The original reason for suggesting making Pillow a requirements is that during the SciPy conference sprints Juan and I found that the unit tests were failing when only the modules required for scikit-image were installed. I think parts of this should be fixed by Pull Request #1060 [1] which will make the matplotlib related tests skippable. This was compounded by the fact that the version of Pillow Continuum's provides for windows is broken, and the version of QT they link matplotlib to has some issues. Let me do some additional testing and see if this is still an issue, if it is I'll open a GitHub issue. If anyone is interested there is an MIT licensed pure Python PNG reader at: https://github.com/drj11/pypng. Cheers, - Jonathan Helmus [1] https://github.com/scikit-image/scikit-image/pull/1060 On 07/16/2014 05:25 PM, Juan Nunez-Iglesias wrote:

@stefanv, despite the possibility of getting imread included in Anaconda, I would argue that it's better to go with the most widely supported package by default. Will you also campaign to have it included in Canopy, NeuroDebian, Python(x,y), etc.?

@jjhelmus and I also looked into grabbing BSD-licensed, pure Python readers for TIFF, PNG, and JPEG, and including them in our code, to fall back on when none of the plugins are available. As I recall we found a candidate for PNG but I don't remember continuing on the search for a JPEG reader...

Patrick, thanks for the input! Hopefully the Pillow package will be updated in Anaconda soon! I would not be happy shipping a release with a vaguely-supported workaround as a dependency.

On Wed, Jul 16, 2014 at 10:36 AM, Patrick Snape <patricksnape@gmail.com <mailto:patricksnape@gmail.com>> wrote:

There is currently a bug in the Anaconda repositories whereby Anaconda ship, by default, a package they call 'imaging'. This seems to be some custom version of Pillow, as it still namespaces the PIL Image class. Pillow, on the other hand, is incorrectly linked against the libraries that Anaconda ships on Windows, and so lots of things will fail. I've opened this bug here:

https://github.com/ContinuumIO/anaconda-issues/issues/30#issuecomment-424669...

I've resolved this at the moment by requiring the `imaging` package instead of `pillow` within my conda build script (for a project I maintain).

On Friday, 11 July 2014 17:31:39 UTC+1, Juan Nunez-Iglesias wrote:

Hey Stefan,

Pillow is in conda, imread is not. My experimentation over the last 15 minutes suggests that it will be much more of a challenge to install imread than pillow. (I'm also not sure whether imread supports Python 3 and that's why things are failing.)

I also just saw @jjhelmus's email so maybe PIL should throw its hat into the ring...

On Fri, Jul 11, 2014 at 10:44 AM, Stéfan van der Walt <ste...@sun.ac.za> wrote:

Hi Juan

On Fri, Jul 11, 2014 at 5:32 PM, Juan Nunez-Iglesias <jni....@gmail.com> wrote: > Cheap solution: make Pillow a requirement.

My suggestion last time was to make 'imread' a requirement--what do you think?

Stéfan

-- You received this message because you are subscribed to the Google Groups "scikit-image" group. To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

-- You received this message because you are subscribed to the Google Groups "scikit-image" group. To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image+unsubscribe@googlegroups.com <mailto:scikit-image+unsubscribe@googlegroups.com>. For more options, visit https://groups.google.com/d/optout.

-- You received this message because you are subscribed to the Google Groups "scikit-image" group. To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image+unsubscribe@googlegroups.com <mailto:scikit-image+unsubscribe@googlegroups.com>. For more options, visit https://groups.google.com/d/optout.