Mailman 3 March 2010 - NumPy-Discussion

subclassing ndarray in python3
by Darren Dale March 11, 2010

March 11, 2010

Now that the trunk has some support for python3, I am working on making Quantities work with python3 as well. I'm running into some problems related to subclassing ndarray that can be illustrated with a simple script, reproduced below. It looks like there is a problem with the reflected operations, I see problems with __rmul__ and __radd__, but not with __mul__ and __add__: import numpy as np class A(np.ndarray): def __new__(cls, *args, **kwargs): return np.ndarray.__new__(cls, *args, **kwargs) class B(A): def __mul__(self, other): return self.view(A).__mul__(other) def __rmul__(self, other): return self.view(A).__rmul__(other) def __add__(self, other): return self.view(A).__add__(other) def __radd__(self, other): return self.view(A).__radd__(other) a = A((10,)) b = B((10,)) print('A __mul__:') print(a.__mul__(2)) # ok print(a.view(np.ndarray).__mul__(2)) # ok print(a*2) # ok print('A __rmul__:') print(a.__rmul__(2)) # yields NotImplemented print(a.view(np.ndarray).__rmul__(2)) # yields NotImplemented print(2*a) # ok !!?? print('B __mul__:') print(b.__mul__(2)) # ok print(b.view(A).__mul__(2)) # ok print(b.view(np.ndarray).__mul__(2)) # ok print(b*2) # ok print('B __add__:') print(b.__add__(2)) # ok print(b.view(A).__add__(2)) # ok print(b.view(np.ndarray).__add__(2)) # ok print(b+2) # ok print('B __rmul__:') print(b.__rmul__(2)) # yields NotImplemented print(b.view(A).__rmul__(2)) # yields NotImplemented print(b.view(np.ndarray).__rmul__(2)) # yields NotImplemented print(2*b) # yields: TypeError: unsupported operand type(s) for *: 'int' and 'B' print('B __radd__:') print(b.__radd__(2)) # yields NotImplemented print(b.view(A).__radd__(2)) # yields NotImplemented print(b.view(np.ndarray).__radd__(2)) # yields NotImplemented print(2+b) # yields: TypeError: unsupported operand type(s) for +: 'int' and 'B'

2 2

multiprocessing shared arrays and numpy
by Jesper Larsen March 11, 2010

March 11, 2010

Hi people, I was wondering about the status of using the standard library multiprocessing module with numpy. I found a cookbook example last updated one year ago which states that: "This page was obsolete as multiprocessing's internals have changed. More information will come shortly; a link to this page will then be added back to the Cookbook." http://www.scipy.org/Cookbook/multiprocessing I also found the code that used to be on this page in the cookbook but it does not work any more. So my question is: Is it possible to use numpy arrays as shared arrays in an application using multiprocessing and how do you do it? Best regards, Jesper

6 21

Calling routines from a Fortran library using python
by Nils Wagner March 11, 2010

March 11, 2010

Hi all, I have a static library (*.a) compiled by gfortran but no source files. How can I call routines from that library using python ? Any pointer would be appreciated. Thanks in advance. Nils

8 33

numarray iterator question
by Neal Becker March 11, 2010

March 11, 2010

This is a bit confusing to me: import numpy as np u = np.ones ((3,3)) for u_row in u: u_row = u_row * 2 << doesn't work print u [[ 1. 1. 1.] [ 1. 1. 1.] [ 1. 1. 1.]] for u_row in u: u_row *= 2 << does work [[ 2. 2. 2.] [ 2. 2. 2.] [ 2. 2. 2.]] Naively, I'm thinking a *= b === a = a * b. Is this behavior expected? I'm asking because I really want: u_row = my_func (u_row)

3 2

Backwards slicing including the first element
by Jerome Esteve March 10, 2010

March 10, 2010

Dear all, Is there a way to give an integer value to j when using a[i:j:-1] so that the first element of the array can be included in the slice ? I would like to use some code like a[i:i-k:-1] to get a slice of length k. The numpy documentation seems to suggest that j=-1 should work: "Assume n is the number of elements in the dimension being sliced. Then, if i is not given it defaults to 0 for k > 0 and n for k < 0 . If j is not given it defaults to n for k > 0 and -1 for k < 0 . If k is not given it defaults to 1." But a[i:i-k:-1] is empty if i-k is -1. The workaround is a[i::-1][:k], is there something simpler ? Many thanks in advance, Jerome.

4 4

printing structured arrays
by Bruce Schultz March 10, 2010

March 10, 2010

Hi, I've just started playing with numpy and have noticed that when printing a structured array that the output is not nicely formatted. Is there a way to make the formatting look the same as it does for an unstructured array? Here an example of what I mean: data = [ (1, 2), (3, 4.1) ] dtype = [('x', float), ('y', float)] print '### ndarray' a = numpy.array(data) print a print '### structured array' a = numpy.array(data, dtype=dtype) print a Output is: ### ndarray [[ 1. 2. ] [ 3. 4.1]] ### structured array [(1.0, 2.0) (3.0, 4.0999999999999996)] Thanks Bruce

7 16

How to apply a function to an ndarray over a given dimension
by Jorge Scandaliaris March 10, 2010

March 10, 2010

Hi, First, excuse me if I am over-optimizing, but I am curious if there exist a way to apply a function to an ndarray over a given dimension. In case I don't make myself clear, I have an array of shape( n,2,2) where each row represents a 2 by 2 covariance matrix, and I want to perform the eigenvalue decomposition of each row. Right now I do it with list comprehensions: import numpy as np import scipy as sp C = np.arange(10*2*2).reshape(10,2,2) ed = [sp.linalg.eig(r) for r in C[:]] Is there a better way, along the lines of vectorize, of doing this? Cheers, Jorge

1 0

Memory leak in signal.convolve2d? Alternative?
by David Reichert March 10, 2010

March 10, 2010

Hi, I just reported a memory leak with matrices, and I might have found another (unrelated) one in the convolve2d function: import scipy.signal from numpy import ones while True: scipy.signal.convolve2d(ones((1,1)), ones((1,1))) Is there an alternative implementation of a 2d convolution? On the long run I'd be interested in using GPU acceleration, but for now I'd just like to get my code to run without running out of memory... Cheers! David The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

3 5

PSF GSoC 2010 (Py3K focus)
by Jarrod Millman March 10, 2010

March 10, 2010

Hello, Given the interest in participating in the GSoC this summer, I am forwarding a very interesting email from Titus Brown. If you are interested in doing a GSoC or mentoring, please read his email carefully. Basically, the PSF will be focuing on Py3K-related projects. Given Pauli's work on Py3K support for NumPy, I think we might be in a good position to move forward on porting the rest of our stack to Py3K. So we should focus on projects to: 1. finish porting and testing NumPy with Py3K 2. port and test SciPy with Py3K 3. port and test matplotlib with Py3K 4. port and test ipython with Py3K 5. etc. Given the PSF's stated emphasis this year, it probably doesn't make sense to pursue any non-Py3K projects. Jarrod ---------- Forwarded message ---------- From: C. Titus Brown <ctb(a)msu.edu> Date: Tue, Mar 2, 2010 at 6:12 AM Subject: [SoC2009-mentors] [ctb(a)msu.edu: GSoC 2010 - it's on!] To: soc2009-mentors(a)python.org ----- Forwarded message from "C. Titus Brown" <ctb(a)msu.edu> ----- Date: Wed, 24 Feb 2010 12:54:52 -0800 From: "C. Titus Brown" <ctb(a)msu.edu> To: psf-members(a)python.org Cc: gsoc2010-mentors(a)python.org Subject: GSoC 2010 - it's on! Hi all, it's that time of year again, and Google has decided to run the Google Summer of Code again! http://groups.google.com/group/google-summer-of-code-discuss/browse_thread/… http://socghop.appspot.com/ Arc Riley has stepped up to run it for the PSF again this year, and I'm backstopping him. If you are interested in mentoring or kibbitzing on those who are, please sign up for the soc2010-mentors mailing list here, http://mail.python.org/mailman/listinfo/soc2010-mentors This year we're proposing to solicit and prioritize applications for Python 3.x -- 3K tools, porting old projects, etc. Python 2.x projects will be a distinct second. There will be no "core" category this year, although obviously if someone on one of the core teams wants to push a project it'll help! If you have an idea for a project, please send it to the -mentors list and add it to the wiki at http://wiki.python.org/moin/SummerOfCode/2010 We're also going to change a few things up to make it more useful to the PSF. Specifically, - the foundation is going to *require* 1 blog post/wk from each student. - we're going to hire an administrative assistant to monitor the students. - the student application process will be a bit more rigorous and job-app like; the Django SF has been doing this for at least one round and they claim that it results in much better and more serious students. - we'll be focusing on student quality more than on project egalitarianism. If project X can recruit three fantastic students to one fantastic and one mediocre student for project Y, then project X gets three and project Y gets one. The hope is that this will make the GSoC much more useful for Python than it has been in the past. Arc will be posting something to the www.python.org site and python-announce soon, too. Followups to soc2010-mentors. cheers, --titus -- C. Titus Brown, ctb(a)msu.edu ----- End forwarded message -----

7 8

Memory leak with matrices?
by David Paul Reichert March 9, 2010

March 9, 2010

Hi, I've got two issues: First, the following seems to cause a memory leak, using numpy 1.3.0: a = matrix(ones(1)) while True: a += 0 This only seems to happen when a is a matrix rather than an array, and when the short hand '+=' is used. Second, I'm not sure whether that's a bug or whether I just don't understand what's going on, but when a is a column array, (e.g. a = ones((10, 1))), then a -= a[0,:] only subtracts from a[0, 0], whereas not using the short hand or using something else than a on the righthand side seems to subtract from all rows as expected. Thanks a lot, David -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

4 4