Ensure that environment.yml works with future updates

[adehecq]

Make sure that the environment.yml allows running the code as it was used for the RAGMAC intercomparison (last change in June 2022):

• Hard-code geoutils and xdem version, which have changed a lot
• hard-code rasterio version as it was causing an error
• hard-code numba version as it was raising a warning
• hard-code python, numpy, scipy, pandas and toml versions for safety.

Hopefully this will not be too restrictive for any type of architecture...

TODO:

• add a test to ensure that some outputs are well reproduced with the set environment.

[adehecq]

With specified environment, the code does not run. For example python scripts/main_experiment1.py -c AT_Hintereisferner yields:

• a deprecation warning: "NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details."
• the initial statistics seems to fail with an empty table in data/processed/experiment_1/AT_Hintereisferner/init_stats.csv, so no valid DEMs are found. The issue is not obvious because it occurs during the multiprocessing, but is related to a call to rasterio: "*** ValueError: 'bilinear' is not a valid rasterio.warp.Resampling method. Valid methods: ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '10', '11', '12', '13', '14']"

[adehecq]

Teh above issues are solved by freezing the rasterio and numba versions. I also freezed the python, numpy, scipy, pandas and toml versions but trying to not be too restrictive.