Contents

Initial Setup & Maintenance 🛠️

Initial Setup & Maintenance 🛠️

The tech stack that I had to create and deploy this book was:

  • A Windows environment

  • PyCharm Community Edition

  • GitHub Desktop

Some mistakes I made in setup include:

  • Installing and running Jupyter Book on Python 3.9

  • Trying to install a g-zipped version of Python 3.7 to install the latest version (Python 3.7.12)

  • Navigating to the gh-pages branch

Some challenges I faced in setup include:

  • Deploying the sample book

  • Activating a virtual environment in PyCharm after navigating between Git branches

Let’s go through these mistakes in more detail!

Installing and running Jupyter Book on Python 3.9

The latest version of Python today is Python 3.10, but I routinely use Python 3.9 on my local machine.

I managed to install jupyter-book and successfully ran the command:

jupyter-book create mynewbook/

So imagine the surprise I experienced, following this tutorial when I ran

jupyter-book build book/

and got this problem.

I wish I started by reading the warning here and then navigating to Working on Windows

Lesson 1:

Learn how to set up Jupyter Book based on your OS environment.

Trying to install the latest version of Python 3.7

When I found that Python 3.7 was required, my natural inclination was to go here and zoom in to Python 3.7.12:

py1

It came as a .tgz which I unzipped on Administrator Windows cmd with the following code:

tar -xvzf C:\Users\CT\Downloads\Python-3.7.12.tgz -C C:\Users\CT\Downloads

When I unzipped it, I felt stuck because I couldn’t activate it without a crazy amount of searching.

There had to be a better way.

So I decided to search for versions of Python 3.7 that had a .exe file for Windows here settling on Python 3.7.9:

py2

I clicked the Download Windows x86-64 executable installer because my machine is 64-bit. I installed it with default installation instructions, and I was able to create a virtual environment venv in my project that ran on Python 3.7 as follows:

I changed to the main directory, BTS-Continuous-Learning then ran

virtualenv -p python3.7 venv

All that’s left is activating venv and checking that the correct version of Python is running with the code below:

py --version

Finally, in Github Desktop after committing the files created with

jupyter-book create mynewbook/

I added all venv files to .gitignore.

Lesson 2:

Use a .exe installer to install Python on your machine; avoid zipped files.

Activating a virtual environment in PyCharm

The first time (when the only Git branch that I had was main, it was easy enough to activate the virtual environment in PyCharm. However, when I navidated to the gh-pages branch in GitHub Desktop, I started getting a weird error in the terminal like

pycharm1

When I upgraded my PyCharm community Edition program, the error changed to this:

pycharm2

Following the suggestion, this happened:

pycharm3

So what was the solution?

I navigated first to the File menu, then clicked on Settings resulting in the following screen:

pycharm4

I then navigated to Tools, then Terminal, and changed the Shell path from Powershell’s path to cmd’s path like so:

pycharm5

Problem solved!

pycharm6

Work within the environment and clear the workspace with cls command.

Lesson 3:

Use a cmd engine within Pycharm instead of Powershell to activate the virtual environment.

Deployment of the site

Updating the book was easy enough with the jb build --all book/ command once the virtual environment was activated.

I then

  • Pushed any changes in book, _static and images folders on the main branch to remote

  • Installed ghp-import with the pip install ghp-import command

  • Changed into the directory containing the files with cd book/ command

  • Created a gh-pages branch (when the command ran the first time) or pushed the changes to the gh-pages branch with the ghp-import -n -p -f _build/html command

Warning:

When running the ghp-import -n -p -f _build/html command, DO NOT forget to add -n; the reason why is well explained here .

Initially, after deploying the page, I would navigate to the gh-pages branch, but I realised that it was not necessary since running the ghp-import -n -p -f _build/html command automatically sends local changes to remote in the gh-pages branch.


previous

Jupyter Books explored 📙

next

Customizing the Front End 🌼