dump($response->getProtocolVersion());

Response Headers

Response headers can be accessed by a set of methods.

  • hasHeader(string) returns whether a given header is present.
  • getHeader(string) returns the first header with the given name or null if no such header is present.
  • getHeaderArray(string) returns all headers with the given name, possibly an empty array.
  • getHeaders() returns all headers as an associative array, see below.

getHeaders() Format

[
    "header-1" => [
        "value-1",
        "value-2",
    ],
    "header-2" => [
        "value-1",
    ],
]

Response Body

getBody() returns a Payload, which allows simple buffering and streaming access.

$chunk = $response->getBody()->read(); reads only a single chunk from the body while $contents = $response->getBody()->buffer() buffers the complete body. Please refer to the Payload documentation for more information.

Request, Original Request and Previous Response

getRequest() allows access to the request corresponding to the response. This might not be the original request in case of redirects. getOriginalRequest() returns the original request sent by the client. This might not be the same request that was passed to Client::request(), because the client might normalize headers or assign cookies. getPreviousResponse allows access to previous responses in case of redirects, but the response bodies of these responses won’t be available, as they’re discarded. If you need access to these, you need to disable auto redirects and implement them yourself.

Interceptors

Interceptors allow customizing the HttpClient behavior in a composable fashion. Use cases range from adding / removing headers from a request / response and recording timing information to more advanced use cases like a fully compliant HTTP cache that intercepts requests and serves them from the cache if possible.

use Amp\Http\Client\Client;
use Amp\Http\Client\HttpClientBuilder;
use Amp\Http\Client\Interceptor\SetRequestHeader;
use Amp\Http\Client\Interceptor\SetResponseHeader;
use Amp\Http\Client\Request;

$client = (new HttpClientBuilder)
    ->intercept(new SetRequestHeader('x-foo', 'bar'))
    ->intercept(new SetResponseHeader('x-tea', 'now'))
    ->build();

$response = $client->request(new Request("https://httpbin.org/get"));
$body = $response->getBody()->buffer();

There are two kinds of interceptors with separate interfaces named ApplicationInterceptor and NetworkInterceptor.

Choosing the right interceptor

Most interceptors should be implemented as ApplicationInterceptor. However, there’s sometimes the need to have access to the underlying connection properties. In such a case, a NetworkInterceptor can be implemented to access the used IPs and TLS settings.

Another use-case for implementing a NetworkInterceptor is an interceptor, that should only ever run if the request is sent over the network instead of served from a cache or similar. However, that should usually be solved with the configuration order of the application interceptors.

The big disadvantage of network interceptors is that they have to be rather quick and can’t take too long, because they’re only invoked after the connection has been created and the client will run into a timeout if there’s no activity within a reasonable time.

List of Interceptors

  • AddRequestHeader
  • AddResponseHeader
  • ConditionalInterceptor
  • DecompressResponse
  • FollowRedirects
  • ForbidUriUserInfo
  • IfOrigin
  • ModifyRequest
  • ModifyResponse
  • RemoveRequestHeader
  • RemoveResponseHeader
  • RetryRequests
  • SetRequestHeader
  • SetRequestHeaderIfUnset
  • SetResponseHeader
  • SetResponseHeaderIfUnset
  • SetRequestTimeout
  • CookieHandler
  • PrivateCache

Redirects

If you use HttpClientBuilder, the resulting HttpClient will automatically follow up to ten redirects by default. Automatic following can be customized or disabled (using a limit of 0) using HttpClientBuilder::followRedirects().

Redirect Policy

The FollowRedirects interceptor will only follow redirects with a GET method. If another request method is used and a 307 or 308 response is received, the response will be returned as is, so another interceptor or the application can take care of it. Cross-origin redirects will be attempted without any headers set, so any application headers will be discarded. If HttpClientBuilder is used to configure the client, the FollowRedirects interceptor is the outermost interceptor, so any headers set by interceptors will still be present in the response. It is therefore recommended to set headers via interceptors instead of directly in the request.

Examining the Redirect Chain

All previous responses can be accessed from the resulting Response via Response::getPreviousResponse(). However, the response body is discarded on redirects, so it can no longer be consumed. If you want to consume redirect response bodies, you need to implement your own interceptor.

Cookies

See amphp/http-client-cookies.

Logging

The LogHttpArchive event listener allows logging all requests / responses including detailed timing information to an HTTP archive (HAR).

These log files can then be imported into the browsers developer tools or online tools like HTTP Archive Viewer or Google’s HAR Analyzer.

Be careful if your log files might contain sensitive information in URLs or headers if you submit these files to third parties like the linked services above.

use Amp\Http\Client\HttpClientBuilder;
use Amp\Http\Client\EventListener\LogHttpArchive;

$httpClient = (new HttpClientBuilder)
    ->listen(new LogHttpArchive('/tmp/http-client.har'))
    ->build();

$httpClient->request(...);

HAR Viewer Screenshot

Proxies

See amphp/http-tunnel.

Versioning

amphp/http-client follows the semver semantic versioning specification like all other amphp packages.

Everything in an Internal namespace or marked as @internal is not public API and therefore not covered by BC guarantees.

Security

If you discover any security related issues, please email [email protected] instead of using the issue tracker.

License

The MIT License (MIT). Please see LICENSE for more information.