agencyenterprise / analyzingneuraltimeseries-python Goto Github PK

The original Matlab code from Mike X Cohen has a bunch of useful comments about what's happening in the code. Some of these are Matlab-specific and not useful for the Python translation, but others can be translated to useful Python-specific comments or are language-agnostic. I believe some comments have already made their way into this Python translation, but it is worth going through the Matlab code and adding any useful comments that are missing.

The package nbdev that we use to create the GitHub-pages site only renders the .ipynb notebooks by default. This covers all the chapter code, but there are some functions in .py files that it would also be nice to have available on the GitHub-pages site. This would either involve checking if .py files can be rendered with nbdev or adding a notebook to show all of those functions (or an additional notebook for each function).

In some chapters, some figures reference figures from other chapters. If nbdev has this functionality, it would be nice to add links to those referenced figures. We could also potentially add previous/next chapter buttons to each notebook.
Ex. in chapter 33, there is this placeholder text for figures 33.5/6 that are in chapter 34:

• These figures are generated in the code for chapter 34
  It would be nice to have a link there to take you to those figures.

It is likely people might come to this repository with very little or no Python experience. We would like to make any setup for them to run these notebooks themselves as simple and painless as possible. Adding documentation on how to get and run these notebooks as part of the readme or as a simple installation page would be extremely helpful in this regard (ex. NDK's installation instructions).

This Python translation was mainly done using the Matlab code as reference. Could be useful to compare this implementation's code and figures against the book's text and figures to make sure that it makes sense in that context. Additionally, checking against the other implementations' code/figures could also be beneficial.