Mailman 3 May 2003 - SciPy-User

xplt, dynamic plots question
by Arnd Baecker May 19, 2003

May 19, 2003

Hi, concerning my question below it seems that everything boils down to the observation, that scipy.xplt.pause(1) does not work under Windows. It either crashes pythonwin (the interpreter window dissapears) or the "normal" python hangs, I was told. Personally I am not using windows, but many of our students do and until now we have been able to work around this problem... Background/Details: When doing a computation for several parameters I would like to plot the results for each parameter once they have been computed. A stripped down set of examples (without real computation ;-) is at the end of the email - if you are interested in the one with the failure go straight to dynamik3.py. For linux one solution is to use scipy.xplt.pause(1), see dynamik1.py below. Somehow scipy.xplt.pause(1) seems to act like a flush command for the present plot. A solution which works both under linux and windows (however doing a full redraw) uses xplt.fma, see dynamik2.py However, when trying to do this with a xplt.pli command it seems we cannot avoid the pause(1), see dynamik3.py. Is there a way to get this working under windows as well? Any suggestions are very welcome! Arnd P.S.: sorry for causing so much traffic with these questions, but as we do weekly exercises in the course we are giving they just come up quite frequently now ;-) ########################################## # dynamik1.py # # This one works on my machine under linux (debian) # in a nice dynamic way !! # (i.e. one xplt.plg is done after the other) from Numeric import * import scipy import scipy.xplt as xplt xplt.winkill() xplt.window() t=arange(0.0,0.5,0.1) mat=reshape(arange(100),(10,10)) xplt.limits(0,2000,0,2000) for i in arange(0,2000,1): mat[0,0]=i ew,ef=scipy.linalg.eig(mat) xplt.plg(ew.real,i*ones(len(ew)),type=0,marker=1,msize=2.0) xplt.pause(1) raw_input("enter") ########################################## # dynamik2.py # # This one also works with a dynamic plot # (with a bit of flickering ...) # Also works under windows (as no pause(1) is used). # Due to the redraw only the last set of points is # shown, but this can easily be repaired by storing # the previous points. # from Numeric import * import scipy import scipy.xplt as xplt xplt.winkill() xplt.window(wait=1) t=arange(0.0,0.5,0.1) mat=reshape(arange(100),(10,10)) xplt.limits(0,2000,0,2000) for i in arange(0,2000,1): mat[0,0]=i ew,ef=scipy.linalg.eig(mat) xplt.fma() xplt.plg(ew.real,i*ones(len(ew)),type=0,marker=1,msize=4.0) raw_input("enter") ########################################## # dynamik3.py # # Here we would like a second plot to show # continuously a plot of the matrix. # However, it is only shown at the very end # # Only the pause(1), which does not work under windows # will lead to the right result. from Numeric import * import scipy import scipy.xplt as xplt xplt.winkill() xplt.window(wait=1) t=arange(0.0,0.5,0.1) mat=reshape(arange(100),(10,10)) xplt.limits(0,2000,0,2000) for i in arange(0,2000,1): xplt.window(0) mat[0,0]=i ew,ef=scipy.linalg.eig(mat) xplt.fma() xplt.plg(ew.real,i*ones(len(ew)),type=0,marker=1,msize=4.0) xplt.window(1) xplt.pli(mat) #xplt.pause(1) # ONLY with this we get a dynamic pli(mat). raw_input("enter")

1 0

special.erf underflow
by Arnd Baecker May 19, 2003

May 19, 2003

Hi, is there a way to prevent the error message in the following example In [32]: import scipy In [33]: scipy.special.erf(300) erfc underflow error Out[33]: 1.0 ((On the one hand it is nice to be informed about that - on the other hand the result -1.0 is just the one I would expect. Eg. for exp one just gets In [34]: scipy.special.exp(-10000) Out[34]: 0.0 Acutally for exp this is better than Numeric's exp which gives a math range error in this case: In [35]: Numeric.exp(-1000) OverflowError: math range error )) Arnd

3 6

Once more with feeling
by Joe Cooper May 18, 2003

May 18, 2003

Hi all, This time I think I've really got it working. Apologies for the previous two bungled emails (I had the wrong python2 package installed, how embarrassing). Thanks for your patience. -- Joe Cooper <joe(a)enthought.com>

2 4

AttributeError: 'ft_spec' module has no attribute 'EXC_WRAP'
by Nils Wagner May 16, 2003

May 16, 2003

This is the output of /usr/local/bin/python setup.py build wagner@lisa:~/cvs/scipy> /usr/local/bin/python setup.py build ### Little Endian detected #### Traceback (most recent call last): File "setup.py", line 111, in ? setup_package() File "setup.py", line 90, in setup_package config_list += map(get_separate_package_config,separate_packages) File "setup.py", line 74, in get_separate_package_config return get_package_config(name,'') File "setup.py", line 68, in get_package_config config = mod.configuration(parent) File "Lib_chaco/freetype/setup_freetype.py", line 23, in configuration mod = build_ft._ft_ext_module() File "/export/home/wagner/cvs/scipy/Lib_chaco/freetype/src/build_ft.py", line 22, in _ft_ext_module mod = library_weave.weave_wrappers(mod) File "/export/home/wagner/cvs/scipy/Lib_chaco/freetype/src/library_weave.py", line 23, in weave_wrappers ext_code = ft_spec.EXC_WRAP % ext_code AttributeError: 'ft_spec' module has no attribute 'EXC_WRAP' Nils

2 4

problem with plt
by Paul Casteels May 16, 2003

May 16, 2003

This is my configuration : Z:\>python ActivePython 2.2.1 Build 222 (ActiveState Corp.) based on Python 2.2.1 (#34, Apr 15 2002, 09:51:39) [MSC 32 bit (Intel)] on win32 Type "help", "copyright", "credits" or "license" for more information. >>> import Numeric >>> Numeric.__version__ '23.0' >>> import wxPython >>> wxPython.__version__ '2.4.0.7' >>> import scipy >>> scipy.__version__ '0.2.0_alpha_196.4128' If I try : >>> import gui_thread >>> from scipy import plt >>> plt.plot((1,2,3)) I get a new window labeld Figure0 immediately followed by an Application error : The instruction at 0x00c0577b referenced memory at 0x00000000 the memory could not be read. Do I need to use other versions for wxPython or Numeric ? Best regards and thanks, Paul Casteels Paul.Casteels(a)ua.ac.be Tel: +32.3.8202455 Fax: +32.3.8202470 University of Antwerp Dpt.Physics Universiteitsplein 1 B-2610 Wilrijk Belgium

4 4

constrained optimization
by John Hunter May 15, 2003

May 15, 2003

I need to do a bounded nonlinear minimization problem. I've been using scipy.optimize.leastsq, and apply artificial bounds by increasing the error when then params escape their bounds, but this isn't working well enough. I see that scipy.optimize.anneal has bounds, but am wondering if there are any bounded optimizers available for general nonlinear least squares, like a bounded Lev-Marq alogirithm. Thanks! John Hunter

5 6

Iterative solution of systems of linear equations
by Nils Wagner May 15, 2003

May 15, 2003

Hi all, I wonder if scipy will support iterative solvers like (CG, BiCG, BiCGstab, CGS, TFQMR, GMRES, etc) ? Nils

1 0

Installing Scipy
by imaginee1＠gmx.net May 15, 2003

May 15, 2003

Hi everyone, I would like to install Scipy but so far, I haven't gotten anywhere. System specifications are on the bottom. Since I'm using debian linux I first tried the debian package which is for python2.1, but this was quite unstable and I want to use scipy under python2.2 . So here goes: Binary install: >>> from scipy.xplt import * Traceback (most recent call last): File "<stdin>", line 1, in ? File "/home/martin/tmp/linux2/2.2/lib/python2.2/site-packages/scipy_base/ppimport.py", line 211, in __getattr__ File "/home/martin/tmp/linux2/2.2/lib/python2.2/site-packages/scipy_base/ppimport.py", line 183, in _ppimport_importer AssertionError: <module 'scipy.xplt' from '/usr/lib/python2.2/site-packages/scipy/xplt/__init__.pyc'> Source Compilation: signum:/usr/src/SciPy-0.2.0_alpha_196.4120# python setup.py install Traceback (most recent call last): File "setup.py", line 111, in ? setup_package() File "setup.py", line 90, in setup_package config_list += map(get_separate_package_config,separate_packages) File "setup.py", line 74, in get_separate_package_config return get_package_config(name,'') File "setup.py", line 67, in get_package_config mod = __import__('setup_'+os.path.basename(name)) ImportError: No module named setup_scipy_distutils CVS Compilation: signum:/usr/src/cvs/scipy# python setup.py install ### Little Endian detected #### Traceback (most recent call last): File "setup.py", line 111, in ? setup_package() File "setup.py", line 90, in setup_package config_list += map(get_separate_package_config,separate_packages) File "setup.py", line 74, in get_separate_package_config return get_package_config(name,'') File "setup.py", line 68, in get_package_config config = mod.configuration(parent) File "Lib_chaco/kiva/setup_kiva.py", line 60, in configuration config_list.append(get_package_config('agg','kiva')) File "Lib_chaco/kiva/setup_kiva.py", line 37, in get_package_config config = mod.configuration(parent) File "/usr/src/cvs/scipy/Lib_chaco/kiva/agg/setup_agg.py", line 15, in configuration mod = build_agg.kiva_ext_module() File "/usr/src/cvs/scipy/Lib_chaco/kiva/agg/src/build_agg.py", line 24, in kiva_ext_module mod = affine_matrix_weave.weave_wrappers(mod) File "/usr/src/cvs/scipy/Lib_chaco/kiva/agg/src/affine_matrix_weave.py", line 65, in weave_wrappers result = zeros(6,typecode=Float) TypeError: zeros() takes no keyword arguments GREAT! System specifications: # python -c 'import os,sys;print os.name,sys.platform' posix linux2 (Debian woody) # gcc -v Reading specs from /usr/lib/gcc-lib/i386-linux/2.95.4/specs gcc version 2.95.4 20011002 (Debian prerelease) # g77 --version GNU Fortran 0.5.25 20010319 (prerelease) # python -c 'import sys;print sys.version' 2.2.1 (#1, Sep 7 2002, 14:34:30) # python -c 'import Numeric;print Numeric.__version__' 21.0 # f2py -v 2.32.225-1419 # python scipy_core/scipy_distutils/system_info.py atlas_info: FOUND: libraries = ['lapack', 'f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/lib/3dnow/atlas', '/usr/lib/3dnow'] blas_info: NOT AVAILABLE blas_src_info: NOT AVAILABLE dfftw_info: NOT AVAILABLE dfftw_threads_info: NOT AVAILABLE djbfft_info: NOT AVAILABLE fftw_threads_info: NOT AVAILABLE lapack_info: NOT AVAILABLE lapack_src_info: NOT AVAILABLE sfftw_info: NOT AVAILABLE sfftw_threads_info: NOT AVAILABLE x11_info: FOUND: libraries = ['X11'] library_dirs = ['/usr/X11R6/lib'] include_dirs = ['/usr/X11R6/include'] # python scipy_core/scipy_distutils/command/build_flib.py Not found/available: Absoft Fortran compiler Not found/available: SGI Fortran compiler Not found/available: Forte Fortran compiler Not found/available: Sun Fortran compiler Not found/available: Intel Fortran compiler Not found/available: Itanium Fortran compiler Not found/available: NAG Fortran compiler Not found/available: Compaq Fortran compiler Not found/available: VAST Fortran compiler Not found/available: HP Fortran compiler Not found/available: F Fortran compiler Not found/available: Lahey Fortran compiler Gnu version='0.5.25' F77='g77' F77FLAGS=' -Wall -fno-second-underscore -fPIC ' F77OPT=' -O3 -funroll-loops ' LIBS='-lg2c-pic' Thanks for reading this, Nikolai -- +++ GMX - Mail, Messaging & more http://www.gmx.net +++ Bitte lächeln! Fotogalerie online mit GMX ohne eigene Homepage!

3 5

io.write_array
by Nils Wagner May 15, 2003

May 15, 2003

Hi all, I wonder, if I can write the output of w = linalg.eigvals(A+mu*S,B) for varying mu to a file.with the following structure Each row should be related to a fixed mu - the columns should represent the eigenvalues mu0 w[0] ... w[n] mu1 w[0] .. w[n] mum w[0] ... w[n] How can I do that with io.write_array ? Nils

1 0

RE: unable to import scipy
by Vernon May 14, 2003

May 14, 2003

With regards to my previous post: I was generating errors when trying to import scipy >>>from scipy import * At that time I was using scipy 0.2.0 and numeric 21. I just installed numeric 22, and now I am generating the same error. Thanks again for the help and I hope my problem helps other people with similar issues. __________________________________ Do you Yahoo!? The New Yahoo! Search - Faster. Easier. Bingo. http://search.yahoo.com

1 0