Protlearn

- fixed SOCN dimensionality error in qso.py

- new preprocessing feature: onehot_encode()
- increased efficiency of various features
- no longer automatically deleting all-zero columns of features due to dimension mismatches between training and test data for prediction tasks
- bug fix: socn() - previously, it was only returning the results of the last sequence in the dataset due to a bug in the for-loop over all sequences

Python package for extracting amino acid sequence features from proteins and peptides.