I don't understand the idea that one doesn't lose one's scripts when using virtualenv?

The virtual environment created by virtualenvprovides a mechanism to change the order of directories searched to find files to execute. Without the virtualenv, all files executed would come from the global system directories. Therefore, only files placed into the venv directory through pip install are at risk of being removed when a venv directory is deleted. Your code created and held in other directories is independent of the venv directory or the whether it is "activated" or not.

Let's assume you create a virtualenv using virtualenv venv. This command simply creates a directory to hold specific versions of executables: python, pip, site-packages, etc. This directory, by itself, means nothing.

When the .\venv\Scripts\activate (windows) or the . .\venv\bin\activate (linux) is run, these directories are added to the beginning of the search path so these files are found before system-level files.

Typically, the additional files in the venv directory come from running pip install while the venv is activated, meaning that the venv directory is in the search path, causing the venv version of pip to run which knows to install the modules in the venv site-packages directory. This is where they will be found when running the venv version of python.

Your local app.py, Django, Flask, or other ..py files don't know about the venv, it is the command line interface that is working the magic via the path variables.

Therefore, you don't put your own code in the venv directory. Instead you place your code in what ever development directory you wish as if there wasn't a venv.

Finally, using activate and deactivate only determines whether the local venv versions or the system-level versions of files are run.