Mailman 3 March 2010 - NumPy-Discussion

Bug in logaddexp2.reduce
by T J 04 Apr '10

04 Apr '10

Hi, I'm getting some strange behavior with logaddexp2.reduce: from itertools import permutations import numpy as np x = np.array([-53.584962500721154, -1.5849625007211563, -0.5849625007211563]) for p in permutations([0,1,2]): print p, np.logaddexp2.reduce(x[list(p)]) Essentially, the result depends on the order of the array...and we get nans in the "bad" orders. Likely, this also affects logaddexp.

8 41

Applying formula to all in an array which has value from previous
by Vishal Rana 03 Apr '10

03 Apr '10

Hi, For a numpy array: array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) I do some calculation with 0, 1... and get a value = 2.5, now use this value to do the repeat the same calculation with next element for example... 2.5, 2 and get a value = 3.1 3.1, 3 and get a value = 4.2 4.2, 4 and get a value = 5.1 .... .... and get a value = 8.5 8.5, 9 and get a value = 9.8 So I should be getting a new array like array([0, 2.5, 3.1, 4.2, 5.1, ..... 8.5,9.8]) Is it where numpy or scipy can help? Thanks Vishal

2 3

Dealloat Numy arrays in C
by Jeremy Lewi 01 Apr '10

01 Apr '10

Hello, I'm passing a numpy array into a C-extension. I would like my C-extension to take ownership of the data and handle deallocating the memory when it is no longer needed. (The data is large so I want to avoid unnecessarily copying the data). So my question is what is the best way to ensure I'm using the correct memory deallocator for the buffer? i.e the deallocator for what ever allocator numpy used to allocate the array? Thanks Jeremy Jeremy Lewi Engineering Scientist The Intellisis Corporation jlewi(a)intellisis.com

2 1

Set values of a matrix within a specified range to zero
by Sean Mulcahy 31 Mar '10

31 Mar '10

Hello all, I'm relatively new to numpy. I'm working with text images as 512x512 arrays. I would like to set elements of the array whose value fall within a specified range to zero (eg 23 < x < 45). Any advice is much appreciated. Sean

5 7

Fourier transform
by Pascal 31 Mar '10

31 Mar '10

Hi, Does anyone have an idea how fft functions are implemented? Is it pure python? based on BLAS/LAPACK? or is it using fftw? I successfully used numpy.fft in 3D. I would like to know if I can calculate a specific a plane using the numpy.fft. I have in 3D: r(x, y, z)=\sum_h^N-1 \sum_k^M-1 \sum_l^O-1 f_{hkl} \exp(-2\pi \i (hx/N+ky/M+lz/O)) So for the plane, z is no longer independant. I need to solve the system: ax+by+cz+d=0 r(x, y, z)=\sum_h^N-1 \sum_k^M-1 \sum_l^O-1 f_{hkl} \exp(-2\pi \i (hx/N+ky/M+lz/O)) Do you think it's possible to use numpy.fft for this? Regards, Pascal

3 3

numpy.distutils/f2py: forcing 8-bit reals
by David Warde-Farley 31 Mar '10

31 Mar '10

Hi, In my setup.py, I have from numpy.distutils.misc_util import Configuration fflags= '-fdefault-real-8 -ffixed-form' config = Configuration( 'foo', parent_package=None, top_path=None, f2py_options='--f77flags=\'%s\' --f90flags=\'%s\'' % (fflags, fflags) ) However I am still getting stuff returned in 'real' variables as dtype=float32. Am I doing something wrong? Thanks, David

2 5

Re: [Numpy-discussion] Interpolation question
by Kevin Dunn 31 Mar '10

31 Mar '10

> Message: 5 > Date: Sun, 28 Mar 2010 00:24:01 +0000 > From: Andrea Gavana <andrea.gavana(a)gmail.com> > Subject: [Numpy-discussion] Interpolation question > To: Discussion of Numerical Python <numpy-discussion(a)scipy.org> > Message-ID: > <d5ff27201003271724o6c82ec75v225d819c84140b46(a)mail.gmail.com> > Content-Type: text/plain; charset=ISO-8859-1 > > Hi All, > > I have an interpolation problem and I am having some difficulties > in tackling it. I hope I can explain myself clearly enough. > > Basically, I have a whole bunch of 3D fluid flow simulations (close to > 1000), and they are a result of different combinations of parameters. > I was planning to use the Radial Basis Functions in scipy, but for the > moment let's assume, to simplify things, that I am dealing only with > one parameter (x). In 1000 simulations, this parameter x has 1000 > values, obviously. The problem is, the outcome of every single > simulation is a vector of oil production over time (let's say 40 > values per simulation, one per year), and I would like to be able to > interpolate my x parameter (1000 values) against all the simulations > (1000x40) and get an approximating function that, given another x > parameter (of size 1x1) will give me back an interpolated production > profile (of size 1x40). [I posted the following earlier but forgot to change the subject - it appears as a new thread called "NumPy-Discussion Digest, Vol 42, Issue 85" - please ignore that thread] Andrea, may I suggest a different approach to RBF's. Realize that your vector of 40 values for each row in y are not independent of each other (they will be correlated). First build a principal component analysis (PCA) model on this 1000 x 40 matrix and reduce it down to a 1000 x A matrix, called your scores matrix, where A is the number of independent components. A is selected so that it adequately summarizes Y without over-fitting and you will find A << 40, maybe A = 2 or 3. There are tools, such as cross-validation, that will help select a reasonable value of A. Then you can relate your single column of X to these independent columns in A using a tool such as least squares: one least squares model per column in the scores matrix. This works because each column in the score vector is independent (contains totally orthogonal information) to the others. But I would be surprised if this works well enough, unless A = 1. But it sounds like your don't just have a single column in your X-variables (you hinted that the single column was just for simplification). In that case, I would build a projection to latent structures model (PLS) model that builds a single latent-variable model that simultaneously models the X-matrix, the Y-matrix as well as providing the maximal covariance between these two matrices. If you need some references and an outline of code, then I can readily provide these. This is a standard problem with data from spectroscopic instruments and with batch processes. They produce hundreds, sometimes 1000's of samples per row. PCA and PLS are very effective at summarizing these down to a much smaller number of independent columns, very often just a handful, and relating them (i.e. building a predictive model) to other data matrices. I also just saw the suggestions of others to center the data by subtracting the mean from each column in Y and scaling (by dividing through by the standard deviation). This is a standard data preprocessing step, called autoscaling and makes sense for any data analysis, as you already discovered. Hope that helps, Kevin > Something along these lines: > > import numpy as np > from scipy.interpolate import Rbf > > # x.shape = (1000, 1) > # y.shape = (1000, 40) > > rbf = Rbf(x, y) > > # New result with xi.shape = (1, 1) --> fi.shape = (1, 40) > fi = rbf(xi) > > > Does anyone have a suggestion on how I could implement this? Sorry if > it sounds confused... Please feel free to correct any wrong > assumptions I have made, or to propose other approaches if you think > RBFs are not suitable for this kind of problems. > > Thank you in advance for your suggestions. > > Andrea. > > "Imagination Is The Only Weapon In The War Against Reality." > http://xoomer.alice.it/infinity77/ > > ==> Never *EVER* use RemovalGroup for your house removal. You'll > regret it forever. > http://thedoomedcity.blogspot.com/2010/03/removal-group-nightmare.html <==

2 2

Replacing NANs
by Vishal Rana 31 Mar '10

31 Mar '10

Hi, In an array I want to replace all NANs with some number say 100, I found a method* **nan_to_num *but it only replaces with zero. Any solution for this? * *Thanks Vishal

3 3

indexing question
by Tom K. 30 Mar '10

30 Mar '10

This one bit me again, and I am trying to understand it better so I can anticipate when it will happen. What I want to do is get rid of singleton dimensions, and index into the last dimension with an array. In [1]: import numpy as np In [2]: x=np.zeros((10,1,1,1,14,1024)) In [3]: x[:,0,0,0,:,[1,2,3]].shape Out[3]: (3, 10, 14) Whoa! Trimming my array to a desired number ends up moving the last dimension to the first! In [4]: np.__version__ Out[4]: '1.3.0' ... In [7]: x[:,:,:,:,:,[1,2,3]].shape Out[7]: (10, 1, 1, 1, 14, 3) This looks right... In [8]: x[...,[1,2,3]].shape Out[8]: (10, 1, 1, 1, 14, 3) and this... In [9]: x[...,[1,2,3]][:,0,0,0].shape Out[9]: (10, 14, 3) ... In [11]: x[:,0,0,0][...,[1,2,3]].shape Out[11]: (10, 14, 3) Either of the last 2 attempts above results in what I want, so I can do that... I just need some help deciphering when and why the first thing happens. -- View this message in context: http://old.nabble.com/indexing-question-tp28083162p28083162.html Sent from the Numpy-discussion mailing list archive at Nabble.com.

5 4

SciPy 2010: Vote on Tutorials & Sprints
by Amenity Applewhite 30 Mar '10

30 Mar '10

Email not displaying correctly? View it in your browser. Hello Amenity, Spring is upon us and arrangements for SciPy 2010 are in full swing. We're already nearing on some important deadlines for conference participants: April 11th is the deadline for submitting an abstract for a paper, and April 15th is the deadline for submitting a tutorial proposal. Help choose tutorials for SciPy 2010... We set up a UserVoice page to brainstorm tutorial topics last week and we already have some great ideas. The top ones at the moment are: Effective multi-core programming with Cython and Python Building your own app with Mayavi High performance computing with Python Propose your own or vote on the existing suggestions here. ...Or instruct a tutorial and cover your conference costs. Did you know that we're awarding generous stipends to tutorial instructors this year? So if you believe you could lead a tutorial, by all means submit your proposal — soon! They're due April 15th. Call for Papers Continues Submitting a paper for to present at SciPy 2010 is easy, so remember to prepare one and have your friends and colleagues follow suit. Send us your abstract before April 11th and let us know whether you'd like to speak at the main conference or one of the specialized tracks. Details here. Have you registered? Booking your tickets early should save you money — not to mention the early registration prices you will qualify for if you register before May 10th. Best, The SciPy 2010 Team @SciPy2010 on Twitter You are receiving this email because you have registered for the SciPy 2010 conference in Austin, TX. Unsubscribe amenity(a)enthought.com from this list | Forward to a friend | Update your profile Our mailing address is: Enthought, Inc. 515 Congress Ave. Austin, TX 78701 Add us to your address book Copyright (C) 2010 Enthought, Inc. All rights reserved.

1 0