How to access user_data json from OAuthenticator?

JupyterHub
,
1

I’m using the GenericOAuthenticator from GitHub - jupyterhub/oauthenticator: OAuth + JupyterHub Authenticator = OAuthenticator
and I’d like to be able to access the groups within the user response json data.
In my jupyterhub_config.py, I got

    c.JupyterHub.authenticator_class = 'oauthenticator.generic.GenericOAuthenticator'

and for the claim groups key, I put in

c.GenericOAuthenticator.claim_groups_key = lambda r: r.get('mygroups')

as I would want it to be Callable so I can access the underlying dict of the response.
The resulting dict would look something like this:

{
    'name': 'Foo Bar',
    'family_name': 'Bar',
    'given_name': 'Foo',
    'mygroups': [
        'group1',
        'group2',
        'group3'
    ]
}

If I try to do c.GenericOAuthenticator.claim_groups_key(), it will give me an error complaining that it needs a parameter ‘r’. But I don’t know which variable should be used for r. I’m not sure if I’m going about this the wrong way, but all I want to do is to be able to get the value of “mygroups”.
When JupyterHub spins up and reads in all the data from jupyterhub_config.py, it runs that lambda function and inserts the user_data dict for r. But how would I be able to access the user_data dict from within c.GenericOAuthenticator? I need to be able to parse it to be able to send it to somewhere else.

Thanks!

2

Basically what I’m trying to do is to have a custom KubeSpawner be able to read the groups that the claim groups key is referring to and pass that to the c.KubeSpawner.profile_list. The custom spawner is included inside of the values.yaml file within the Helm chart.

Any ideas?

3

c.GenericOAuthenticator.claim_groups_key() won’t work if called manually, it’s designed to be called by the GenericOAuthenticator code when it’s running:

As you can see user_data_resp_json is passed to your function.

4

Ok, so trying to access that directly is a non-starter. I wonder if I can use the pre_spawn_start() method to pass the auth_state to the spawner via the environment variable as described here: Authenticators — JupyterHub 1.4.2 documentation

5

If the information you need can be obtained by the authenticate method then yes, you can save it auth_state and retrieve it in the pre_spawn_start hook.

authenticate returns a user_info object with auth_state which includes user_data_resp_json:

so if the information you need is already in user_data_resp_json you shouldn’t need to override authenticate.

6

So it turns out that I didn’t need to override any of the Authenticator() logic.

I did have to enable my auth state in the OAuthenticator:

c.JupyterHub.authenticator_class = 'oauthenticator.generic.GenericOAuthenticator'
c.GenericOAuthenticator.enable_auth_state = True
os.environ['JUPYTERHUB_CRYPT_KEY'] = token_hex(32)

then had to create a userdata hook for the auth state as shown here:

if I do a self.log.debug(auth_state) inside the userdata hook method, then I can see the user_data_resp_json.