Improve protopipe.mva and protopipe-MODELS

[HealthyPear]

Description

• allows for full training scenario, i.e. the whole input dataset is used for training and there is no exported test data (aka the test sample is the one used in the next step of the pipeline and thus the benchmarking will be possible only at that stage)
• improved usage of GridSearchCV algorithm for model cross-validation (exhaustive grid search along hyper-parameter space)
• update of model-related configuration files (and their test version)
• small fixes

[codecov]

Codecov Report

Merging #164 (8057ba5) into master (4f66db8) will decrease coverage by 0.53%.
The diff coverage is 62.13%.

@@            Coverage Diff             @@
##           master     #164      +/-   ##
==========================================
- Coverage   67.19%   66.65%   -0.54%     
==========================================
  Files          24       24              
  Lines        2545     2600      +55     
==========================================
+ Hits         1710     1733      +23     
- Misses        835      867      +32     

Impacted Files	Coverage Δ
protopipe/mva/train_model.py	54.68% <36.36%> (-9.32%)	⬇️
protopipe/scripts/write_dl2.py	80.67% <50.00%> (-0.44%)	⬇️
protopipe/pipeline/io.py	47.27% <57.69%> (+23.93%)	⬆️
protopipe/mva/utils.py	29.07% <63.63%> (-1.58%)	⬇️
protopipe/scripts/build_model.py	84.61% <80.00%> (-0.76%)	⬇️
protopipe/mva/io.py	100.00% <100.00%> (ø)
protopipe/scripts/model_diagnostic.py	7.21% <100.00%> (+0.44%)	⬆️
protopipe/pipeline/utils.py	46.34% <0.00%> (-2.93%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4f66db8...8057ba5. Read the comment docs.