Improved Documentation

[SomeoneInParticular]

Now that this project is inching towards being ready to be used on new projects (not just mine and Jan's), investing into proper documentation to help new devs use, extend, and contribute to the project is probably warranted. It will also help us standardize how new additions are made, and keep everyone on the same page.

Off the top of my head, the following needs to be done to accomplish this library becoming "general use":

• A formal tutorial on how to install and use the program (using the data provided in the testing directory)
• Documentation on the overall structure of the analysis, namely:
  • What each config file (data, model, and study) manages in the analysis.
  • How and when the data is split based on the data config file's contents.
  • How and when the hyper-parameters of the model are determined based on the model config file's contents.
  • How the resulting splits are used to train these models.
  • Where and when the performance metrics defined by the user are calculated based on the study config file's contents.
• Examples of how our data hooks modify the the data (see here).
• Examples of the how the metrics calculated for each model are calculated.

Feel free to suggest more below, and we can add them.

[SomeoneInParticular]

PR #26 is the first step towards this, though it is still WIP.

As of commit 5b4cfbf, the tutorial is mostly done though! Just need to add some discussion on how to read an interpret the results, which is very subject to change as we expand upon the tool.