Mailman 3 March 2008 - SciPy-User

minpack.error / fsolve problem
by Webb Sprague 04 Mar '08

04 Mar '08

I am having a problem with convergence (I think) for an optimization. Every so often I do a non-linear fit of a parameter using fsolve (as input a constant vector of base death rates, a constant vector multiplier of those, and a variable scalar multiplier -- the last is what I am trying to fit) and I get a "minpack.error" with the message that "Error occured while calling the Python function named f" with no other information. The information normally returned from fsolve (ier, message, infodict) are all set to none. I am kind of at a loss for how to proceed, at least without taking a class on optimization algorithms. How do I get more information from fsolve? Is there a better optimization function to use? optimize.golden()? Or ... ? See attached for how I set up the optimizer, and here is the back trace from the web application using this code: TRACEBACK: Traceback (most recent call last): File "/usr/lib64/python2.5/site-packages/mod_python/importer.py", line 1537, in HandlerDispatch default=default_handler, arg=req, silent=hlist.silent) File "/usr/lib64/python2.5/site-packages/mod_python/importer.py", line 1229, in _process_target result = _execute_target(config, req, object, arg) File "/usr/lib64/python2.5/site-packages/mod_python/importer.py", line 1128, in _execute_target result = object(arg) File "/usr/lib64/python2.5/site-packages/mod_python/publisher.py", line 213, in handler published = publish_object(req, object) File "/usr/lib64/python2.5/site-packages/mod_python/publisher.py", line 425, in publish_object return publish_object(req,util.apply_fs_data(object, req.form, req=req)) File "/usr/lib64/python2.5/site-packages/mod_python/util.py", line 554, in apply_fs_data return object(**args) File "/var/www/localhost/htdocs/lcfit/lc.py", line 221, in ProcessRatesCoherent return ProcessRatesCoherent_(req) File "/home/webbs/larry/INTERNET_APPLICATION/LcPageObjects.py", line 425, in __call__ obj = self.targetClass(**formData) File "/home/webbs/larry/INTERNET_APPLICATION/LcSinglePopObject.py", line 530, in __init__ self._do_lc() File "/home/webbs/larry/INTERNET_APPLICATION/LcCoherentPopObject.py", line 205, in _do_lc flattenBx=self.flattenBx, doTS=True)) File "/home/webbs/larry/INTERNET_APPLICATION/LcSinglePopObject.py", line 338, in lcInfer kt = fitMultiKt(ax, bx, copy.copy(kt_unfit), nmx[goodRowsNum,:], lifeTableParams) File "/home/webbs/larry/INTERNET_APPLICATION/LcSinglePopObject.py", line 103, in fitMultiKt fittedKt[i] = fitSingleKt(ax, bx, kt[i], nmx[i,:], lifeTableParams) # ages go across File "/home/webbs/larry/INTERNET_APPLICATION/LcSinglePopObject.py", line 88, in fitSingleKt fittedKt = LcUtil.fitX(func=LcUtil.kt2e0, target=target_e0, ax=ax, bx=bx, lifeTableParams=lifeTableParams) File "/home/webbs/larry/INTERNET_APPLICATION/LcUtil.py", line 460, in fitX "funcKwargs: %s\n" % pprint.pformat(funcKwargs)) Exception: Caught exception and weird failure in fitting kt to empirical e_0. e: "Error occured while calling the Python function named f". error __doc__: "None". type: "<class 'minpack.error'>". out: "None". infodict: "None". ier: "None". mesg: "None". funcArgs: () funcKwargs: {'ax': array([-4.231837, -6.92667 , -7.537601, -7.668495, -6.554903, -5.971107, -5.759917, -5.514402, -5.230182, -4.900118, -4.559831, -4.244538, -3.937256, -3.602081, -3.269744, -2.901267, -2.51867 , -2.118842, -1.746651, -1.402097, -1.085182, -0.795903, -0.534262, -0.300259]), 'bx': array([ 0.135533, 0.126908, 0.117685, 0.095685, 0.029012, 0.019818, 0.025011, 0.003613, -0.00145 , -0.03515 , -0.0542 , -0.062684, -0.051495, -0.039603, -0.029531, -0.011641, -0.002807, -0.018365, -0.028617, -0.033565, -0.033207, -0.027544, -0.016576, -0.000302]), 'lifeTableParams': {'ageCutoff': 80, 'beginFuncParam': 0.0, 'endFuncParam': 0.0, 'extensionMethod': 'mxExtend_Boe', 'gender': 'combined', 'ltFuncType': 'ex'}}

2 3

Newbie broadcasting question
by John Reid 04 Mar '08

04 Mar '08

This works fine: In [49]: numpy.zeros((3,2))+numpy.array([1,2]) Out[49]: array([[ 1., 2.], [ 1., 2.], [ 1., 2.]]) but this doesn't: In [50]: numpy.zeros((3,2))+numpy.array([1,2,3]) --------------------------------------------------------------------------- ValueError Traceback (most recent call last) C:\Dev\MyProjects\Bio\Python\site_dpm\<ipython console> in <module>() ValueError: shape mismatch: objects cannot be broadcast to a single shape > <ipython console>(1)<module>() what is the simplest way to get this addition to work? I would like the following array([[ 0., 0.], [ 0., 0.], [ 0., 0.]]) + array([1,2,3]) = array([[ 1., 1.], [ 2., 2.], [ 3., 3.]]) In general my arrays are held in variables and are not constructed on the fly so while this works numpy.zeros((3,2)).T+numpy.array([1,2,3]) which works fine with the '+' operator but not with '+='. This doesn't work: In [54]: a=zeros((3,2)) In [55]: a.T += numpy.array([1,2,3]) --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) C:\Dev\MyProjects\Bio\Python\site_dpm\<ipython console> in <module>() AttributeError: attribute 'T' of 'numpy.ndarray' objects is not writable > <ipython console>(1)<module>() My guess is that I should do a = (a.T + numpy.array([1,2,3])).T but I wonder if there is a more efficient way. Am I missing something? Thanks, John.

3 2

change array index order
by BK 03 Mar '08

03 Mar '08

Hi everybody, I am still kind of a newbie but I've been using scipy for some time now for post-processing of CFD surface ASCII data. Over the time I wrote myself a bunch of small tools doing this and that. Lately there was a version change in the CFD code I am using and the output format changed slightly. Old format was I,J-grid data with N variables in the order N,I,J (N being the fastet running index). The new data format uses I,J,N ordering (with I running fastest). So, is there a way to tell scipy to somehow 'map' one index direction to another one, like: data_new[n,j,i] -> data_old[j,i,n] without copying all the data? This way I could just change my IO routine and keep all the other tools unchanged. And I could easily handle new and old version of output files. I suspect this is a real newbie-question, but so far I haven't been able to figure it out myself. Thanks for your help, and thanks for that great toos scipy! Best regards Bene

3 2

Varias figuras en aplicacion en local
by Natali Melgarejo Diaz 03 Mar '08

03 Mar '08

Buenas a todos, He hecho una aplicacion grafica en pyQt4 q al hacer click en los botones tiene que hacer graficas de funciones predefinidas en otras clases, bueno, al hacer click en el 2do boton, el programa ya no responde. y yo quisiera q el usuario pueda ver ambas figuras o las que sean, al aprecer solo se puede hacer una ventana por una tal y como lo tengo implementado. Usé Threads pero el problema persiste, y las figuras las hago usando pylab con figure(1), figure(2) ..etc pero al hacer show() parece q se raya :S Alguien ha tenido q hacer algo parecido ? como lo solucionó? Gracias x adelantado! Natali ;)

1 0

Maximally distinguishable colour map
by David Warde-Farley 02 Mar '08

02 Mar '08

Hi folks, This isn't so much a SciPy question as it is a matplotlib question (or even a general "scientific computing question"), but I figure someone else has got to have run across the problem. Essentially I'm trying to choose a set of colours to display a graph plot that maximizes the visual differentiability of them. Basically, I'm going to be producing a graph with a lot of edges (in fact, a multigraph, i.e. more than one edge between a pair of nodes) and it gets pretty hard to distinguish colours that are very "close" together. Can anyone perhaps point me at a colour map that maximizes the visual difference between consecutive colours, or at least a way of constructing one? The method I have now works fairly well (every 43rd colour in the HSV map in matplotlib) but it unfortunately includes yellow early in the sequence (I should mention this is going to be on a white background). Thanks, David

2 3

Lawson-Hanson's Non-Negative Least Squares (NNLS) algorithm
by Bing 01 Mar '08

01 Mar '08

Hi Dmitrey, Thanks for your response to my previous email. I just handcoded a Python extension of Lawson-Hanson's Non-Negative Least Squares (NNLS) algorithm using C++ based on the following source code: http://www.cs.utexas.edu/~suvrit/work/progs/nnls.html Seems it is working fine for me. The extension is straightforward and please feel free to email me (bing.jian(a)gmail.com) if you are interested in this extension. Thanks! Bing

1 0

nonnegative linear squares (NNLS, lsqnonneg) in Python
by Jian, Bing (MED US) 01 Mar '08

01 Mar '08

Hi, I am wondering if there is a non-negative linear squares solver in scipy/numpy which is equivalent to the lsqnonneg() in MATLAB? If not, then probably I need to write my own extensions based on existing C code. Thanks! Bing ---------------------------------------------------------------------------- This message and any included attachments are from Siemens Medical Solutions and are intended only for the addressee(s). The information contained herein may include trade secrets or privileged or otherwise confidential information. Unauthorized review, forwarding, printing, copying, distributing, or using such information is strictly prohibited and may be unlawful. If you received this message in error, or have reason to believe you are not authorized to receive it, please promptly delete this message and notify the sender by e-mail with a copy to Central.SecurityOffice(a)siemens.com Thank you

2 1

How to tell scipy setup that I have a INTEL fortran ATLAS/BLAS/LAPACK instead of g77
by Berthold "Höllmann" 29 Feb '08

29 Feb '08

(Message resent, i was not subscribed, so yesterdays version reached the moderator only) No matter what I do, I can't tell scipy to use the INTEL fortran API conventions instead of the g77 conventions for fortran routine names containing underscores: hoel@pc047299:scipy-0.6.0 nm /usr/local/gltools/linux/lib/libf77blas_ifc91.so.3.8|grep atl_f77wrap_dtrsv 0000cfe0 T atl_f77wrap_dtrsv_ hoel@pc047299:scipy-0.6.0 nm build/lib.linux-i686-2.5/scipy/linsolve/_zsuperlu.so| grep atl_f77wrap_dtrsv U atl_f77wrap_dtrsv__ How can I set up scipy in a way that superlu tries to access atl_f77wrap_dtrsv_ instead of atl_f77wrap_dtrsv__? Kind regards Berthold Höllmann -- Germanischer Lloyd AG CAE Development Vorsetzen 35 20459 Hamburg Phone: +49(0)40 36149-7374 Fax: +49(0)40 36149-7320 e-mail: berthold.hoellmann(a)gl-group.com Internet: http://www.gl-group.com This e-mail and any attachment thereto may contain confidential information and/or information protected by intellectual property rights for the exclusive attention of the intended addressees named above. Any access of third parties to this e-mail is unauthorised. Any use of this e-mail by unintended recipients such as total or partial copying, distribution, disclosure etc. is prohibited and may be unlawful. When addressed to our clients the content of this e-mail is subject to the General Terms and Conditions of GL's Group of Companies applicable at the date of this e-mail. If you have received this e-mail in error, please notify the sender either by telephone or by e-mail and delete the material from any computer. GL's Group of Companies does not warrant and/or guarantee that this message at the moment of receipt is authentic, correct and its communication free of errors, interruption etc. Germanischer Lloyd AG, 31393 AG HH, Hamburg, Vorstand: Dr. Hermann J. Klein, Dr. Joachim Segatz, Vorsitzender des Aufsichtsrats: Dr. Wolfgang Peiner

2 3