Update MSG-4 visible calibration coefficients

[sfinkens]

Update MSG-4 visible calibration coefficients to latest version from May 2023: https://msgcpp.knmi.nl/solar-channel-calibration.html

Edit 1

For some reason creating the conda environment on macos failed (see https://github.com/foua-pps/level1c4pps/actions/runs/5785938956):

ImportError: dlopen(...libmambapy/bindings.cpython-311-darwin.so, 0x0002): Library not loaded:
...
Error: The process '/Users/runner/miniconda3/condabin/mamba' failed with exit code 1

Switching to mambaforge seems to solve the problem.

Edit 2

Now creating the conda env with Python-3.8 takes forever...

I noticed some warnings about deprecated node 12 (https://github.blog/changelog/2023-06-13-github-actions-all-actions-will-run-on-node16-instead-of-node12-by-default/) so I updated actions/checkout and codecov/codecov-action to version 3. But that doesn't speed things up...

• Tests added
• Tests passed: Passes pytest level1c4pps
• Passes flake8

[codecov-commenter]

Codecov Report

Merging #81 (1c8cc0f) into main (1ed7063) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main      #81   +/-   ##
=======================================
  Coverage   77.61%   77.61%           
=======================================
  Files          21       21           
  Lines        1452     1452           
  Branches      107      120   +13     
=======================================
  Hits         1127     1127           
  Misses        301      301           
  Partials       24       24           

Flag	Coverage Δ
unittests	77.61% <ø> (ø)

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

Files Changed	Coverage Δ
level1c4pps/calibration_coefs.py	100.00% <ø> (ø)
level1c4pps/tests/test_seviri2pps.py	96.76% <ø> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[ninahakansson]

Can we just remove support for python 3.8? Or do you still need to run level1c4pps for python-3.8?

[sfinkens]

Can we just remove support for python 3.8? Or do you still need to run level1c4pps for python-3.8?

Forwarding this to @yeigenbrodt 🙂

[yeigenbrodt]

Can we just remove support for python 3.8? Or do you still need to run level1c4pps for python-3.8?

I see no reason to support 3.8.

[sfinkens]

I simplified the satpy version requirement a little (> 0.41), which speeds up the build a bit:

OS	Python	Build time old [min]	Build time new [min]
ubuntu	3.9	inf	22
ubuntu	3.10	116	11
ubuntu	3.11	13	3
macos	3.9	inf	46
macos	3.10	37	22
macos	3.11	14	8

So at least from the Python-3.11 jobs you get a quick feedback. But 46 minutes for Python-3.9 is way too long... Any idea how to speed this up?

[mraspaud]

I see this is already using mamba, which I was going to suggest. I'm very surprised it takes such a long time with mamba though...
As for ways to improve this, maybe the requirements file should be trimmed down? Satpy has hard requirements on pyresample, dask and xarray at least, so these can go.

[sfinkens]

As for ways to improve this, maybe the requirements file should be trimmed down?

That helped a lot! Now at 6 minutes max 🙂

@ninahakansson From my point of view this is ready to merge now.