Guzzle site authenticator

This package is a plugin for guzzle 5.x. It provides a subscriber that can authenticate requests by posting login information.

It comes up as a Symfony bundle and a generic php lib.

Installation

Using composer

Add the package to your requirements using composer: composer require bdunogier/guzzle-site-authenticator.

If you're using the Symfony fullstack, add BD\GuzzleSiteAuthenticatorBundle\BDGuzzleSiteAuthenticatorBundle to your kernel class.

Usage

The guzzle subscriber, Guzzle\AuthenticatorSubscriber, must be attached to the Guzzle client. It is provided by the bundle as @bd_guzzle_site_authenticator.authenticator_subscriber:

$client = new GuzzleHttp\Client(['defaults' => ['cookies' => new FileCookieJar('/tmp/cookiejar.json')]]);
$client->getEmitter()->attach(
    $container->get('bd_guzzle_site_authenticator.authenticator_subscriber')
);

Cookies handling

The CookieJar passed to the guzzle client defaults is important: it will be used read/write cookies received by Guzzle, and is required for authentication to work.

Send a request with Guzzle. If the request's host has a SiteConfig that requires configuration (see below), the plugin will try to log in to the site if it does not have a cookie yet. After a request, if the response contains the not logged in text (matched by xpath), it tries to login again, and retries the request.

Site configuration

$siteConfig = new BD\GuzzleSiteAuthenticator\SiteConfig\SiteConfig([
    'host' => 'example.com',
    'loginUri' => 'http://example.com/login',
    'usernameField' => 'username',
    'passwordField' => 'password',
    'extraFields' => ['action' => 'login'],
    'notLoggedInXpath' => "//div[@class='not-logged-in']",
    'username' => "johndoe",
    'password' => "unknown",
]);

SiteConfig objects are returned by a SiteConfigBuilder. The library comes with a default ArraySiteConfigBuilder, that accepts a list of site config properties array, indexed by host. With the bundle, its contents can be configured using the bd_guzzle_site_authenticator.site_config container variable:

# config.yml
parameters:
    bd_guzzle_site_authenticator.site_config:
        example.com:
            host: "example.com"
            loginUri: "http://example.com/login"
            usernameField: "username"
            passwordField: "password"
            extraFields: {action: login}
            notLoggedInXpath: "//div[@class='not-logged-in']"
            username: "johndoe"
            password: "unknown"
        otherexample.com:
            host: ...

Implementations

Used by wallabag, a read it later web application, to fetch content from sites that require a login.

It implements a custom SiteConfigBuilder, based on sites configuration provided by j0k3r/graby.