This is what works for me currently. Note that the direnv wiki addresses various methods for managing python with direnv: https://github.com/direnv/direnv/wiki/Python

1. Install direnv and pyenv. I use homebrew for these tasks. There may be setup and/or shell integration steps for these tools that I'm omitting (I can't remember), so check their docs.
2. Determine the python version(s) you will likely be using and install them with pyenv at the command line (i.e., in Terminal).
   • pyenv versions will show you the versions already locally installed.
   • pyenv install --list will show you all the versions available to install.
   • pyenv install 3.12.4 will install that particular version for use on your local system.
3. Using the Terminal, move to or create and move to a directory that will contain your python project (package or script or whatever).
4. Configure the directory using direnv. To specify the python version you want to use as the virtual environment for python work in that directory (and its children), create a .envrc file and put the line layout pyenv 3.12.5 (or whatever python version you need) and save the file.
5. Follow the direnv prompt at the command line to "Run direnv allow to approve its content."
6. The first time you do this, direnv will invoke pyenv to set up a python virtual environment inside a .direnv subfolder it creates for this purpose.
7. Subsequently, when you cd into this directory, direnv will activate the virtual environment. When you cd out of this directory, direnv will deactivate the virtual environment.
8. When in the directory with the virtual environment activated, you can add packages and do updates in the normal ways, e.g., pip install -U pip.
9. Recent versions of Visual Studio Code correctly detect and use your direnv-managed virtual environment if started from the command line in the directory using code ..