pandas2numpy VS yadm

I wrote pandas2numpy [0] it helps you turn a pandas dataframe into one (or two if you have categorical variables) numpy tensors ready to be pushed into a deep learning framework.

It can deal with continuous data, categorical data, normalisation of the data, log scale and NA.

I found it really useful to work on tabular data and, since it uses numpy, it is agnostic to the deep-learning framework used.

[0]: https://github.com/nestordemeure/pandas2numpy

Everyone hand-rolls their own dotfile management system, but YADM already does everything you need:

https://yadm.io/

I've been working around this using tools built on top of git like [yadm](https://github.com/TheLocehiliosan/yadm) and relying on `ls-files` to list all my tracked dotfiles and their paths.

Still having everything in one place would make things much simpler. Great idea!

I wonder if the program i use to manage my dotfiles could help manage your scripts and extend your setup to all your desktops? Its called yadm (https://yadm.io/) it makes it so easy to have a laptop and a desktop or two.

I really like that one but still prefer yadm because you can just edit your files as usual and then yadm add them wherever you are.

If you haven’t already, this is the time to install a tool like yadm and get your computer configuration into version control. Your command-line tools can be managed by yadm directly, your system settings can mostly be managed with a yadm bootstrap script that runs things like defaults write, and the software you install can be managed with a Brewfile that the yadm bootstrap script uses to install software with Homebrew. Don’t manually download Xcode, use xcodes to do it.

I personally use YADM. It's basically a git repo on my home folder, that only tracks what I explicitly set. And you can setup bootstraps to do what you said, install a bunch of stuff or make custom changes. In it's essence, it's a set of bash/sh files that are executed sequentially when you launch the yadm bootstrap command.