Mailman 3 August 2008 - SciPy-User

[ Python(x,y) ] New release : 2.0.1
by Pierre Raybaut 14 Aug '08

14 Aug '08

Hi all, As you may already know, Python(x,y) is a free scientific-oriented Python Distribution based on Qt and Eclipse providing a self-consistent scientific development environment. Release 2.0.1 is now available on http://www.pythonxy.com. (Full release, and update patch) Note: ITK will be updated in the next release (3.6 -> 3.8) and GDCM will be included. Changes history 08-14-2008 - Version 2.0.1 : ** Added ** o Python(x,y) is now available in two versions (due to disk quota limits - and to the unusual size of the 2.0.1 update, the Basic Edition will not be available online, at least in the next few weeks): Full Edition (all Python packages are installed, some are optional like ETS or ITK) and Basic Edition (with essential Python libraries only: PyQt4, NumPy, SciPy, IPython and matplotlib) o SWIG 1.3.36 - SWIG is a compiler that integrates C and C++ with several languages including Python o Pyrex 0.9.8.4 - Pyrex is a language for writing Python extension modules (Note: Cython - which is based on Pyrex - is already included in the distribution) o xy 1.0.2 - xy is a module that gathers all Python(x,y) tools ** Updated ** o Enthought Tool Suite 2.8.0 o NumPy 1.1.1 o matplotlib 0.98.3 o Pywin32 2.12 o pp (Parallel Python) 1.5.5 o SymPy 0.6.1 o PyTables 2.0.4 o Eclipse 3.4.0 (CDT 5.0) o PyDev 1.3.19 (Eclipse plugin) o Qt Eclipse Integration 1.4.1 o Photran 4.0b4 (Eclipse plugin) o Wicked Shell 2.0.4 (Eclipse plugin) o StartExplorer 0.0.4 (Eclipse plugin) o Notepad++ 5.0.3 (and added Python script execution shortcut: Shift+F1) o Console 2 installer: checking if an old configuration file exists (and deleting it) before installation o Console 2 configuration: window transparency has been disabled because of display bugs with TVTK o IPython(x,y) profile: added customizable startup script ** Corrected ** o IPython : IPython(x,y) profile startup script can now be customized o PyQt4: installation folder is added to PATH, allowing to use directly pyrcc4.exe, pylupdate.exe, ... o Missing documentation in the following packages: Cython, GDAL, DAP, MDP, PyXML, MinGW o Minor bug in package uninstallers: dialog box with an error message but without any consequence o ITK module installer: Visual C++ 2008 libraries installer "vcredist.exe" has a known bug which will be fixed in release 2008 SP1 - some temporary files are erroneously copied to the system root - meanwhile, these files are now deleted at the end of the installation process Regards, Pierre Raybaut

2 1

need lapack/atlas/fftw
by Rohit Garg 14 Aug '08

14 Aug '08

Hi all, I have installed scipy,numpy,lapack,atlas and fftw from the standard fedora repositories. Does anybody here know with options were they configured? I mean, when I am going to use the linalg module, am I using the lapack which uses ATLAS beneath it? I read in the list archives that scipy depends upon lapack and blas. Can I just set an environment variable somewhere to make it use ATLAS and FFTW? (probably not, but gotta ask). In case the answer is no, I guess the only way to do it is compile every thing by hand :( I am using fedora 9, 64 bit on dual core AMD machine. My machine supports sse2 Off topic - I understand that numpy does not make use of SSE and multithreading. If that is the case, I propose using the framewave library for implementing both of them. Their goal is not to bias it towards any particular company's chips and license is apache, so that should be no problem. I ask it here because numpy page has no links for dev lists and because I am eager to write code to achieve this. Cheers, -- Rohit Garg Junior Undergraduate Department of Physics Indian Institute of Technology Bombay

2 4

fmin_bfgs - Loglik Estimation leads to strange error
by Benjamin Eberlei 13 Aug '08

13 Aug '08

Hello everybody, i am quite new to numpy/scipy and currently porting some algorithms i wrote for Ox to python and run into a very strange error inside the bfgs optimization algorithm I cannot track down. I want to estimate Log Liklihood functions from different distributions. I prepared a small snippet: class fLoglik(object): def __init__(self, y): self._y = asmatrix(y).T pass def __call__(self, params): print params y = self._y a = _exp(-(params[0] * y)) return asmatrix(a) f = fLoglik( _log(data[:,16]) ) params = array(ones((1,1))) max = fmin_bfgs(f, params) data is a matrix so the input comes in correct. The fLoglik.__call__ is called twice before a ValueError occurs exactly as in the following backtrace: [ 1.] [ 1.00000001] Traceback (most recent call last): File "/var/www/workspace/pydiplom/src/diploma/weibull.py", line 42, in <module> max = fmin_bfgs(f, params) File "/usr/lib/python2.5/site-packages/scipy/optimize/optimize.py", line 723, in fmin_bfgs gfk = myfprime(x0) File "/usr/lib/python2.5/site-packages/scipy/optimize/optimize.py", line 95, in function_wrapper return function(x, *args) File "/usr/lib/python2.5/site-packages/scipy/optimize/optimize.py", line 617, in approx_fprime grad[k] = (f(*((xk+ei,)+args)) - f0)/epsilon ValueError: setting an array element with a sequence. What is weird to me, the first two function calls to fLoglik are correct, that is the output [1] and [ 1.00000001] you see comes from the call func. Does somebody maybe know thats going on?

1 0

SciPy 2008 hotel recommendation
by Ryan Krauss 13 Aug '08

13 Aug '08

I saw the link on nearby accommodations for the SciPy conference, but I don't know anything about the area. I will not have a car. Can anyone recommend a nearby (probably walking distance) hotel that is decent, reasonably priced, and not in a scary part of town (maybe that is a silly question, but I have never been to Pasadena)? Thanks, Ryan

4 4

Usage of PyDSTool
by Fabrice Silva 13 Aug '08

13 Aug '08

Hi, I used to deal with numerical solving of dynamical systems using ODE integrators provided by SciPy as VODE for example. I recently found PyDSTool which seems to be an excellent and almost complete tool for investigating bifurcation and other stuff... The point is that I do not even now how to begin! I do have a description of the dynamical system as y'=f(y,t) (N-dimensional vector y), this function being written in fortran for speed purpose and wrapped to python using f2py. How can I define a Model with such thing ? Is there any tutorial that could help ? I only found the example files which are not very explicit... Thanks -- Fabrice Silva LMA CNRS

4 10

Re: [SciPy-user] sparse csr_matrix memory error
by Dinesh B Vadhia 12 Aug '08

12 Aug '08

Hi Nathan For future reference, how did you arrive at the 864MB? Dinesh -------------------------------------------------------------------------------- Message: 2 Date: Sun, 10 Aug 2008 18:39:13 -0400 From: "Nathan Bell" <wnbell(a)gmail.com> Subject: Re: [SciPy-user] sparse csr_matrix memeory error To: "SciPy Users List" <scipy-user(a)scipy.org> Message-ID: <d05265cb0808101539l862ece2i8e97894afc522f37(a)mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1 On Sun, Aug 10, 2008 at 5:29 PM, Dinesh B Vadhia <dineshbvadhia(a)hotmail.com> wrote: > I'm obtaining a memory error when creating a large sparse csr matrix as > follows: > > I = 680000 > J = 900000 > nnz = 72000000 > row = numpy.empty(nnz, dtype='intc') > column = numpy.empty(nnz, dtype='intc') > # read (i,j) data into row and column > data = scipy.ones(nnz, dtype='intc') Together, the arrays for row,column, and data take 864MB of memory. You need approximately 2x that to do the conversion to CSR. > I'm running the program under Windows XP with over 2gb memory. Any thoughts > on what the problem is? Yes, your matrix is simply too large. -- Nathan Bell wnbell(a)gmail.com http://graphics.cs.uiuc.edu/~wnbell/

2 1

fit confidence intervals from minpack.leastsq, or odrpack.ODR?
by Zane Selvans 11 Aug '08

11 Aug '08

I have an observation, L, which consists of the shape of a particular one dimensional feature (a line on the surface of a sphere). I have a model of the process that I think may have generated the feature. Aside from the feature, L, the model has one parameter, B (an angle: 0 < B < pi). For a given feature and parameter value, I can calculate a metric f(B,L) describing how well my observation matches the model. f(B,L) has the following properties: - Small values of f(B,L) indicate agreement with my model, large f(B,L) indicate disagreement. - f(B,L) is periodic (with a wavelength of pi radians) - f(B,L) may have several local minima - 0 < f(B,L) < pi/2 I need to somehow quantify three things: i) At its best, how good is my model at explaining the observation (i.e. is it good enough to be significant?) ii) For what value of the input parameter does my model do the best job of explaining the observation? iii) How unique is that best value (i.e. are there many other values that do almost as well?) Currently, I'm using the global minimum of f(B,L) for i, and the value of B which results in the global minimum for ii. I'm kind of stuck on iii though. My current idea is to fit f(B,L) to some periodic function (e.g. cosine), and use the width of the 95% confidence interval of that fit as an indication of its uniqueness. If I use a function with the same wavelength (pi) as f(B,L), set its amplitude to be whatever the observed amplitude of f(B,L), and its phase such that the minimum of both f(B,L) and the cosine... I'll get a fit of some confidence. Or alternatively, I could allow the fitting function to determine the phase, and instead of using the global minimum of f(B,L) as the value which determines the best value of B, I could use the minimum of the fit cosine. Or I could just not worry about whether or not they're the same, and use the best-fit confidence interval as the measure of uniqueness. Does that sound at all like the right way to go about this? If so, which fitting/minimization module is more appropriate/easier to use (if what I want is the confidence interval, ultimately) minpack.leastsq or odrpack.ODR. I see that leastsq returns a covariance matrix, and that it's possible somehow to turn that into a confidence interval... and it looks like you can get a confidence interval (sd_beta) directly from ODR. Thanks for any insight... -- Zane Selvans Amateur Earthling http://zaneselvans.org zane(a)ideotrope.org 303/815-6866 PGP Key: 55E0815F

1 0

installing numpy on ubuntu
by Michael Hearne 11 Aug '08

11 Aug '08

I'm trying to install numpy/scipy/matplotlib/basemap on a fresh installation of Ubuntu. The available versions of each I've found on Ubuntu: numpy: Version: 1:1.0.4-6ubuntu3 scipy: Version: 0.6.0-8ubuntu1 matplotlib: Version: 0.91.2-0ubuntu1 Basemap (a matplotlib toolkit) does not appear in Ubuntu, it seems. The problem is that the supported version of Basemap depends on matplotlib versions 0.98 and greater, and for that I need numpy 1.1. I have two questions: 1) Is there an apt-get repository for the more recent versions of numpy (or any of these other packages) 2) Failing 1, should the building of numpy on an Ubuntu system be a straightforward "configure;make;make install"? God, I hope so. Thanks, Mike -- ------------------------------------------------------ Michael Hearne mhearne(a)usgs.gov (303) 273-8620 USGS National Earthquake Information Center 1711 Illinois St. Golden CO 80401 Senior Software Engineer Synergetics, Inc. ------------------------------------------------------

3 2

sparse csr_matrix memeory error
by Dinesh B Vadhia 10 Aug '08

10 Aug '08

I'm obtaining a memory error when creating a large sparse csr matrix as follows: I = 680000 J = 900000 nnz = 72000000 row = numpy.empty(nnz, dtype='intc') column = numpy.empty(nnz, dtype='intc') # read (i,j) data into row and column data = scipy.ones(nnz, dtype='intc') A = sparse.csr_matrix((data, (row, column)), shape=(I,J)) The traceback is: Traceback (most recent call last): File "C:\... \ijdata.py", line 72, in <module> A = sparse.csr_matrix((data, (row, column)), shape=(I,J)) File "C:\Python25\Lib\site-packages\scipy\sparse\compressed.py", line 55, in __init__ other = self.__class__( coo_matrix(arg1, shape=shape) ) File "C:\Python25\Lib\site-packages\scipy\sparse\compressed.py", line 39, in __init__ arg1 = arg1.asformat(self.format) File "C:\Python25\Lib\site-packages\scipy\sparse\base.py", line 211, in asformat return getattr(self,'to' + format)() File "C:\Python25\Lib\site-packages\scipy\sparse\coo.py", line 278, in tocsr indices = empty(self.nnz, dtype=intc) MemoryError I'm running the program under Windows XP with over 2gb memory. Any thoughts on what the problem is? Dinesh

2 1

concatenate array with number
by Nicolas Chopin 09 Aug '08

09 Aug '08

Hi list, I want to do this: x = concatenate( (x,x[-1]) ) i.e. append to 1d array x its last element. However, the only way I managed to do this is: x = concatenate( (x,array(x[-1],ndmin=1)) ) which is a bit cryptic. (if you remove ndmin, it does not work.) 1. Is there a better way? 2. Could concatenate accept floating point numbers as arguments for convenience? Thanks in advance, Nicolas <http://www.stats.bris.ac.uk/%7Emanxac/> <http://www.stats.bris.ac.uk/%7Emanxac/>

3 3