Backtester: Feature Calculation and Model Driving

[echavemann]

This builds on top of #84 , and will require 84 to be done to be completed.

We need to drive a FeatureGenerator, a subclass of BaseFeatureGenerator that the user will write (see #66), and then run this through the LASSO container as well. Note that we should probably add a backtester config at this point as well, which can take in parameters such as online retrain frequency and fee rate. After calculating the prediction, it should drive a BaseStrategy subclass with the prediction and have the list of actions ready for execution.