Validation of credentials without querying anything?

[din14970]

I'm trying to write unit tests that will only run if valid credentials are found on the system running the tests. I'm using the get_api_key function but this will only query the information that is found on the system, not validate it. Is there an easy way to validate the credentials using the API?

[cristian-codorean]

@din14970, what about something like the following, valid for API client versions 1.6.3 or older:

import requests

from ecmwfapi.api import ANONYMOUS_APIKEY_VALUES
from ecmwfapi.api import Connection


def test_anonymous_apikey_using_requests():
    token, url, email = ANONYMOUS_APIKEY_VALUES

    response = requests.get(
        "{}/{}".format(url, "who-am-i"),
        headers={
            "Accept": "application/json",
            "From": email,
            "X-ECMWF-KEY": token,
        },
    )

    assert response.status_code == 200


def test_anonymous_apikey_using_connection():
    token, url, email = ANONYMOUS_APIKEY_VALUES

    connection = Connection(url, email=email, key=token)

    try:
        connection.call("{}/{}".format(url, "who-am-i"))
    except:
        assert False

Note one possible test directly queries the API using Python's requests. The other test uses the Connection object in the library so that the call to the API, and whatever else the library does, is abstracted.

Might look into adding these tests to the library itself.