Ralf Gommers | 1 Sep 12:36 2014
Picon

EuroSciPy sprint report

Hi all,

We had a great sprint at EuroSciPy yesterday, with about 20 people working on Scipy and about 25 more on scikit-learn, scikit-image and Vispy. The result at the end of the day for Scipy was:

  - 24 pull requests submitted
  - 20 pull requests merged
  - 17 issues closed
  - 2 new issues created

A couple of highlights:

  - Fred Ludlow implemented bitwise operations on sparse matrices (PR 3941)
  - Julian Taylor implemented releasing the GIL from filter and interpolation functions in ndimage (PR 3943)
  - Aaren O'Leary implemented API changes needed before integrating discrete wavelets into scipy (issue 3931).
  - Chris Kerr removed linpack_lite from scipy.integrate (PR 3916). (done right before the sprint actually, but removing 15000 lines of Fortran code is worth a mention anyway!).

It was great to see a lot of new contributors as well as meet people who have been contributing for a while face-to-face. Thanks a lot to everyone who contributed, and especially to Pietro for the excellent organization!

Cheers,
Ralf


_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Moritz Beber | 29 Aug 12:13 2014
Picon

nested setup.py scripts

Dear all,

I want to generate a package with a submodule structure similar to what numpy and scipy use. (Or do you recommend not doing that?)  I have read the following pieces of documentation but I'm still unclear about how the main setup.py script discovers the nested scripts and gets the configuration values from those. Is this documented somewhere or can anyone point me to how this is done?

Thank you in advance,
Moritz

P.S.: What I've read:
https://github.com/numpy/numpy/blob/master/doc/DISTUTILS.rst.txt
http://docs.scipy.org/doc/scipy-dev/reference/hacking.html
http://docs.scipy.org/doc/scipy/reference/api.html
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Sri Krishna | 26 Aug 14:36 2014
Picon

To use C code or Cython code?

Hi,

I'm new to the Scipy-Dev mailing list, looking to contribute wherever I can. I was looking through the open issues and saw this issue, regarding a speed-up for the convolve2d function.

My confusion arises from the SciPy coding guidelines which states that using Cython is much preferable to using plain C/C++/Fortran.

Would it be desirable then to change the C code of signal/firfilter.c to a Cythonized code?

Thanks,
Krishna

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Jeff Grasty | 26 Aug 00:55 2014
Picon

Nyquist Filters

Hi,

One of the features that I have found missing in SciPy are functions to design nyquist and root-nyquist
filters, such as raised cosine and root-raised cosine filters.  I have written several functions for this
purpose and was curious if anyone thought was a greater need for this. 

Thanks,
Jeff
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Ralf Gommers | 19 Aug 00:20 2014
Picon

sprint <at> EuroSciPy, Aug 31

Hi all,

Here is a reminder that on Sunday 31 August, there will be a Scipy sprint at EuroSciPy (in Cambridge, UK). Details can be found at https://www.euroscipy.org/2014/program/sprints/

Newcomers to Scipy development are very welcome; actually one of the main goals of the sprint is to help new people to get started. Last year's sprint was excellent - 20 people joined and we still have all-time highs in the commits per month and contributors per month graph to show for it: https://www.openhub.net/p/scipy

If you have time and will be at EuroSciPy: please join!

Cheers,
Ralf

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Johann Goetz | 15 Aug 16:59 2014
Picon

Histogram as its own class

Hello,
I'm a long-time user of scipy doing mostly multivariate big-data (several terabytes) analysis in the high-energy physics realm. One thing I've found useful was to promote the histogram to it's own class. Instead of creating yet another package, I have a mind to include it into the scipy.stats module and I would like some feed-back. I.e. is this the right place for such an object?

I have some documentation, but not enough I would say, and the classes are currently buried in my "pyhep" project, but they are easily extracted out.

https://bitbucket.org/theodoregoetz/pyhep/wiki/Home

Here are some details:

The histograms I am addressing are N-dimensional over a continuous-domain (floating-point data, no gaps - though bins can have value inf or nan if need-be) along each axis. The axes need not be uniform.

There are two classes: HistogramAxis and Histogram. The Axes are always floating point, but the histogram's data can be any dtype (default: np.int, a "cast" to float is done when dividing two histograms). I make use of np.histogramdd() and store the data along with the uncertainty. Many operations are supported including adding, subtracting, multiplying, dividing, bin-merging, cutting/clipping along one or more axes, projecting along an axis, iterating over an axis, filling from a sample with or without weights.

Most of power in this package is in the fitting method of the histogram which makes use of scipy.curve_fit(). It handles missing data (when a bin is inf or nan), can include the uncertainty in the fit, and calculates a goodness of fit.

On top of this, I have free functions to plot 1D and 2D histograms using matplotlib, as well as functions to handle reading in large HDF5 files. These are auxiliary and may not fit into scipy directly.

Thank you all,
Johann. 
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Moritz Beber | 13 Aug 17:08 2014
Picon

Proposal for a new function nanpdist that treats NaNs as missing values

Dear all,

As suggested in this github issue (https://github.com/scipy/scipy/issues/3870), I would like to discuss the merit of introducing a new function nanpdist into scipy.spatial. I have also brought up the problem in the following previous e-mail (http://comments.gmane.org/gmane.comp.python.scientific.devel/18956) and on SO (http://stackoverflow.com/questions/24781461/compute-the-pairwise-distance-in-scipy-with-missing-values).

Warren suggested three ways to tackle this problem:
  1. Don't change anything--the users should clean up their data!
  2. nanpdist
  3. Add a keyword argument to pdist that determines how nan should be treated.

Clearly, I don't favor the first option since I believe missing values can be important pieces of information, too. I slightly tend towards option two because adding a keyword will further complicate an already very long pdist function.

I'm happy to submit a pull request if there is a consensus that something should be done.

Best,

Moritz

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Manoj Kumar | 11 Aug 17:04 2014
Picon

Fastest way to multiply a sparse matrix with another numpy array

Hello,

I was wondering what is the fastest way (format) to multiply a sparse matrix with a numpy array. Intuitively, a csr format multiplied with a numpy array which is fortran contiguous seems to be the fastest, but I have ran a few benchmarks and it seems otherwise. It is also mentioned here
http://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.sparse.csc_matrix.html that using csr matrices "may" be faster.


In [5]: X Out[5]: <11314x130107 sparse matrix of type '<type 'numpy.float64'>' with 1787565 stored elements in Compressed Sparse Row format> In [6]: _, n_features = X.shape In [9]: w_c = np.random.rand(n_features, 10) In [10]: w_f = np.asarray(w_c, order='f') In [13]: csc = sparse.csc_matrix(X) In [30]: %timeit X * w_f 10 loops, best of 3: 40.5 ms per loop In [31]: %timeit X * w_c 10 loops, best of 3: 37.3 ms per loop In [32]: %timeit csc * w_c 10 loops, best of 3: 24.3 ms per loop In [33]: %timeit csc * w_f 10 loops, best of 3: 27.3 ms per loop
It seems here, using a csc matrix is faster with a C-contiguous numpy array which is completely non-intuitive to me. Are there any hard rules for this? or is it data dependent?

Sorry for my noobish questions!
--
Regards,
Manoj Kumar,
GSoC 2014, Scikit-learn
Mech Undergrad
http://manojbits.wordpress.com
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Sai Rajeshwar | 27 Jul 10:28 2014
Picon

convolution using numpy/scipy using MKL libraries

hi all,

   Im trying to implement 3d convolutional networks.. for which I wanted to use convolve function from scipy.signal.convolve or fftconvolve..  but looks like both of them doesnot use MKL libraries..  is there any implementation of convolutoin which uses MKL libraries or MKL-threaded  so that code runs faster.

thanks a lot in advance

with regards..

M. Sai Rajeswar
M-tech  Computer Technology
IIT Delhi
----------------------------------Cogito Ergo Sum---------
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
nicky van foreest | 23 Jul 16:40 2014
Picon

scipy.sparse versus pysparse

Hi, 

I am doing some testing between scipy.sparse and pysparse on my ubuntu machine. Some testing reveals that pysparse is about 9 times faster in matrix-vector multiplication that scipy.sparse. Might there be anything specific I forgot to do during scipy's installation (I just ran apt-get install python-scipy)? Is there another simple explanation for this difference? I prefer to use scipy.sparse for its cleaner api, but a factor 9 in speed is considerable. 

thanks 

Nicky
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev
Johannes Kulick | 21 Jul 11:06 2014
Picon

Pull Request: Dirichlet Distribution

Hi,

I sent a pull request, that implements a Dirichlet distribution. Code review
would be appreciated!

https://github.com/scipy/scipy/pull/3815

Best,
Johannes Kulick

--

-- 
Question: What is the weird attachment to all my emails?
Answer:   http://en.wikipedia.org/wiki/Digital_signature
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev <at> scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-dev

Gmane