19  Setup Git

At some point I wanted to test the script on my local machine before running it in Colab (see next chapter and section) but copying the files was redundant and prone to errors.
In addition I discovered that once I shared my notebook with some else directly from Colab all outputs will be lost.

The solution to these problems I arrived was using Git, a distributed version control system.

Firstly, I opened a GitHub repository and from my local machine I uploaded the preprocessing and the input data.
On the VM the process is slightly more difficult:

1. I setup my credentials and path to GitHub. Credentials are to used only during the push to GitHub

GIT_USERNAME = "dado330"
GIT_MAIL = "messina.davide.statistician@gmail.com"
GIT_REPO = "https://github.com/dado330/AE-WGAN-GP"

2. I cloned the repository on the VM. I didn’t need any form of authentication since the repo is public.

!git clone $GIT_REPO

3. I inserted a cell to pull changes I’ve tested locally.

%%bash
cd /content/AE-WGAN-GP
git pull origin main

4. This cell makes Git aware of our credentials

!git config --global user.email $GIT_MAIL
!git config --global user.name $GIT_USERNAME

5. To push changes we need a token. You can easily generate one in your GitHub account with at least Repo scope. (For example: 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).
   Then set the origin where to push the commits to a combinations of token and repository link as below.

%%bash
cd /content/AE-WGAN-GP
git remote set-url origin "https://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@github.com/dado330/AE-WGAN-GP.git"

6. Here I stage all changes. Commit them with a message and then push the commit to the origin defined previously.

%%bash
cd /content/AE-WGAN-GP
git add -A
git commit -m "Updated comments and generated data"
git push -u origin main

The Jupyter notebook was saved on the repository using the internal options of Google Colab.

Now all files related to the projects are publicly accessible and included the notebook’s outputs.