Closed doxenix closed 1 year ago
Hi @doxenix ,
actually there's a mistake in the docs. We have 4 classes in gcp.py
:
class MLFlowGoogleIAMCredentialsProvider(DynamicConfigProvider)
class MLFlowGoogleOAuthCredentialsProvider(DynamicConfigProvider)
class MLFlowGoogleIAMRequestHeaderProvider(RequestHeaderProviderWithKedroContext)
class MLFlowGoogleOauthRequestHeaderProvider(MLFlowGoogleIAMRequestHeaderProvider)
For dynamic authorization you should use the *RequestHeaderProvider
classes, the other two are for legacy, "static" authorization method. Sorry for that. Could you change the docs instead of the code in the PR, since you're already on it? 🙂
FYI, I've discussed the "dynamic authorization" issue with the author of kedro-mlflow
and this functionality was ported to the kedro-mlflow
plugin natively:
https://kedro-mlflow.readthedocs.io/en/stable/source/04_experimentation_tracking/01_configuration.html#authentication-with-expiring-tokens
Hi @marrrcin Thanks for explanation. Docs is corrected right now.
class MLFlowGoogleOAuthCredentialsProvider
ingcp.py
file should inheritance fromRequestHeaderProviderWithKedroContext
Based on documentation - doc If we will choose OAuth authorization and do like is desribed right now:
settings.py
we get error:
AssertionError: Provider class needs to be a subclass of RequestHeaderProviderWithKedroContext