MNT: Fix warnings in test suite and adds support for numpy 2.0

[Gui-FernandesBR]

Pull request type

• Code maintenance (refactoring, formatting, tests)

Checklist

• Tests for the changes have been added (if needed)
• Docs have been reviewed and added / updated
• Lint (black rocketpy/ tests/) has passed locally
• All tests (pytest tests -m slow --runslow) have passed locally
• CHANGELOG.md has been updated (if relevant)

Current behavior

Numpy 2.0 was officially released today, but we cannot run rocketpy with that new version because some methods were deprecated.
Also, running pytest in our repo usually raises a lot of annoying warnings.

New behavior

Fixed, fixed, fixed! Please see the commit history.

Breaking change

• No

Additional information

• https://numpy.org/devdocs/numpy_2_0_migration_guide.html#
• The optional dependency timezonefinder cannot be installed with numpy 2.0. See this issue: Can't install timezonefinder with numpy 2.0 jannikmi/timezonefinder#234

[Gui-FernandesBR]

This PR was actually partially problematic: https://github.com/RocketPy-Team/RocketPy/pull/552/files
The 3D trajectory is currently being distorted. Fixed too!

[Gui-FernandesBR]

Something is wrong here. The aritmetic priority is mixed, especially when running tests on macOS.
I can't find an easy solution right now.

[codecov]

Codecov Report

Attention: Patch coverage is 80.00000% with 8 lines in your changes missing coverage. Please review.

Project coverage is 73.69%. Comparing base (6af67ac) to head (2a74fa4).
Report is 15 commits behind head on develop.

Files	Patch %	Lines
rocketpy/environment/environment.py	20.00%	4 Missing ⚠️
rocketpy/plots/flight_plots.py	75.00%	2 Missing ⚠️
rocketpy/_encoders.py	0.00%	1 Missing ⚠️
rocketpy/mathutils/function.py	96.15%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #623      +/-   ##
===========================================
+ Coverage    73.60%   73.69%   +0.09%     
===========================================
  Files           70       70              
  Lines        10292    10304      +12     
===========================================
+ Hits          7575     7594      +19     
+ Misses        2717     2710       -7     

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

[Gui-FernandesBR]

Something is wrong here. The aritmetic priority is mixed, especially when running tests on macOS. I can't find an easy solution right now.

Ok, I have narrowed down the problem, it is a bit clearer now. However, I don't have a solution yet.

The documentation tests are breaking on Python 3.12 with numpy 2.0. Essentially, this is a type promotion problem that was possibly caused by NEP50. This may help to find a solution: https://numpy.org/neps/nep-0050-scalar-promotion.html#alternatives

However, I think the easiest solutions for us would be the removal of the problematic tests.

Below we have an example:

(.venv12) PS RocketPy> pytest rocketpy --doctest-modules
======================================================= test session starts ========================================================
platform win32 -- Python 3.12.4, pytest-8.2.2, pluggy-1.5.0
rootdir: RocketPy
configfile: pyproject.toml
plugins: cov-5.0.0
collected 34 items

rocketpy\environment\environment.py F.F                                                                                       [  8%]
rocketpy\mathutils\function.py FF....                                                                                         [ 26%]
rocketpy\mathutils\vector_matrix.py ................                                                                          [ 73%]
rocketpy\tools.py .........                                                                                                   [100%]

============================================================= FAILURES =============================================================
_________________________ [doctest] rocketpy.environment.environment.Environment.calculate_density_profile _________________________
3084         None
3085
3086         Examples
3087         --------
3088         Creating an Environment object and calculating the density
3089         at Sea Level:
3090
3091         >>> env = Environment()
3092         >>> env.calculate_density_profile()
3093         >>> env.density(0)
Expected:
    1.225000018124288
Got:
    np.float64(1.225000018124288)

RocketPy\rocketpy\environment\environment.py:3093: DocTestFailure
_____________________________ [doctest] rocketpy.environment.environment.Environment.set_gravity_model _____________________________
556         a :class:`Function` object representing the gravity model.
557
558         Examples
559         --------
560         Let's prepare a `Environment` object with a constant gravity
561         acceleration:
562
563         >>> g_0 = 9.80665
564         >>> env_cte_g = Environment(gravity=g_0)
565         >>> env_cte_g.gravity([0, 100, 1000])
Expected:
    [9.80665, 9.80665, 9.80665]
Got:
    [np.float64(9.80665), np.float64(9.80665), np.float64(9.80665)]

RocketPy\rocketpy\environment\environment.py:565: DocTestFailure
_____________________________________ [doctest] rocketpy.mathutils.function.Function.get_value _____________________________________
829         >>> f2.get_value([1, 2, 3], [1, 2, 3])
830         [2, 8, 18]
831         >>> f2.get_value([5], [5])
832         [50]
833
834         Testing with ndarray source (1 dimension):
835         >>> f3 = Function(
836         ...    [(0, 0), (1, 1), (1.5, 2.25), (2, 4), (2.5, 6.25), (3, 9), (4, 16)]
837         ... )
838         >>> f3.get_value(2)
Expected:
    4.0
Got:
    np.float64(4.0)

RocketPy\rocketpy\mathutils\function.py:838: DocTestFailure
_______________________________ [doctest] rocketpy.mathutils.function.Function.is_strictly_bijective _______________________________
2617         -------
2618         result : bool
2619             True if the Function is "strictly" bijective, False otherwise.
2620
2621         Examples
2622         --------
2623         >>> f = Function([[0, 0], [1, 1], [2, 4]])
2624         >>> f.isbijective()
2625         True
2626         >>> f.is_strictly_bijective()
Expected:
    True
Got:
    np.True_

[phmbressan]

The new promotion rules make it rather difficult to support doctest of both numpy<2.0 and the newest one, since there is a type difference in the __repr__ check: True (for older numpy) and np.True_.

The PR #624 bumps the minimum Python version to 3.9, allowing all the docstest to be executed with numpy>=2.0 allowing for consistency. The code itself (and the tests) still work fine on numpy<2.0.

Therefore, that PR should be merged first.