A simpler tool for creating venvs in a central location
mkenv is a simpler tool for creating and maintaining virtualenvs in a central location.
It consists of the 5% of virtualenvwrapper that I actually use, and is in some ways meant to more closely complement virtualenv rather than completely wrapping or hiding it.
$ pip install mkenv
Usage is similar to mkvirtualenv, although mkenv passes arguments directly through to virtualenv:
$ mkenv nameofvenv -- -p pypy
will create a virtual environment in an appropriate platform-specific data directory, or in the directory specified by WORKON_HOME for compatibility with virtualenvwrapper.
A common use case for virtualenvs is for single-purpose installations, e.g.:
“I want to install fabric and give it its own virtualenv so that its dependencies can be independently upgraded, all while still being able to use the fab binary globally”.
mkenv supports a --link option for this use case:
$ mkenv -i fabric --link fab
will create a virtualenv for fabric (in the same normal location), but will symlink the fab binary from within the virtualenv into your ~/.local/bin directory.
(You may have heard of pipsi which is a similar tool for this use case, but with less customization than I would have liked.)
I also find mktmpenv useful for quick testing. To support its use case, mkenv currently supports a different but similar style of temporary virtualenv.
$ venv=$(mkenv -t)
in your shell will create (or re-create) a global temporary virtualenv, and print its bin/ subdirectory (which in this case will be then stored in the venv variable). It can subsequently be used by, e.g.:
$ $venv/pip ...
You may prefer using:
$ cd $(mkenv -t)
as your temporary venv workflow if you’re into that sort of thing instead.
The global virtualenv is cleared each time you invoke mkenv -t. Unless you care, unlike virtualenvwrapper’s mktmpenv, there’s no need to care about cleaning it up, whenever it matters for the next time, it will be cleared and overwritten.
mkenv may support the more similar “traditional” one-use virtualenv in the future, but given that it does not activate virtualenvs by default (see below), the current recommendation for this use case would be to simply use the virtualenv binary directly.
The 5 Minute Tutorial
Besides the mkenv for named-virtualenv creation and mkenv -t for temporary-virtualenv creation described above:
$ findenv name foo
will output (to standard output) the path to a virtualenv with the given name (see also --existing-only), and:
$ rmenv foo
will remove it.
There are a number of other slight variants, see the --help information for each of the three binaries.
Real documentation to come (I hope)
Why don’t I use virtualenvwrapper?
virtualenvwrapper is great! I’ve used it for a few years. But I’ve slowly settled on a much smaller subset of its functionality that I like to use. Specifically:
I don’t like activating virtualenvs.
virtualenvs are magical and hacky enough on their own, and piling activation on top just makes things even more messy for me, especially when moving around between different projects in a shell. Some people use cd tricks to solve this, but I just want simplicity.
I don’t need project support.
I’ve never attached a project to a virtualenv. I just use a naming convention, naming the virtualenv with the name of the repo (with simple coersion), and then using dynamic directory expansion in my shell to handle association.
Basically, I just want a thing that is managing a central repository of virtualenvs for me. So that’s what mkenv does.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|File Name & Checksum SHA256 Checksum Help||Version||File Type||Upload Date|
|mkenv-2.3.1-py2-none-any.whl (17.7 kB) Copy SHA256 Checksum SHA256||2.7||Wheel||Jun 16, 2017|
|mkenv-2.3.1.tar.gz (10.8 kB) Copy SHA256 Checksum SHA256||–||Source||Jun 16, 2017|