Made get_conn method async for KiotaRequestAdapterHook to fix AsyncToSync error

[dabla]

This is a fix for "An error occurred: You cannot use AsyncToSync in the same thread as an async event loop - just await the async function directly." in Airflow 3, as get_conn should be async.

closes: #54598

---

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

[gopidesupavan]

@dabla thanks overall LGTM, some tests needs to be fixed.

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

Offcourse this change must for this operator to work with AF3.

[gopidesupavan]

cc: @eladkal FYI.

[dabla]

@dabla thanks overall LGTM, some tests needs to be fixed.

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

Offcourse this change must for this operator to work with AF3.

We could but normally this is an "internal" method, as user would normally not interact with it when using operator of even the hook as there you have the run method. WDYT as get_conn is expected to be async now.

[eladkal]

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

This is not good. I don't think we can break Airflow 2.10 comparability for this do we?

[dabla]

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

This is not good. I don't think we can break Airflow 2.10 comparability for this do we?

I don't think it's broken for Airflow 2 as this is "internal kitchen stuff", I can test it to make sure in Airflow 2.

[dabla]

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

This is not good. I don't think we can break Airflow 2.10 comparability for this do we?

If they directly interact with get_conn then yes it's would be broken, but that would be a bad practise normally you should use the run method. I can add a new get_async_conn method and keep the original for backward compatibility.

[dabla]

BTW this is complete breaking change for AF2 users or anyone using these hooks methods , because get_conn is now async awaitable. so its better to add some compatibility and throw warning to users ?

This is not good. I don't think we can break Airflow 2.10 comparability for this do we?

I'll made get_conn synced again and create new method get_async_conn and deprecate the former one.

[eladkal]

If they directly interact with get_conn then yes it's would be broken, but that would be a bad practise normally you should use the run method. I can add a new get_async_conn method and keep the original for backward compatibility.

I'm not into the details of this change. If this works on AF2 and we have reasonable claim to classify it as bug fix then this is fine. If mitigation steps or special instructions into what was changed are needed then we need a log entry to specify this change (a key indication to know if this is required is: if we expect that as a result of releasing this change workflows will behave differently)

[dabla]

If they directly interact with get_conn then yes it's would be broken, but that would be a bad practise normally you should use the run method. I can add a new get_async_conn method and keep the original for backward compatibility.

I'm not into the details of this change. If this works on AF2 and we have reasonable claim to classify it as bug fix then this is fine. If mitigation steps or special instructions into what was changed are needed then we need a log entry to specify this change (a key indication to know if this is required is: if we expect that as a result of releasing this change workflows will behave differently)

Well I just kept original method and added new async one like other providers do, it wasn't big of a deal after all, and so we are certain we don't break anything, @gopidesupavan was right by mentioning this.

[eladkal]

Well I just kept original method and added new async one like other providers do, it wasn't big of a deal after all, and so we are certain we don't break anything, @gopidesupavan was right by mentioning this.

Cool

[dabla]

Will adapt PowerBI related triggers also so it's aligned with MSGraph as they use the same code.

[dabla]

@eladkal Should be ready now

[gopidesupavan]

Thanks for the update, instead of removing the properties its better to throw depreciation warning. my only concern is all these are public facing methods or properties, not how users interact with these so its better with deprecation warning. WDYT?

[dabla]

Thanks for the update, instead of removing the properties its better to throw depreciation warning. my only concern is all these are public facing methods or properties, not how users interact with these so its better with deprecation warning. WDYT?

I've re-added get_conn with deprecation.

[eladkal]

LGTM
cc @gopidesupavan

[gopidesupavan]

Lovely thanks @dabla :)