Skip to main content

Datasette plugin that authenticates users based on existing domain cookies

Project description

datasette-auth-existing-cookies

PyPI CircleCI License

Datasette plugin that authenticates users based on existing domain cookies.

When to use this

This plugin allows you to build custom authentication for Datasette when you are hosting a Datasette instance on the same domain as another, authenticated website.

Consider a website on www.example.com which supports user authentication.

You could run Datasette on data.example.com in a way that lets it see cookies that were set for the .example.com domain.

Using this plugin, you could build an API endpoint at www.example.com/user-for-cookies which returns a JSON object representing the currently signed-in user, based on their cookies.

The plugin can protect any hits to any data.example.com pages by passing their cookies through to that API and seeing if the user should be logged in or not.

You can also use subclassing to decode existing cookies using some other mechanism.

Configuration

This plugin requires some configuration in the Datasette metadata.json file.

It needs to know the following:

  • Which domain cookies should it be paying attention to? If you are authenticating against Dango this is probably ["sessionid"].
  • What's an API it can send the incoming cookies to that will decipher them into some user information?
  • Where should it redirect the user if they need to sign in?

Example configuration setting all three of these values looks like this:

{
    "plugins": {
        "datasette-auth-existing-cookies": {
            "api_url": "http://www.example.com/user-from-cookies",
            "auth_redirect_url": "http://www.example.com/login",
            "original_cookies": ["sessionid"]
        }
    }
}

With this configuration the user's current sessionid cookie will be passed to the API URL (as a regular cookie header). That URL should then return either an empty JSON object if the user is not currently signed in:

{}

Or a JSON object representing the user if they ARE signed in:

{
    "id": 123,
    "username": "simonw"
}

This object can contain any keys that you like - the information will be stored in a new signed cookie and made available to Datasette code as the "auth" dictionary on the ASGI scope.

I suggest including at least an id and a username.

Templates

You probably want your user's to be able to see that they are signed in. The plugin makes the auth data from above directly available within Datasette's templates. You could use a custom base.html template (see template documentation) that looks like this:

{% extends "default:base.html" %}

{% block extra_head %}
<style type="text/css">
.hd .logout {
    float: right;
    text-align: right;
    padding-left: 1em;
}
</style>
{% endblock %}

{% block nav %}
    {{ super() }}
    {% if auth and auth.username %}
        <p class="logout">
            <strong>{{ auth.username }}</strong> &middot; <a href="https://www.example.com/logout">Log out</a>
        </p>
    {% endif %}
{% endblock %}

Other options

  • require_auth. This defaults to True. You can set it to False if you want unauthenticated users to be able to view the Datasette instance.
  • cookie_secret. You can use this to set the signing secret that will be used for the cookie set by this plugin (you should use secret configuration values for this). If you do not set a secret the plugin will create one on first run and store it in an appropriate state directory based on your operating system (the user_state_dir according to appdirs).
  • cookie_ttl. The plugin sets its own cookie to avoid hitting the backend API for every incoming request. By default it still hits the API at most every 10 seconds, in case the user has signed out on the main site. You can raise or lower the timeout using this setting.
  • next_secret. See below.

Login redirect mechanism

If the user does not have a valid authentication cookie they will be redirected to an existing login page.

This page is specified using the auth_redirect_url setting.

Given the above example configuration, the URL that the user should be sent to after they log in will be specified as the ?next= parameter to that page, for example:

http://www.example.com/login?next=http://foo.example.com/

It is up to you to program the login endpoint such that it is not vulnerable to an Unvalidated redirect vulnerability.

One way to do this is by verifying that the URL passed to ?next= is a URL that belongs to a trusted website. Django's own login view does this by verifying that the URL hostname is on an approved list.

Another way to do this is to use the next_secret configuration parameter to set a signing secret for that URL. This signing secret will be used to construct a ?next_sig= signed token using the Python itsdangerous module, like this:

?next_sig=Imh0dHBzOi8vZGVtby5leGFtcGxlLmNvbS9mb28vYmFyIg.7JdhRCoP7Ow1cRF1ZVengC-qk6c

You should use Datasette's secret configuration values mechanism to set this secret from an environment variable, like so:

{
    "plugins": {
        "datasette-auth-existing-cookies": {
            "api_url": "http://www.example.com/user-from-cookies",
            "auth_redirect_url": "http://www.example.com/login",
            "original_cookies": ["sessionid"],
            "next_secret":  {
                "$env": "NEXT_SECRET"
            }
        }
    }
}

You can verify this secret in Python code for your own login form like so:

from itsdangerous import URLSafeSerializer, BadSignature

def verify_next_sig(next_sig):
    signer = URLSafeSerializer(next_secret)
    try:
        decoded = signer.loads(next_sig)
        return True
    except BadSignature:
        return False

If you want to roll your own signing mechanism here you can do so by subclassing ExistingCookiesAuth and over-riding the build_auth_redirect(next_url) method.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page