Hi Julien,

there indeed are other ways of managing python on your system, and the best one is probably conda.


One of the most useful features of conda is that it separates your system python (packages installed by pacman as dependencies for other programs) and your user python (packages you want to use in your personal projects).

Conda can act as a package manager for your user python installation, so you can play around with different versions without messing up your system python.

Once you get conda set up, you can install scikit image with:

conda install scikit-image

Switching to a different way of managing your python installations may seem overwhelming at first, but I promise it will save you a lot of headache in the long run.

As an arch user myself, here are a few tips:

 * Let pacman handle your system python (never use pip as root)
 * Don't use AUR to install python packages (unless another AUR package
   depends on them)
 * Once you start using conda, you might encounter problems when
   building AUR packages. This is because AUR packages are built in
   your user environment, but they might require a system python
   dependency. If you encounter this, just remove conda directory from
   your PATH and try rebuilding the package.

Conda is a well established tool and you'll find plenty of documentation and tutorials for it online.
This page should be a good place to start:


Best of luck!


On 14/04/18 12:22, Julien Derr wrote:
Hi everyone,

Is there any arch linux user around ? I was wondering what is the most straight forward way to install skimage ?

I tried the AUR repository without success, then I tried PIP with no success neither is there a third way, or could you tell me which one you used, (In this case I will tell you the detail of my errors ...)

thanks a lot,


scikit-image mailing list

scikit-image mailing list

Reply via email to