FiberHMM v1.2

Major changes:

1. Added a new parameter to train_model and apply_model, -e. This allows reads to be masked by n bp at either end with 0% methylation probability. This is necessary due to the inability of fibertools to call methylations in the first and last 8 bp of the read.
2. Added a folder with example input, intermediate, and output files to allow for testing of the scripts.
3. Added a folder to store pretrained models.

Small changes:

1. Fixed a bug where column names of BlockSizes/BlockStarts were swapped in the final output
2. Standardized the -b flag across all scripts
3. Removed a reference in the Readme to multiple input files for apply_model
4. Added parameter -e, as noted above.
5. Changed apply_model.py to output a single bed file instead of individual bed files for each chromosome (matching apply_model_multiprocess.py)