Reorganize testing

[PProfizi]

This PR reorganizes tests according to a testing matrix defined by the core developer team.
The goal is to spread different test requirements at different steps of the development process to both increase development ease and ensure a thorough robust testing.

The testing configuration to implement is:

On commit to a draft PR:

• unit testing on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.

On "ready for review", on commit to a non-draft PR, or on push to master, thus required to pass for merging on master:

• unit testing, on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• examples, on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• compatibility with the latest PyDPF-Post, on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• unit testing, on every older standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• documentation, on the latest standalone, for Python 3.7, on windows-latest.

On push of a release tag, thus required to pass for releasing:

• unit testing, on the latest standalone, for all supported Python versions, on windows-latest and ubuntu-latest.
• examples, on the latest standalone, for all supported Python versions, on windows-latest and ubuntu-latest.
• compatibility with the latest PyDPF-Post, on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• unit testing, on every older standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• compatibility with the latest PyDPF-Post, on every older standalone, for Python 3.7, on windows-latest and ubuntu-latest.
• documentation, on the latest standalone, for Python 3.7, on windows-latest.
• gatebin test, on the latest standalone, for Python 3.7, on windows-latest and ubuntu-latest.

On a weekly basis, run automatically:

• The same as for a release (without the part for drafting a release).

We also give the capability to run manually, with choice of standalone, OS, and Python versions to test:

• Unit tests.
• Examples.
• Compatibility with the latest PyDPF-Post
• Documentation
• Gatebin test.

[codecov]

Codecov Report

Merging #470 (a33faa5) into master (c7f29e4) will increase coverage by 0.07%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #470      +/-   ##
==========================================
+ Coverage   88.00%   88.07%   +0.07%     
==========================================
  Files          68       68              
  Lines        7568     7574       +6     
==========================================
+ Hits         6660     6671      +11     
+ Misses        908      903       -5     

[anslpa]

@PProfizi shouldn't we use a global variable for all the "231" references that might avoid to have a lot to change each release?

[PProfizi]

@anslpa I do not think I can put an expression in the default value of workflow inputs.
The pipelines will have to be updated in a few places anyway.
Maybe NOT putting a default value is the solution, so that only the CI, and CI_release pipelines should be updated.