This is a list of user-contributed tips for making virtualenv and virtualenvwrapper even more useful. If you have tip to share, drop me an email or post a comment on this blog post and I'll add it here.

Via Stephan Sokolow

While the virtualenv activate script does attempt to provide an indicator in the prompt, it has various shortcomings, and cannot be customized.

However, it does also set a shell variable named VIRTUAL_ENV which can be used as the basis for disabling the built-in prompt indicator and substituting an improved one, as a customization to .bashrc or .zshrc:

virtualenv_prompt() {
    # If not in a virtualenv, print nothing
    [[ "$VIRTUAL_ENV" == "" ]] && return

    # Distinguish between the shell where the virtualenv was activated
    # and its children
    local venv_name="${VIRTUAL_ENV##*/}"
    if typeset -f deactivate >/dev/null; then
        echo "[${venv_name}] "
    else
        echo "<${venv_name}> "
    fi
}

# Display a "we are in a virtualenv" indicator that works in child shells too
VIRTUAL_ENV_DISABLE_PROMPT=1
PS1='$(virtualenv_prompt)'"$PS1"

This basic example works in both bash and zsh and has the following advantages:

1. It will also display in sub-shells, because it works by having the shell detect an active virtualenv, rather than by having the activate script modify the prompt for just the current shell instance.
2. It will clearly indicate if you're in a subshell, where the virtualenv will still apply, but the deactivate command will be missing.

However, if you are using zsh, a better example of what the design is capable of can be constructed by taking advantage of zsh's built-in support for easily adding color and right-aligned segments to prompts:

zsh_virtualenv_prompt() {
    # If not in a virtualenv, print nothing
    [[ "$VIRTUAL_ENV" == "" ]] && return

    # Distinguish between the shell where the virtualenv was activated
    # and its children
    local venv_name="${VIRTUAL_ENV##*/}"
    if typeset -f deactivate >/dev/null; then
        echo "[%F{green}${venv_name}%f] "
    else
        echo "<%F{green}${venv_name}%f> "
    fi
}

setopt PROMPT_SUBST PROMPT_PERCENT

# Display a "we are in a virtualenv" indicator that works in child shells too
VIRTUAL_ENV_DISABLE_PROMPT=1
RPS1='$(zsh_virtualenv_prompt)'

From Nat (was blogger.com/profile/16779944428406910187):

I also added the command 'rehash' to $WORKON_HOME/postactivate and $WORKON_HOME/postdeactivate as I was having some problems with zsh not picking up the new paths immediately.

Via James:

In the postmkvirtualenv script I have the following to create a directory based on the project name, add that directory to the python path and then cd into it:

proj_name=$(basename $VIRTUAL_ENV)
mkdir $HOME/projects/$proj_name
add2virtualenv $HOME/projects/$proj_name
cd $HOME/projects/$proj_name

In the postactivate script I have it set to automatically change to the project directory when I use the workon command:

proj_name=$(basename $VIRTUAL_ENV)
cd ~/projects/$proj_name

Justin Abrahms posted about some code he added to his shell environment to look at the directory each time he runs cd. If it finds a .venv file, it activates the environment named within. On leaving that directory, the current virtualenv is automatically deactivated.

Via rizumu (was rizumu.myopenid.com):

I have this postmkvirtualenv to install the get a basic setup.

$ cat postmkvirtualenv
#!/usr/bin/env bash
curl -O http://python-distribute.org/distribute_setup.p... />python distribute_setup.py
rm distribute_setup.py
easy_install pip==dev
pip install Mercurial

Then I have a pip requirement file with my dev tools.

$ cat developer_requirements.txt
ipdb
ipython
pastescript
nose
http://douglatornell.ca/software/python/Nosy-1.0.tar.gz
coverage
sphinx
grin
pyflakes
pep8

Then each project has it's own pip requirement file for things like PIL, psycopg2, django-apps, numpy, etc.

Via mae:

This is supposed to be executed after workon, that is as a postactivate hook. It basically overrides cd to know about the VENV so instead of doing cd to go to ~ you will go to the venv root, IMO very handy and I can't live without it anymore. If you pass it a proper path then it will do the right thing.

cd () {
    if (( $# == 0 ))
    then
        builtin cd $VIRTUAL_ENV
    else
        builtin cd "$@"
    fi
}

cd

And to finally restore the default behaviour of cd once you bailout of a VENV via a deactivate command, you need to add this as a postdeactivate hook:

unset -f cd

Via Michael:

When you use a temporary virtualenv via mktmpenv or if you have a :ref:`plugins-post_deactivate` hook, you have to actually run deactivate to clean up the temporary environment or run the hook, respectively. It's easy to forget and just exit the shell. Put the following in ~/.bash_logout (or your shell's equivalent file) to always deactivate environments before exiting the shell:

[ "$VIRTUAL_ENV" ] && deactivate