scrape abstraction for input batch size requirements

[timothyyu]

limit n size for input batch for stateful = True
i.e. loop abstraction to split 10 hours of data into 1 hour each, or 15min x4 per hour for training and validation; related to train_test_split issue

[timothyyu]

closely related to #22

[timothyyu]

related to #50;
needs work due to fixed batch input size due to RNN Stateful = True

Relevant:
https://stackoverflow.com/questions/43702481/why-does-keras-lstm-batch-size-used-for-prediction-have-to-be-the-same-as-fittin

Tensorflow can predict with different batch and step size than keras; Keras cannot

Alternate: create two models in keras:

1. First model used for training
2. Second model used for predictions, but share weights from both models