Hyper parameter optimization for sliding window arguments enabled

• feat!: removed data preparation out of predictor class, sub_seq, steps_past, steps_future need now to be defined in each model method
  • allows for advanced hyper parameter tuning
• fix: removed tensor board activation logic bug

fix: removed dead pandas imports

• fix: removed dead pandas imports
• chore: added tensorflow as base requirement

imbrium 2.0.0

• adapted keras_core
• removed internal hyperparameter tuning
• removed encoder-decoder architectures
• improved layer configuration via dictionary input
• split data argument into target and feature numpy arrays

Configurable layer depths and optimizer

• added depth parameter to architectures
• added optimizer configuration support
• added optimizer configuration to seeker

Regularization, Dropout, TensorBoard and improved performance graph

• added Tensorboard support
• changed show_performance plot to show loss and metric values
• added optional dropout and regularization layers to architectures

Hyper-parameter optimization added

• removed batch_size parameter from fit_model method
• hyperparameter optimization added via the Optuna library

Updated dependencies

• imbrium supports now:
  • python 3.7, 3.8, 3.9, 3.10

imbrium revision

• added tests for new utils module

• name changes of predictor classes:

univarstandard => univarpure
BasicMultStepUniVar => PureUni
univarhybrid => univarhybrid (unchanged)
HybridMultStepUniVar => HybridUni
multivarstandard => multivarpure
BasicMultSTepMultVar => PureMulti
multivarhybrid => multivarhybrid (unchanged)
HybridMultStepMultVar => HybridMulti

• tox added

• outsourced Binder demo notebook to https://github.com/maxmekiska/ImbriumTesting-Demo

• new README.md

Code refactoring, minor name change

0.1.8

• major refactoring of code base
• minor name change from Imbrium to imbrium

Code format change, Minor refactoring

• improved source code format

• improved scaling selector mechanism