Handle prefix parameters in Learner

[timokau]

Description

See #169 for context. This fix is a bit "hacky, works for now". We will likely have to rewrite or significantly modify this code anyway as part of #125.

How Has This Been Tested?

Lints & tests, CI.

Does this close/impact existing issues?

Fixes #169.

Types of changes

• [x] Bug fix (non-breaking change which fixes an issue)
• [ ] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• [x] My code follows the code style of this project.
• [ ] My change requires a change to the documentation.
• [ ] I have updated the documentation accordingly.
• [ ] I have added tests to cover my changes.

[timokau]

Example:

[ins] In [1]: from csrank.objectranking import FATEObjectRanker; fate = FATEObjectRanker()
Using TensorFlow backend.
/nix/store/kax45bpa01hh152r06d0x049yb5pjwxn-python3-3.7.9-env/lib/python3.7/site-packages/sklearn/utils/deprecation.py:143: FutureWarning: The sklearn.datasets.samples_generator module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.datasets. Anything that cannot be imported from sklearn.datasets is now part of the private API.
  warnings.warn(message, FutureWarning)

[ins] In [2]: fate.get_params()
Out[2]: 
{'activation': 'selu',
 'batch_size': 256,
 'kernel_initializer': 'lecun_normal',
 'kernel_regularizer': <function keras.regularizers.l2(l=0.01)>,
 'loss_function': <function csrank.losses.identifiable.<locals>.wrap_loss(y_true, y_pred)>,
 'metrics': (<function csrank.metrics.zero_one_rank_loss_for_scores_ties(y_true, s_pred)>,),
 'n_hidden_joint_layers': 32,
 'n_hidden_joint_units': 32,
 'n_hidden_set_layers': 2,
 'n_hidden_set_units': 2,
 'optimizer': keras.optimizers.SGD,
 'random_state': None,
 'optimizer__learning_rate': 0.01,
 'optimizer__momentum': 0.0,
 'optimizer__nesterov': False}

[ins] In [3]: fate.set_params(optimizer__learning_rate=0.02)

[ins] In [4]: fate.get_params()
Out[4]: 
{'activation': 'selu',
 'batch_size': 256,
 'kernel_initializer': 'lecun_normal',
 'kernel_regularizer': <function keras.regularizers.l2(l=0.01)>,
 'loss_function': <function csrank.losses.identifiable.<locals>.wrap_loss(y_true, y_pred)>,
 'metrics': (<function csrank.metrics.zero_one_rank_loss_for_scores_ties(y_true, s_pred)>,),
 'n_hidden_joint_layers': 32,
 'n_hidden_joint_units': 32,
 'n_hidden_set_layers': 2,
 'n_hidden_set_units': 2,
 'optimizer': keras.optimizers.SGD,
 'random_state': None,
 'optimizer__learning_rate': 0.02,
 'optimizer__momentum': 0.0,
 'optimizer__nesterov': False}