New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Conda install #1052

Merged
merged 7 commits into from Mar 23, 2017

Conversation

Projects
None yet
4 participants
@arokem
Member

arokem commented May 18, 2016

This is now possible, thanks to conda-forge: https://github.com/conda-forge/dipy-feedstock

@arokem

This comment has been minimized.

Member

arokem commented May 18, 2016

As a side note, once the conda recipe for VTK on Python 3 is ready (conda-forge/staged-recipes#453), we will be fully functional on Python 3!

@arokem arokem added this to the 0.12 milestone May 21, 2016

#. Some of the visualization methods require the VTK_ library and this can be installed separately (but only on Python 2)::

conda install vtk

This comment has been minimized.

@Garyfallidis

Garyfallidis May 21, 2016

Member

Why only on python 2? Still they haven't updated their vtk to vtk 7?

This comment has been minimized.

@Garyfallidis

Garyfallidis May 21, 2016

Member

Ah okay I saw your message. There is a PR that adds it soon.
Do you know how easy it is to move from conda-forge to the actual conda? For example for vtk you don't need to do conda config --add channels conda-forge. You can just do conda install vtk. Is it like with neurodebian? For example you start adding packages in neurodebian and then they are moved when ready to debian?

This comment has been minimized.

@arokem

arokem May 23, 2016

Member

It might be that there are some advantages to not needing to specify the
channel to install from, but this has some advantages as well. Especially
with regards to maintenance of the recipe, because it's all done on the CI
services (see here for details: https://conda-forge.github.io/).

On Sat, May 21, 2016 at 4:45 PM, Eleftherios Garyfallidis <
notifications@github.com> wrote:

In doc/installation.rst
#1052 (comment):

@@ -21,6 +21,20 @@ If you are on Debian or Ubuntu Linux we recommend you try

.. _install-packages:

+Using Anaconda:
+===============
+
+On all platforms, you can use Anaconda_ to install Dipy. To do so issue the following two commands in a terminal::
+

  • conda config --add channels conda-forge
  • conda install dipy

+#. Some of the visualization methods require the VTK_ library and this can be installed separately (but only on Python 2)::
+

  • conda install vtk

Ah okay I saw your message. There is a PR that adds it soon.
Do you know how easy it is to move from conda-forge to the actual conda?
For example for vtk you don't need to do conda config --add channels
conda-forge. You can just do conda install vtk. Is it like with
neurodebian? For example you start adding packages in neurodebian and then
they are moved when ready to debian?


You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
https://github.com/nipy/dipy/pull/1052/files/d75ae785855126af805a6c47683115f737f80331#r64141989

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented May 23, 2016

Hi @arokem. Don't we need to add a travis bot to test this? And possibly buildbots too in different OSes? Or Anaconda has its own systems for testing the projects?

@arokem arokem force-pushed the arokem:conda-install branch 2 times, most recently from 07d925a to dc75252 May 24, 2016

@arokem

This comment has been minimized.

Member

arokem commented May 24, 2016

Testing this on travis here doesn't make sense, because the conda install only changes when we cut a release. We can add testing on the feedstock itself, if we want that. Might be a good idea.

@arokem

This comment has been minimized.

Member

arokem commented May 24, 2016

After some bungling about, I created another repository to do conda installations from, and test those. It's still somewhat work in progress, but there are already some issues we'll need to address. Primarily this one #852. If we had some reliable way to ascertain whether the python being used is the Anaconda python, we could fall back to the implementation we have here: https://github.com/nipy/dipy/blob/master/dipy/reconst/shm.py#L160-L166

@arokem

This comment has been minimized.

Member

arokem commented May 24, 2016

Oh - and the repo location: https://github.com/arokem/dipy-tester

@arokem

This comment has been minimized.

Member

arokem commented Oct 21, 2016

Any reason not to merge this one?

conda config --add channels conda-forge
conda install dipy

#. Some of the visualization methods require the VTK_ library and this can be installed separately (but only on Python 2)::

This comment has been minimized.

@Garyfallidis

Garyfallidis Nov 3, 2016

Member

Here @arokem adds a number at the beginning of the line. Check the View button and you will understand what I mean. Also you may want to say that with VTK 7 python 3 will be supported too.

@arokem

This comment has been minimized.

Member

arokem commented Nov 3, 2016

Yep. Removed number (it doesn't make sense to number only one item 😁), and removed the python 2 comment.

@arokem arokem force-pushed the arokem:conda-install branch from cce682c to 1ced507 Nov 3, 2016

@arokem

This comment has been minimized.

Member

arokem commented Nov 3, 2016

And rebased

@coveralls

This comment has been minimized.

coveralls commented Nov 3, 2016

Coverage Status

Coverage remained the same at 87.982% when pulling 1ced507 on arokem:conda-install into 7302303 on nipy:master.

@coveralls

This comment has been minimized.

coveralls commented Nov 3, 2016

Coverage Status

Coverage remained the same at 87.982% when pulling 1ced507 on arokem:conda-install into 7302303 on nipy:master.

@codecov-io

This comment has been minimized.

codecov-io commented Nov 3, 2016

Codecov Report

Merging #1052 into master will increase coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1052      +/-   ##
==========================================
+ Coverage    85.8%   85.82%   +0.01%     
==========================================
  Files         215      217       +2     
  Lines       26001    26045      +44     
  Branches     2668     2671       +3     
==========================================
+ Hits        22310    22352      +42     
- Misses       3038     3040       +2     
  Partials      653      653
Impacted Files Coverage Δ
dipy/workflows/tests/test_masking.py 91.66% <ø> (ø)
dipy/workflows/mask.py 95% <ø> (ø)
dipy/io/image.py 87.5% <ø> (+6.25%)

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 658f6ab...a258281. Read the comment docs.

@coveralls

This comment has been minimized.

coveralls commented Nov 7, 2016

Coverage Status

Coverage remained the same at 87.982% when pulling 96c2e41 on arokem:conda-install into 7302303 on nipy:master.

@coveralls

This comment has been minimized.

coveralls commented Nov 7, 2016

Coverage Status

Coverage remained the same at 87.982% when pulling 96c2e41 on arokem:conda-install into 7302303 on nipy:master.

On all platforms, you can use Anaconda_ to install Dipy. To do so issue the following two commands in a terminal::

conda config --add channels conda-forge
conda install dipy

This comment has been minimized.

@Garyfallidis

Garyfallidis Dec 20, 2016

Member

What about just writing only one command?

conda install dipy -c conda-forge

And then when dipy gets linked in the main conda repo we can just remove the last two words.


Some of the visualization methods require the VTK_ library and this can be installed separately (for the time being only on Python 2.7)::

conda install vtk

This comment has been minimized.

@Garyfallidis

Garyfallidis Dec 20, 2016

Member

And if we do conda install dipy first do we need to install vtk too? Or it get's directly installed?

This comment has been minimized.

@Garyfallidis

Garyfallidis Dec 20, 2016

Member

Interestingly when I tried to install dipy in Windows it asked me to upgrade vtk.

@coveralls

This comment has been minimized.

coveralls commented Dec 21, 2016

Coverage Status

Coverage increased (+0.04%) to 88.025% when pulling a10017d on arokem:conda-install into 7302303 on nipy:master.

@arokem

This comment has been minimized.

Member

arokem commented Dec 21, 2016

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Dec 21, 2016

The following NEW packages will be INSTALLED:

    dipy:      0.11.0-np111py27_0 conda-forge
    future:    0.15.2-py27_0      conda-forge
    nibabel:   2.0.2-py27_vc9_0   conda-forge [vc9]
    pydicom:   0.9.9-py27_0       conda-forge

The following packages will be UPDATED:

    conda:     4.2.11-py27_0                  --> 4.2.13-py27_0    conda-forge
    vtk:       6.3.0-py27_1                   --> 7.0.0-py27_vc9_1 conda-forge [vc9]
@arokem

This comment has been minimized.

Member

arokem commented Dec 21, 2016

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Dec 21, 2016

It didn't complain.

@arokem

This comment has been minimized.

Member

arokem commented Dec 21, 2016

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Feb 17, 2017

@arokem I believe we can merge this one. Correct?

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

@arokem arokem force-pushed the arokem:conda-install branch from a10017d to a258281 Feb 17, 2017

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Feb 17, 2017

Or we still need to check issues in specific OS?

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

Now rebased, for your convenience

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

Windows is still on and off broken.

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

But that's not an issue with this PR.

@Garyfallidis

This comment has been minimized.

Member

Garyfallidis commented Feb 17, 2017

So, this is on the Anaconda side?

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

@arokem

This comment has been minimized.

Member

arokem commented Feb 17, 2017

@coveralls

This comment has been minimized.

coveralls commented Feb 17, 2017

Coverage Status

Coverage increased (+0.01%) to 88.318% when pulling a258281 on arokem:conda-install into 658f6ab on nipy:master.

@Garyfallidis Garyfallidis merged commit 02ce52f into nipy:master Mar 23, 2017

4 checks passed

codecov/patch Coverage not affected when comparing 658f6ab...a258281
Details
codecov/project 85.82% (+0.01%) compared to 658f6ab
Details
continuous-integration/travis-ci/pr The Travis CI build passed
Details
coverage/coveralls Coverage increased (+0.01%) to 88.318%
Details

ShreyasFadnavis pushed a commit to ShreyasFadnavis/dipy that referenced this pull request Sep 20, 2018

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment