PERF: refactor atef.bin cli entrypoint to defer imports

[tangkong]

Description

• takes a knife to atef.bin to defer imports until absolutely necessary

In more detail:

• moves all build_arg_parser type functions to a module separate from the subcommand logic
• atef.bin.main now grabs those parsers to create the top-level parser, avoiding the meaty functional imports
• the main sub-commands are given an additional layer of wrapping, and atef.bin.main:main now calls a partial function to get the atef.bin.<subcommand>:main function, which then gets fed the parsed kwargs.

Motivation and Context

I keep vendoring this cli hook and never improving it. This was more of a proof to myself that it could work. If we like it I can start using it elsewhere or even maybe cookiecutter it.

I should really work on more important tasks but I refactored it in my brain during a post-dinner walk and wanted to put it to paper

How Has This Been Tested?

Before

$ time atef config -h
...
real    0m8.487s
user    0m6.721s
sys     0m1.158s

After

$ time atef config -h
...
real    0m0.518s
user    0m0.184s
sys     0m0.075s

Aside from this, lots of interactive testing to make sure all the entrypoints are working.

Where Has This Been Documented?

This PR

Pre-merge checklist

• Code works interactively
• Code contains descriptive docstrings, including context and API
• New/changed functions and methods are covered in the test suite where possible
• Code has been checked for threading issues (no blocking tasks in GUI thread)
• Test suite passes locally
• Test suite passes on GitHub Actions
• Ran docs/pre-release-notes.sh and created a pre-release documentation page
• Pre-release docs include context, functional descriptions, and contributors as appropriate

[ZLLentz]

This seems reasonable to me and is a big speedup due to some of the slower imports in some of the code paths. My only in-line review comment is a typo check.

I have mixed feelings about the various variants/options for a subparser scheme like this one. The previous arrangement had some performance problems but functionally made clear separations between the implementations of the different subcommands. Part of me wonders if there was a way to keep the subparsers a little more separate without needing to change so much, something like:

atef.bin.subcmd

import argparse

def build_parser():
    ...

def main(**kwargs):
    from .subcmd_main import subcmd_main
    subcmd_main(**kwargs)

And then the overall structure doesn't change too much and we still import a bunch of subparsers but the heavy imports are deferred and the subparsers aren't collocated.

But maybe there are even cleaner ways as well.

This is all just musing. I'm happy to merge with the typo fixed.

[ZLLentz]

Suggested change

	- improves the performance of the CLI entrypoint, defering functional imports as longas possible
	- improves the performance of the CLI entrypoint, deferring functional imports as long as possible

[tangkong]

Hmm I might like that formulation more. I'll give that a whirl

[tangkong]

superceded by #257