Rework python-cmethods: replace the CMethods class by the new adjust function

[btschwertfeger]

To improve the performance of applied bias corrections, the complete CMethods class was removed. Individual techniques can't be executed directly by calling their function anymore. Also the adjust_3d function was removed.

Instead, a new function named adjust was added which acts like the old adjust_3d function but can handle single and multidimensional data sets and will return the bias corrected data as xarray.Dataset. By using xarrays ufunc function instead of iterating over entries per dimension, the execution time could be reduced drastically.

• Since functions like linear_scaling, variance_scaling and so on are now internal functions and should not be called directly, they are not part of the documentation anymore. Instead, a new section in the documentation was added - which describes the implemented methods and provides some examples.

More changes:

• split init.py into different modules
• reworked the unit tests
• updated examples
• updated README
• updated documentation
• chunked data can now be adjusted more efficiently

Follow-up: Create issue for improving the DQM function (#48 (comment)).

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Comparison is base (9878005) 98.23% compared to head (e1868d8) 98.08%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #48      +/-   ##
==========================================
- Coverage   98.23%   98.08%   -0.15%     
==========================================
  Files           1        7       +6     
  Lines         283      261      -22     
==========================================
- Hits          278      256      -22     
  Misses          5        5              

Flag	Coverage Δ
unittests	98.08% <98.03%> (-0.15%)	⬇️

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

Files	Coverage Δ
cmethods/__init__.py	100.00% <100.00%> (+1.76%)	⬆️
cmethods/scaling.py	100.00% <100.00%> (ø)
cmethods/static.py	100.00% <100.00%> (ø)
cmethods/types.py	100.00% <100.00%> (ø)
cmethods/core.py	97.22% <97.22%> (ø)
cmethods/distribution.py	98.00% <98.00%> (ø)
cmethods/utils.py	96.55% <96.55%> (ø)

[riley-brady]

Great job with this! I just focused mainly on the apply ufunc implementation. Just a few comments here. It worked on my example data over 1-4D, and very quickly with dask.

[riley-brady]

Alles gute bei mir! Well done.