massive refactoring required

[keighrim]

Because

As there are many instances of "copies" of python scripts/functions/methods in the project root directory, the codebase is really hard to read and review. For example, I spent 30 mins to figure out why my edits were now properly staged until I realized that there are to classify.py files when I first started reviewing the app submission. De-cluttering these stopgap practice will highly improve readability and maintenance of the codebase.

Done when

Only one copy of all classes, modules, methods, and functions related to training and classification are left and properly packaged in a python package.

Additional context

No response

[keighrim]

done via #32

[keighrim]

Reopening this as we change to course and decided to release v1.0 with the duplicate code.

[marcverhagen]

Duplicate code is gone I think. One remaining thing that I can think of is to split classification of images from generation of timeframes (now they are not closely entangled, but still in the same file).

[marcverhagen]

Refactoring will continue, but the massive changes that were needed are all done I think.