[scripts.utils] Also track versions of dependencies on meta information saved with train and predict scripts

[scripts,engine] Implement fixes on evaluation (closes #20), and prepare for MLflow integration (issue #60)

* Use average-precision instead of AUC for precision-recall figures
* Report average precision on tables
* Include dependence on credible (prepare to handle issue #43)
* Also saves result table in JSON format (prepare to handle issue #60)

Update dependence to lightning (closes #61)

Closes #61

See merge request biosignal/software/mednet!20

Cuda usage is now inferred from the model.

Make square centre-padding a model transform

Closes #39 and #23

See merge request biosignal/software/mednet!16

Update tests

See merge request biosignal/software/mednet!18

Fix Issues when running tests on the CI

See merge request biosignal/software/mednet!19

Histograms comme be compared directly due to differences in how the
torch.resize transform is applied depending on the environment.
A statistical approach using pearson coefficients is used instead.
Testing of alexnet and densenet transforms have been temporarily
disabled as histograms or RGB images need to be handled differently.

Databases should now be available on the CI, we do not rely on a zipped
db anymore.

Samples are not square cropped during data loading anymore due to recent
changes. Cropping is model-dependent.