Use pyproject.toml with hatch

[getzze]

Modernize the building and testing framework:

• use pyproject.toml
• use hatch environments
• add test_script to test plotting figures -> added in Compatibility with seaborn 0.13 #155
• untrack docs/build

To build: hatch build or python -m build
To bump version: hatch version minor
To test: hatch run tests:run
To test with coverage: hatch run tests:run-cov && hatch run coverage:run
To build and test docs: hatch run docs:all
To run examples: hatch run examples:run -- --plot_type=boxplot --data_type=df_with_group_and_hue

To test #155 I had to write the test_script.py file and it was very convenient to use hatch environments for testing. So I thought I should actually make a PR about it :)

[codecov]

Codecov Report

❌ Patch coverage is 45.45455% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 85.18%. Comparing base (48a3771) to head (b5360eb).
⚠️ Report is 7 commits behind head on master.

Files with missing lines	Patch %	Lines
statannotations/compat.py	25.00%	6 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (48a3771) and HEAD (b5360eb). Click for more details.

HEAD has 4 uploads less than BASE

Flag	BASE (48a3771)	HEAD (b5360eb)
	5	1

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #156      +/-   ##
==========================================
- Coverage   91.39%   85.18%   -6.22%     
==========================================
  Files          34       16      -18     
  Lines        2465     1505     -960     
  Branches        0      210     +210     
==========================================
- Hits         2253     1282     -971     
+ Misses        212      210       -2     
- Partials        0       13      +13     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[getzze]

I modified a bit the Github action so instead of calling hatch once to run all the tests, the workflow creates a matrix of python and seaborn versions (we could also add an OS matrix if needed, see https://github.com/platformdirs/platformdirs/ for the modifications to do).

Github runs matrix jobs in parallel so the tests are substantially faster.

[getzze]

I forgot to mention that python 3.7 is not tested (hatch was not working for me with this python version). It would be better to drop it, as it is EOL anyway.

[getzze]

Thanks for merging #155, I rebased this PR, all tests are passing now. I will check the jupyter notebook.

Reminder: minimal python version is set to 3.8

[getzze]

The notebook runs correctly with python 3.12.

I added support for python 3.13 (only compatible with seaborn 0.13), the tests are passing.

[getzze]

Coverage is working now!

[getzze]

@trevismd it's ready for review

[getzze]

closes #150

I moved all the source files to a src folder, which avoids some errors when running tests and building docs.

Also I added some linting scripts: typos , end-of-file-fixer and trailing-whitespace

[getzze]

Hi @trevismd ,
I rebased this PR. It seems it has a lot of changes but most of it is getting rid of the build docs folder (that should not be tracked with git).
I believe this PR will help make the project more maintainable by moving to modern packaging with pyproject.toml.

The next step would be to add type annotations and a fast linter (ruff). But for that we need to have a pyproject.toml.

A reverted moving the source to a src folder as it is not critical and it makes the changes less trackable. If needed it can be done in a separate PR that does only that.

[phaleanopsis]

I think it would make sense to merge this PR, and then as a next step work to improve docstrings and type annotations to make the package and documentation a bit more accessible?

[getzze]

I rebased this PR, bumping the minimal python version to 3.10.

[getzze]

@trevismd I fixed some small things (to get the test working with pandas>=3 and getting full coverage).
It's ready for review!