Deprecate old AF2 modules: param, baseoperatorlink, secrets cache

[sunank200]

This PR shows warnings at the top of each of the three re-export modules:
* airflow.models.param
* airflow.models.baseoperatorlink
* airflow.secrets.cache

---

^ 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.

[amoghrajesh]

We have a utility: add_deprecated_classes which can do it for us. Let's use that instead. Refer: https://github.com/apache/airflow/pull/50435/files for example

[amoghrajesh]

Last set of nits. LGTM

[shahar1]

Tests fail :(

[jscheffl]

I propose to backport to 3.0-line!

[potiuk]

This is pretty unintended usage of those utils. The way how it works is that you should add those dicts and call those methods in in __init__.py of the packages which contain deprecated modules - this way we avoid having even the module files. Recently @amoghrajesh made similar assumption and he added docstring to the classes explaining it better.

https://github.com/apache/airflow/blob/main/airflow-core/src/airflow/utils/deprecation_tools.py#L83

See for example the usage https://github.com/apache/airflow/blob/main/airflow-core/src/airflow/decorators/__init__.py - that leaves behind only folder with __init__,py and no module files at all.

[potiuk]

#50510 with better description of how it should be used.

[potiuk]

That's it ! way less visible, and even autocomplete won't work in IDEs :)

[sunank200]

That's it ! way less visible, and even autocomplete won't work in IDEs :)

Thanks @potiuk

I did not realise the way of usage of add_deprecated_classes before.

[potiuk]

That's it ! way less visible, and even autocomplete won't work in IDEs :)

Thanks @potiuk

I did not realise the way of usage of add_deprecated_classes before.

Yeah it was a bit "use by following other usages" rather than "by docs" -> I hope the update to docstring I added will make it clearer what is the intended usage :)

[amoghrajesh]

This is pretty unintended usage of those utils. The way how it works is that you should add those dicts and call those methods in in __init__.py of the packages which contain deprecated modules - this way we avoid having even the module files. Recently @amoghrajesh made similar assumption and he added docstring to the classes explaining it better.

https://github.com/apache/airflow/blob/main/airflow-core/src/airflow/utils/deprecation_tools.py#L83

See for example the usage https://github.com/apache/airflow/blob/main/airflow-core/src/airflow/decorators/__init__.py - that leaves behind only folder with __init__,py and no module files at all.

Ah! My bad for the review. I encountered the same issue but didn't look carefully enough.

[amoghrajesh]

Alright, looks great now!

[github-actions]

Backport failed to create: v3-0-test. View the failure log Run details

Status	Branch	Result
❌	v3-0-test

You can attempt to backport this manually by running:

cherry_picker d6e8418 v3-0-test

This should apply the commit to the v3-0-test branch and leave the commit in conflict state marking
the files that need manual conflict resolution.

After you have resolved the conflicts, you can continue the backport process by running:

cherry_picker --continue

[amoghrajesh]

Manual cherry pick: #50588

[amoghrajesh]

This can be 3.1, not an imp bug fix.