Deprecate kadi load_ska_dir() in favor of parse_cm load_dir_from_load_name

[taldcroft]

Description

The kadi.commands.core function load_ska_dir() has been superceded by parse_cm.paths.load_dir_from_load_name(). This PR re-implements the kadi load_ska_dir() function as a call to load_dir_from_load_name() and issues a deprecation FutureWarning.

Interface impacts

This function is now deprecated and using it will cause unit tests to fail or issue warnings to the console.
From https://github.com/search?q=org%3Asot%20ska_load_dir&type=code it seems that only mica was using this (fixed in sot/mica#288).

Testing

Unit tests

• Mac
  With Use parse_cm load_dir_from_load_name and test mica#288 installed:

(razl) ➜  kadi git:(deprecate-ska-load-dir) git rev-parse HEAD  
1d6011d8e0cd5ba9a2babc8dc57b1755d5e465ff
(razl) ➜  kadi git:(deprecate-ska-load-dir) pytest            
============================================================================= test session starts ==============================================================================
platform darwin -- Python 3.10.8, pytest-7.2.1, pluggy-1.0.0
rootdir: /Users/aldcroft/git, configfile: pytest.ini
plugins: timeout-2.1.0, anyio-3.6.2
collected 230 items                                                                                                                                                            

kadi/commands/tests/test_commands.py ......................................................................................                                              [ 37%]
kadi/commands/tests/test_states.py ......................x.............................................x.......................                                          [ 77%]
kadi/commands/tests/test_validate.py ....................                                                                                                                [ 86%]
kadi/tests/test_events.py ..........                                                                                                                                     [ 90%]
kadi/tests/test_occweb.py ......................                                                                                                                         [100%]

================================================================== 228 passed, 2 xfailed in 98.30s (0:01:38) ===================================================================

Independent check of unit tests by [REVIEWER NAME]

• [PLATFORM]:

Functional tests

No functional testing.

[jeanconn]

I think you've used FutureWarning in recent work (like https://github.com/sot/chandra_cmd_states/blob/efd4c947c6508c63df232344d58696fa9d11a6eb/chandra_cmd_states/__init__.py#L17) so maybe that's our convention?

[javierggt]

Looks good to me. The change is simple enough. The tests pass.

I think one can use FutureWarning instead of the numpy warning. What is the reason not to?

[javierggt]

What is the issue with using FutureWarning here? (as you suggest) It would seem to me that this is exactly what it is meant for, but perhaps I'm missing something.

Strictly speaking, maybe the intended use is to create our own deprecation warning class and make it inherit from FutureWarning, though I am not proposing to do that.

[taldcroft]

I guess I like that the numpy one has the word Deprecation in it, but I see the point and have changed this.