[Azure DataStore] Handle storage options as secrets

[hayesgb]

This add the ability to pass standard dictionary keys from fsspec's storage_options parameter into mlrun.run.get_dataitem() as secrets.

Enabling this will more easily allow users engaging in exploratory analysis to leverage the mlrun api to fetch data_items from Azure by enabling the following

storage_options={'account_name': "<NAME>", 'credential': <CREDENTIAL>}
df = mlrun.run.get_dataitem("az://CONTAINER/myfile.parquet", secrets=storage_options).as_df()

[Hedingber]

@hayesgb Looks good to me, just note you have some conflict

[hayesgb]

Fixed merge conflict. Thanks!

[Hedingber]

@hayesgb I now noticed that the test file you added is almost an exact duplicate of test_azure_blob.py
I feel like we can merge them without too much effort
Looks like you'll need to make the verify_auth... method to behave a bit different whether it's env vars or storage options (can be determined by whether one of the params starts with AZURE_)
And in the tests simply always pass storage options to secrets, just in the case of env vars, it will be None
WDYT ?

[hayesgb]

@Hedingber -- I considered combining the tests, but was concerned about the interpretability. The two approaches serve entirely different use-cases, and I was worried that if the tests were blended (storing secrets or retrieving environmental variables vs user-passed credentials, which will most likely be done during exploratory analysis) it would lead to confusion and create potential maintainability issues. Thoughts?

[Hedingber]

@hayesgb I see your point but I feel like leaving the tests duplicated will hurt maintainability more 😬 , @theSaarco WDYT ?

[hayesgb]

@Hedingber -- No worries. I consolidated the tests into a single file and updated the PR. LMKWYT.

[Hedingber]

@hayesgb Looks great!
merged
Big appreciation for the detailed comment in the verify auth method 👏 👏