Mailman 3 October 2007 - NumPy-Discussion

strange problem compiling against ndarrayobject.h
by Vincent Schut Oct. 31, 2007

Oct. 31, 2007

Btw forgot some info: gentoo linux, amd64 numpy 1.0.3.1 gdal from svn, updated today python 2.5.1 Cheers, Vincent.

1 0

MTL 4
by Charles R Harris Oct. 31, 2007

Oct. 31, 2007

Just a note that an alpha version of the hitherto mostly invisible MTL4 library has been released. It is template based and has automatic bindings to blas. You can download it from http://www.osl.iu.edu/research/mtl/mtl4/. Chuck

2 1

trailing max
by John Hunter Oct. 30, 2007

Oct. 30, 2007

In financial time series, it is very common to keep track of things like a trailing N day max, trailing N day average, etc. Generally, for a 1D array x, I'd like to be able to efficiently compute a new len(x) vector where y[i] = func(x[i-N:]) and I need to be able to handle edge effects (eg where i<N). I'm thinking of writing some extension code to handle this, but since this is a common task, I thought I'd elicit some suggestions about how to best handle this. One approach might to create a very large 2D array with each row the appropriate masked (to handle edge effects) shifted array, and then one could do sums and maxes over the columns. For a length 1000 vector with a max lag of 250, you'd have 250K elements, which is probably manageable, but I'm still not sure how to best create this shifted array and if there are better approaches. JDH

5 4

More intel about the bug "Reference count error detected"
by Matthieu Brucher Oct. 30, 2007

Oct. 30, 2007

Hi, Instead of using : xp = (x-self.loc) * (1/self.scale) I've decided to use : xp = (x-self.loc) / self.scale and by some miracle, the bug does not show up in almost all tests. Before this bug, the program seems to be hung in a DOUBLE_substract function() for quite some time (umathmodule.c file). Perhaps it is this function that triggers to many references or dereferencing them too much ? If someone wants to take a look, I may be able to wrap what is needed in a single (but big) file. Matthieu -- French PhD student Website : http://miles.developpez.com/ Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92

1 1

numpy version and numpy.asarray behavior issue
by Ray S Oct. 29, 2007

Oct. 29, 2007

I just installed 1.0.3.1 on top of Enthought's and asarray() works. But... Although the creation of an array from an address via a Dummy class is kosher in one process (as in the previous attachment), it fails across processes - the array is created, but gives a "Python has generated errors" window if the second process even attempts to read. It can seem to work across processes with mmap.mmap() and tags (used in Windows) def arrSharedMemory(shape, dtype, tag="PySharedMemory", access=None): ## Windows only ! share memory between different ## processes if same `tag` is used. itemsize = N.dtype(dtype).itemsize count = N.product(shape) size = count * itemsize import mmap sharedmem = mmap.mmap(0, size, tag, access) a=N.frombuffer(sharedmem, dtype, count) a.shape = shape return a I guess I'll use mmap unless someone can point out otherwise.... Thanks, Ray

1 0

numpy version and numpy.asarray behavior issue
by Ray S Oct. 29, 2007

Oct. 29, 2007

I have 2 PCs with 2 different installs: ActivePython 2.4.3 Build 12 with numpy version '1.0b1' and Enthought 2.4.3 (1.0.0 #69) with numpy version '0.9.7.2476' The attached runs Ok on numpy v1.0, but on Enthought's, print a1[0] gives: IndexError: 0-d arrays can't be indexed. It seems that the 0.9.7 numpy.asarray is not creating a true array from Dummy class in the code below. Enthought only comes with 0.9.9.2706 (now). When was the asarray behavior supported, or, is there some other issue I'm missing? I'll use Activestate's distro if needed, but I'd like to keep Enthought for that one... def fromaddress(address, dtype, shape, strides=None): """ Create a numpy array from an integer address, a dtype or dtype string, a shape tuple, and possibly strides. """ import numpy # Make sure our dtype is a dtype, not just "f" or whatever. dtype = numpy.dtype(dtype) class Dummy(object): pass d = Dummy() d.__array_interface__ = dict( data = (address, False), typestr = dtype.str, descr = dtype.descr, shape = shape, strides = strides, version = 3, ) return numpy.asarray(d) Thanks, Ray

1 0

Getting an item in an array with its coordinates given by another array
by Matthieu Brucher Oct. 29, 2007

Oct. 29, 2007

Hi, I'm trying to get an item in an array when I only have another array giving the position. For instance: >>> c = numpy.arange(0., 3*4*5).reshape((3,4,5)) >>> c***(numpy.array((2,3), dtype=int)) [55, 56, 57, 58, 59] I'm trying to figure what to put between c and ((2,3)). I supposed that the take method would be what I wanted but it is not... The [] operator does not work on array(but it seems to give the expected result with a tuple or a list). Is there a method for this ? Matthieu -- French PhD student Website : http://miles.developpez.com/ Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92

5 11

Error in deallocation ?
by Matthieu Brucher Oct. 29, 2007

Oct. 29, 2007

Hi I keep on getting this error : *** Reference count error detected: an attempt was made to deallocate 7 (l) *** It happens on numpy calls (multiplications, call to inner(), ...), but I didn't find the real reason. I'm using Numpy '1.0.4.dev3875' with Python 2.5.1. Someone has a hint to solve this ? Matthieu

4 23

[numpy.scons] Some really basic starting info
by David Cournapeau Oct. 29, 2007

Oct. 29, 2007

Hi, Since several people have been asking me how to use scons to build python modules, I have started a wiki page on numpy wiki: http://scipy.org/scipy/numpy/wiki/NumpyScons This is really preliminary, but it should be enough to get you started if you have simple needs (building a Ctypes extension, checking for the presence of some libraries, etc...). cheers, David

1 0

NumPy 1.0.4 release
by Jarrod Millman Oct. 29, 2007

Oct. 29, 2007

Hi all, After speaking with Travis, I think that we can release NumPy 1.0.4 by the end of the month. 1.0.3 came out almost 5 months ago and there have been a number of bug-fixes and other improvements since then. Please take a look at the 1.0.4 roadmap: http://scipy.org/scipy/numpy/milestone/1.0.4 There are currently 23 active tickets that could use some attention. Hopefully, we can close as many of these as possible before the release (some need fixes, some can just be closed if appropriate). If you see something that must be addressed before the release, mark it with a severity level of 'blocker'. Here is the proposed release schedule: Monday (10/22) * Development Freeze. After this point, only bug fixes should be checked in. Sunday (10/28) * Release Tagged. At this point, I will be asking for help from packagers and testers. Thursday (11/1) * Release Announced. Thanks, -- Jarrod Millman Computational Infrastructure for Research Labs 10 Giannini Hall, UC Berkeley phone: 510.643.4014 http://cirl.berkeley.edu/

4 11