MIXLB Follow up

The following "to-do" items all remain and are apparent after implementing the MIXLB model in PyTorch:

• unit tests for InputMixlB, especially the from_df method.
• moving MIXLB notebooks from notebooks/miscellaneous to notebooks/.
• creating a docker image for the repository.
• creating a reproducible workflow that:
  1. Performs all necessary raw data transformations for
     1. model estimation and
     2. forecasting.
  2. Estimates all models (MNL, MIXLA, MIXLB).
  3. Creates all model checking outputs, for each model.
  4. Creates all replication outputs, for each model.
  5. Copying all outputs used in the accompanying article into locations that are easily accessible by the article latex document and easily packaged for ArXiv.
  6. Compiles the accompanying article.
  7. Creates the zip archive for upload to ArXiv.
• making the accompanying research article fully reproducible (eliminating manually entry/edit of values, charts, and figures).