Vlad Apetrei authored Jun 17, 2019 and David Monllaó committed Sep 09, 2019

Adds multi-class capabilities to prediction processors
as well as multi-classification unit tests

Use the new method to check that the installed Python package has high
enough yet still API-compatible version.

This method is to be used for checking that a compatible version of the
moodlemlbackend package is installed on the server. The package is
expected to use the semantic versioning scheme (semver.org).

0.0.4 update fixes an incorrectly specified tensorflow dependency

- Basic unit test for minimum machine learning backends requirements
- Warning return messages now include not enough data
- Clear models when the predictions processor is changed
- Refined the name of a couple of constants / methods

- Method names renamed to avoid interface changes once
  we support regression and unsupervised learning
- Adding regressor interface even if not implemente
- predictor interface comments expanded
- Differentiate model's required accuracy from predictions quality
- Add missing get_callback_boundary call
- Updated datasets' metadata to allow 3rd parties to code
  regressors themselves
- Add missing option to exception message
- Include target data into the dataset regardless of being a prediction
  dataset or a training dataset
- Explicit in_array and array_search non-strict calls
- Overwrite discrete should_be_displayed implementation with the binary one
- Overwrite no_teacher get_display_value as it would otherwise look
  wrong
- Other minor fixes

Part of MDL-57791 epic.

Part of MDL-57791 epic.

Part of MDL-57791 epic.