Upgrade pants, Makefile, CI, and scripts

[haydenbaker]

pants

The latest version of pants is fairly different from the version we were using - much has changed.

• linters/fixers/etc options (in pants.toml) have changed a bit
• lockfile recommendations have changed, we'll use a single lockfile for the repo (python-default.lock)
  • install_from_resolve must be set to use plugins (like pytest-async, pytest-cov, etc.) for supported tools
    • Setting this means that pants uses the requirements given in the resolve
• pants no longer recommends checking in the pants runner script, and instead recommends using a get-pants installer script to pull in the latest binary
  • It's packaged as a PEX, and no longer requires a specific version of python (3.9) to be installed to run it
  • The binary will use the pants-version as specified by the pants.toml file
• Since we use make, pants will be installed to the repo root and can be ran using ./pants (similar to before)
• The init-pants action is now used, it will set up pants for CI
  • The installation dir is not configurable in the action, so a wrapper script is copied in the repo-root as part of the pants make-recipe IF it is detected that we're in a CI environment (CI env-var is set to true)
• pants fix runs all the formatters/fixers, whereas pants format only runs the formatters

make

As part of upgrading pants, several new recipes are added. Also, to make CI more ergonomic, I've transitioned our workflows to use make as much as possible.

examples

An example model (examples/weather) is added with configuration to generate a rest-json python client

• With the smithy-cli installed, you can run smithy build to generate the client

isort issue

When upgrading pants, isort begins to behave strangely

• An issue was opened for this: (isort no longer discovers configuration after 2.14 to 2.16 upgrade pantsbuild/pants#19516)
• However, It is more than likely related to this: (isort behavior depends on which files are available in the sandbox pantsbuild/pants#15069)
  • When executing separate commands for each package (instead of ::), isort will correctly distinguish 1p and 3p import statements (i.e. pytest)
  • Since this is a bug, we'll need to use separate commands for each python package (until #15069 is fixed)

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (94fdd7d) 94.62% compared to head (08ec194) 94.64%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #188      +/-   ##
===========================================
+ Coverage    94.62%   94.64%   +0.02%     
===========================================
  Files           35       35              
  Lines         1675     1682       +7     
===========================================
+ Hits          1585     1592       +7     
  Misses          90       90              

Flag	Coverage Δ
unittests	94.64% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JordonPhillips]

Suggested change

	uses: actions/checkout@v3
	uses: actions/checkout@v4

[JordonPhillips]

Suggested change

	uses: actions/checkout@v1
	uses: actions/checkout@v4

[haydenbaker]

Whoops, I meant to update all the usages of checkout. Good catch!