[Scikit-learn-general] Building sklearn for different python versions (development)

Discussion:

WENDLINGER Antoine

2016-01-25 16:23:05 UTC

Hi everyone,

I am using conda in order to be able to test scikit-learn for different
python versions. The problem I have is whenever I build scikit-learn for
one version of python it messes up the installation of scikit learn for the
other version, and when I try to run tests I get the "It seems that
scikit-learn has not been built correctly." error message.
What do you guys use to be able to test scikit-learn on different python
versions ?

Regards,

Antoine

Jacob Vanderplas

2016-01-25 18:46:20 UTC

Permalink

Hi Antoine,
For this type of thing I use conda environments:
http://conda.pydata.org/docs/using/envs.html

The other thing to keep in mind is that if you're installing from the same
source directory, you'll need to do a clean install each time; i.e. type
``python setup.py clean`` before typing ``python setup.py install``.

Hope that helps,
Jake

Jake VanderPlas
Senior Data Science Fellow
Director of Research in Physical Sciences
University of Washington eScience Institute

On Mon, Jan 25, 2016 at 8:23 AM, WENDLINGER Antoine <

Post by WENDLINGER Antoine
Hi everyone,
I am using conda in order to be able to test scikit-learn for different
python versions. The problem I have is whenever I build scikit-learn for
one version of python it messes up the installation of scikit learn for the
other version, and when I try to run tests I get the "It seems that
scikit-learn has not been built correctly." error message.
What do you guys use to be able to test scikit-learn on different python
versions ?
Regards,
Antoine
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Scikit-learn-general mailing list
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

WENDLINGER Antoine

2016-01-25 19:56:29 UTC

Permalink

Hello Jake,

Thanks for your answer. I am already using conda, and having to clean and
rebuild everything everytime I want to switch versions is what I want to
avoid, since it's a bit long to accomplish. Using two different source
folders seems a bit silly since I would like to test changes I just made in
the code, and I don't see an easy way to maintain the changes in two
different directories.

Regards,

Antoine

Post by Jacob Vanderplas
Hi Antoine,
http://conda.pydata.org/docs/using/envs.html
The other thing to keep in mind is that if you're installing from the same
source directory, you'll need to do a clean install each time; i.e. type
``python setup.py clean`` before typing ``python setup.py install``.
Hope that helps,
Jake
Jake VanderPlas
Senior Data Science Fellow
Director of Research in Physical Sciences
University of Washington eScience Institute
On Mon, Jan 25, 2016 at 8:23 AM, WENDLINGER Antoine <

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Scikit-learn-general mailing list
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Andreas Mueller

2016-01-25 22:30:40 UTC

Permalink

Post by WENDLINGER Antoine
Hello Jake,
Thanks for your answer. I am already using conda, and having to clean
and rebuild everything everytime I want to switch versions is what I
want to avoid, since it's a bit long to accomplish. Using two
different source folders seems a bit silly since I would like to test
changes I just made in the code, and I don't see an easy way to
maintain the changes in two different directories.

WENDLINGER Antoine

2016-01-26 12:23:25 UTC

Permalink

Well I did not use travis because I thought it was a little cumbersome to
have to push every little change I made to my Github repo, plus a travis
build takes ~15 min. I was looking for a way to keep the binaries for both
versions with the same source directory (I don't do edit cython files for
the moment) so that i could run some tests for different python environment
without having to wait for the long build process to complete.
But if travis is the way to go i'll use it, thanks for your answers.

Regards,

Antoine

Post by Andreas Mueller

Well you can't really make changes and apply them to different versions.
What if the change was in a cython file? You need to recompile for each
version.
If the change was in a python file, you still need to have the binaries
for all the different python versions around.
Why don't you just use travis?
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Scikit-learn-general mailing list
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Jacob Vanderplas

2016-01-26 13:45:36 UTC

Permalink

I don't see an easy way to maintain the changes in two different

directories.

If both directories are Git repositories linked to a common remote, you
could commit the changes on a branch and then sync them that way.

Jake VanderPlas
Senior Data Science Fellow
Director of Research in Physical Sciences
University of Washington eScience Institute

On Tue, Jan 26, 2016 at 4:23 AM, WENDLINGER Antoine <

Post by Andreas Mueller

WENDLINGER Antoine

2016-01-26 15:55:36 UTC

Permalink

I turns out that using "python setup.py develop" while using conda did
exactly what I wanted (it created the build files in to separate
directories), I was convinced I tried that earlier, sorry about that.

Post by WENDLINGER Antoine

I don't see an easy way to maintain the changes in two different

directories.
If both directories are Git repositories linked to a common remote, you
could commit the changes on a branch and then sync them that way.
Jake VanderPlas
Senior Data Science Fellow
Director of Research in Physical Sciences
University of Washington eScience Institute
On Tue, Jan 26, 2016 at 4:23 AM, WENDLINGER Antoine <

Post by Andreas Mueller