Enh/setup cleanup

[bzah]

Pull Request to resolve #259

• Unit tests cover the changes.
• These changes were tested on real data.
• The relevant documentation has been added or updated.
• A short description of the changes has been added to doc/source/references/release_notes.rst.

Describe the changes you made

• Update project structure to have sources into a dedicated src directory and automated tests in a tests directory. Both are at root level. This means to run the tests we now have to install icclim in the local environement (with for example pip install --editable .). This is considered a best practice according to https://packaging.python.org/en/latest/discussions/src-layout-vs-flat-layout/#src-layout-vs-flat-layout
• Migrate from setup.py and setup.cfg to pyproject.toml.
  This is reccomended by: https://packaging.python.org/en/latest/guides/modernize-setup-py-project/
• Migrate build toolchain from setuptools and flit. Flit is a simpler build toolchain maintained by the python packaging authority. It makes it simpler to build and release packages to pypi.

TODO:

• Complete the pyproject.toml with the linter and formatter configuration
  • For now a few rules are disabled (either commented in select or added to ignore in pyproject.toml).
    They should be added step by step and will help improve the code quality.
• Migrate to ruff for linting
• Update CI to install requirements from toml file instead of requirement for test, logo generation, API generation and doc generation
• Remove ICCLIM_VERSION from constants and use icclim.version instead (investigate if that's possible without issues)

[github-actions]

Report

File	Stmts	Miss	Cover	Missing
/opt/hostedtoolcache/Python/3.9.18/x64/lib/python3.9/site-packages/icclim
_generated_api.py	178	9	95%	17–27
icclim_logger.py	85	29	66%	24, 43, 65, 70–95, 105, 111–142, 150–155
icclim_types.py	13	13	0%	1–21
main.py	283	34	88%	59–65, 125, 134, 415–419, 502, 528–529, 531–532, 534, 536–537, 539–540, 560–567, 637–642, 727, 734, 738–739
utils.py	40	8	80%	14–15, 25–30, 37–41, 43–44
/opt/hostedtoolcache/Python/3.9.18/x64/lib/python3.9/site-packages/icclim/ecad
xclim_binding.py	52	8	85%	16–18, 44, 48, 77, 81, 110, 114
/opt/hostedtoolcache/Python/3.9.18/x64/lib/python3.9/site-packages/icclim/generic_indices
generic_indicators.py	498	98	80%	49–55, 68, 72, 89, 93, 97, 110–113, 117, 191, 333–337, 339–343, 389, 420–421, 723–724, 759, 765, 773–774, 794, 958–962, 964–965, 983–989, 993, 1021, 1023, 1091–1092, 1125–1151, 1179–1199, 1215, 1224–1225, 1252–1284, 1290–1293, 1315
standard_variable.py	48	1	98%	19
threshold.py	411	34	92%	49–54, 209–210, 234, 238, 286, 302–304, 319–323, 327–331, 411, 553–554, 617–622, 717–718, 913, 936–937, 942, 1016–1017, 1020–1024, 1112–1114
/opt/hostedtoolcache/Python/3.9.18/x64/lib/python3.9/site-packages/icclim/models
cf_calendar.py	35	1	97%	20
climate_variable.py	106	21	80%	27–34, 107–113, 168–169, 178, 214–218, 221, 261, 265–269, 296–297, 334
frequency.py	180	15	92%	36–38, 174–177, 321–327, 331, 414–418, 440, 445
global_metadata.py	6	6	0%	1–9
index_config.py	27	5	81%	7–11
index_group.py	35	2	94%	33–37
logical_link.py	21	1	95%	12
operator.py	31	3	90%	10, 15–19
standard_index.py	30	6	80%	7–10, 59, 62, 66
user_index_dict.py	20	20	0%	1–28
/opt/hostedtoolcache/Python/3.9.18/x64/lib/python3.9/site-packages/icclim/pre_processing
in_file_dictionary.py	8	8	0%	1–29
input_parsing.py	205	53	74%	22–28, 81–82, 86–87, 92–97, 111–112, 180, 185–189, 192, 202–203, 214–218, 244, 258, 272, 293–298, 302, 304, 312–317, 322–327, 338–339, 353–360, 369–389, 414
rechunk.py	89	6	93%	21–22, 43–44, 48, 136
TOTAL	2643	381	86%

Test results

Tests	Skipped	Failures	Errors	Time
189	0 💤	0 ❌	0 🔥	1m 34s ⏱️

[bzah]

There is a drop of the code coverage in this PR, from 91% to 86%.
It is without consequences because the only statements that are now uncovered are the types declarations.
These are now bypassed at runtime with the addition of if TYPE_CHECKING in imports.