Mailman 3 November 2007 - SciPy-User

instrumentation control question
by Chris Lee 12 Nov '07

12 Nov '07

Hi, I know this is a bit off-topic but I wondered if anyone here had any experience with the following problem. Part of our work involves controlling a bunch of lab instrumentation, which so far has worked beautifully using python + scipy/numpy + pyVISA. Then we got a stepper motor which does not comes with an activeX control and (as far as I can tell) nothing else. This means using the com extensions for win32 and trying to communicate with the activeX control. All is not progressing well (we can find the control but send no data to it, and it certainly won't talk back to us). Does anyone have any experience with this? If so, would you mind sending me some code snippets for gaining control of the activeX element? Thanks very much and my appologies for the not-so-on-topic post (I actually think instrumentation control would be a great addition to scipy and if I become competent may look at contributing code in that direction). Cheers Chris

3 2

random gaussian array
by sandric ionut 12 Nov '07

12 Nov '07

Hi: I want to create an random Gaussian 2D array with an exact size. For example I want my array to look like: 2 3 4 0 1 1 2 3 3 4 1 1 3 2 5 6 7 8 all the values should be randomly generated from a Gaussian distribution with mean and standard deviation Is it possible to do it with scipy or numpy? Thank you Ionut __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

Bias in numpy.random.multivariate_normal?
by John Reid 12 Nov '07

12 Nov '07

Could someone confirm I've got this right? In a multivariate normal (MVN) the covariance is E(X.X') - E(X).E(X'). I sample many times from a MVN with given mean and covariance. I then calculate the covariance of the sample. This covariance matrix should show no bias to be larger or smaller than the distribution's covariance but I see that some entries are always larger and others are smaller. See this code: from numpy.random import multivariate_normal from numpy.linalg import inv from numpy import outer mu = [10.,10.] precision = [[.6, .3], [.4, .3]] sigma = inv(precision) sample_size = 10000 num_tests = 100 for test in xrange(num_tests): samples = multivariate_normal(mu, sigma, [sample_size]) sample_mean = samples.sum(axis=0)/sample_size #print (sample_mean - mu) > 0. sample_covariance = sum(outer(sample-sample_mean, sample-sample_mean) for sample in samples) / sample_size print (sample_covariance - sigma) > 0 'True' is printed when the entry is larger and False otherwise Did I get this wrong or is this acceptable behaviour for a random variate generator or is it a bug? Thanks, John.

4 7

Problem building SciPy-0.6.0
by Rich Shepard 12 Nov '07

12 Nov '07

I upgraded my notebook from Slackware-11.0 to -12.0 so libc, gcc, etc. are different. I've built ATLAS (and copied the *.a and *.so libraries to /usr/local/lib), then sucessfully built and installed NumPy-1.0.4. But, when I try 'python setup.py install' from the scipy-0.6.0 directory the attempt immediately fails: Traceback (most recent call last): File "setup.py", line 53, in <module> setup_package() File "setup.py", line 28, in setup_package from numpy.distutils.core import setup File "/usr/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module> import linalg File "/usr/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module> from linalg import * File "/usr/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module> from numpy.linalg import lapack_lite ImportError: liblapack.so: cannot open shared object file: No such file or directory liblapack.so is in both /usr/local/lib and ~/python/ATLAS/ATLAS-linux/lib/. What am I missing here? Rich -- Richard B. Shepard, Ph.D. | Integrity Credibility Applied Ecosystem Services, Inc. | Innovation <http://www.appl-ecosys.com> Voice: 503-667-4517 Fax: 503-667-8863

1 1

Sandbox TimeSeries: request
by dieter h. 12 Nov '07

12 Nov '07

Hi Pierre and Matt If I personally had the time to dig in to the module (and extension writing itself), I'd be incorporating this request myself. Functionality that would convert an array of epochs to datetime.datetime|mx.datetime|matplotlib.Dates objects would be a big plus. I suggest it in case you might also find it worthy. In the mean time I'll get by. :-) Thank you -Dieter

2 1

Problem installing scipy 0.6.0 on PPC OS X 10.5 Leopard.
by Daniel Yarlett 10 Nov '07

10 Nov '07

hi, apologies if this is not the appropriate forum for posting this, but i am having some trouble building and installing scipy 0.6.0 on a PPC Mac that i have just upgraded to OS X 10.5 (this version of scipy previously worked fine on the same machine under OS X 10.4). numpy 1.0.4 seems to install fine with "python setup.py install". however, when i try to install scipy 0.6.0 i get the output (tail only) included at the end of this email. here are my compiler versions: gcc -- version gcc (GCC) 4.3.0 20071026 (experimental) gfortran -- version GNU Fortran (GCC) 4.3.0 20071026 (experimental) i remember that on previous versions of OS X you could use gcc_select to specify v3.4 of gcc to make the build go through. unfortunately gcc_select seems to have been deprecated in Leopard so i haven't been able to try that. don't know if that's a red herring or not. also, fwiw, i just upgraded an intel macbook prop from OS X 10.4 to 10.5 and had no problems at all with the pre-existing versions of numpy and scipy. anyway, any help would be greatly appreciated! best, dan. -- END OF OUTPUT FROM "python setup.py install" for scipy 0.6.0: Could not locate executable f95 customize AbsoftFCompiler Could not locate executable f90 Could not locate executable f77 customize IBMFCompiler Could not locate executable xlf90 Could not locate executable xlf customize IntelFCompiler Could not locate executable ifort Could not locate executable ifc customize GnuFCompiler Found executable /usr/local/bin/g77 gnu: no Fortran 90 compiler found gnu: no Fortran 90 compiler found customize GnuFCompiler gnu: no Fortran 90 compiler found gnu: no Fortran 90 compiler found customize GnuFCompiler using build_clib building 'superlu_src' library compiling C sources C compiler: gcc -arch ppc -arch i386 -isysroot /Developer/SDKs/ MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp-precomp - mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 compile options: '-DUSE_VENDOR_BLAS=1 -c' gcc: scipy/linsolve/SuperLU/SRC/colamd.c gcc: unrecognized option '-no-cpp-precomp' cc1: error: unrecognized command line option "-arch" cc1: error: unrecognized command line option "-arch" cc1: error: unrecognized command line option "-Wno-long-double" gcc: unrecognized option '-no-cpp-precomp' cc1: error: unrecognized command line option "-arch" cc1: error: unrecognized command line option "-arch" cc1: error: unrecognized command line option "-Wno-long-double" error: Command "gcc -arch ppc -arch i386 -isysroot /Developer/SDKs/ MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp-precomp - mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 - DUSE_VENDOR_BLAS=1 -c scipy/linsolve/SuperLU/SRC/colamd.c -o build/ temp.macosx-10.3-fat-2.5/scipy/linsolve/SuperLU/SRC/colamd.o" failed with exit status 1

3 2

Inverse of complex matrix
by Roger Herikstad 10 Nov '07

10 Nov '07

Dear scipy users, I am trying to invert a 12 by 12 complex matrix R using scipy.linalg.inv(), but the round-off errors seem a bit too big. If I do scipy.dot(Ri,R), the first row of the resulting matrix looks like this: array([ 0.953125 +0.j , 0.06542969+0.01171875j, 0.08154297+0.15234375j, -0.078125 -0.09375j , 0.19140625-0.0859375j , 0.08203125+0.140625j , -0.08398438+0.0078125j , -0.02539062-0.03125j , -0.06738281+0.015625j , -0.04736328+0.15820312j, -0.09057617+0.02832031j, 0.109375 -0.0625j ]) Is this normal? Inverting a 2x2 matrix produces the expected result. I've attached a pickle of my matrix. Thanks! ~ Roger

4 4

Submitting new code
by Chad Kidder 10 Nov '07

10 Nov '07

I've got some file import classes that I am working on and would eventually like to contribute back to scipy, probably the io package. I've done a little bit of searching and didn't find anything really descriptive. Would anyone like to give me / point to a bit of information on how to go about this? --Chad Kidder

2 1

Problem installing scipy 0.6.0 on PPC OS X 10.5 Leopard.
by Daniel Yarlett 10 Nov '07

10 Nov '07

here's what i get when i type "python -c "import sys; print sys.path": ['', '/Library/Frameworks/Python.framework/Versions/2.5/lib/ python25.zip', '/Library/Frameworks/Python.framework/Versions/2.5/lib/ python2.5', '/Library/Frameworks/Python.framework/Versions/2.5/lib/ python2.5/plat-darwin', '/Library/Frameworks/Python.framework/Versions/ 2.5/lib/python2.5/plat-mac', '/Library/Frameworks/Python.framework/ Versions/2.5/lib/python2.5/plat-mac/lib-scriptpackages', '/Library/ Frameworks/Python.framework/Versions/2.5/lib/python2.5/lib-tk', '/ Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/lib- dynload', '/Library/Frameworks/Python.framework/Versions/2.5/lib/ python2.5/site-packages'] i don't see anything of the form */Extras/* in there, although site- packages is the last in the path. but when i run my default version of python and import numpy i get version 1.0.4, which is the latest one i installed, and probably isn't the default version that comes with leopard: ramscar-lab5:~ yarlett$ python -c "import numpy; print numpy.__version__" 1.0.4 i can try setting $PYTHONPATH if you still think that might help. if so i'd appreciate it if you could let me know what i should set it to. thanks for the speedy response! dan. > Are you using the builtin python? If so, you will need to set your > PYTHONPATH to point to the directory where numpy 1.0.4 is installed. > The reason is that the builtin python comes with python 1.0.1, which > won't build scipy with the gfortran you have. > > There is a somewhat odd issue on Leopard. If you look at sys.path: > > python -c "import sys; print sys.path" > > You will see that the directory where the builtin numpy is (I think it > is something like */Extras/*) is before the site-packages directory > where your numpy 1.0.4 would be installed (unless you used to --home > or --prefix options when running python setup.py install). This means > that unless you set PYTHONPATH you will always get Leopards older > numpy, which will cause scipy to fail on building. > > Note, this is very different than on Tiger where the builtin python > did not include many third party packages. Leopard includes quite a > few (for better or worse). > > Brian > > On Nov 9, 2007 2:33 PM, Daniel Yarlett <yarlett <at> > psych.stanford.edu> wrote: > > hi, > > > > apologies if this is not the appropriate forum for posting this, > but i > > am having some trouble building and installing scipy 0.6.0 on a PPC > > Mac that i have just upgraded to OS X 10.5 (this version of scipy > > previously worked fine on the same machine under OS X 10.4). > > > > numpy 1.0.4 seems to install fine with "python setup.py install". > > > > however, when i try to install scipy 0.6.0 i get the output (tail > > only) included at the end of this email. > > > > here are my compiler versions: > > > > gcc -- version > > gcc (GCC) 4.3.0 20071026 (experimental) > > > > gfortran -- version > > GNU Fortran (GCC) 4.3.0 20071026 (experimental) > > > > i remember that on previous versions of OS X you could use > gcc_select > > to specify v3.4 of gcc to make the build go through. unfortunately > > gcc_select seems to have been deprecated in Leopard so i haven't > been > > able to try that. don't know if that's a red herring or not. > > > > also, fwiw, i just upgraded an intel macbook prop from OS X 10.4 to > > 10.5 and had no problems at all with the pre-existing versions of > > numpy and scipy. > > > > anyway, any help would be greatly appreciated! > > > > best, > > > > dan. > > > > -- > > > > END OF OUTPUT FROM "python setup.py install" for scipy 0.6.0: > > > > Could not locate executable f95 > > customize AbsoftFCompiler > > Could not locate executable f90 > > Could not locate executable f77 > > customize IBMFCompiler > > Could not locate executable xlf90 > > Could not locate executable xlf > > customize IntelFCompiler > > Could not locate executable ifort > > Could not locate executable ifc > > customize GnuFCompiler > > Found executable /usr/local/bin/g77 > > gnu: no Fortran 90 compiler found > > gnu: no Fortran 90 compiler found > > customize GnuFCompiler > > gnu: no Fortran 90 compiler found > > gnu: no Fortran 90 compiler found > > customize GnuFCompiler using build_clib > > building 'superlu_src' library > > compiling C sources > > C compiler: gcc -arch ppc -arch i386 -isysroot /Developer/SDKs/ > > MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp- > precomp - > > mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 > > > > compile options: '-DUSE_VENDOR_BLAS=1 -c' > > gcc: scipy/linsolve/SuperLU/SRC/colamd.c > > gcc: unrecognized option '-no-cpp-precomp' > > cc1: error: unrecognized command line option "-arch" > > cc1: error: unrecognized command line option "-arch" > > cc1: error: unrecognized command line option "-Wno-long-double" > > gcc: unrecognized option '-no-cpp-precomp' > > cc1: error: unrecognized command line option "-arch" > > cc1: error: unrecognized command line option "-arch" > > cc1: error: unrecognized command line option "-Wno-long-double" > > error: Command "gcc -arch ppc -arch i386 -isysroot /Developer/SDKs/ > > MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp- > precomp - > > mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 - > > DUSE_VENDOR_BLAS=1 -c scipy/linsolve/SuperLU/SRC/colamd.c -o build/ > > temp.macosx-10.3-fat-2.5/scipy/linsolve/SuperLU/SRC/colamd.o" failed > > with exit status 1 > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user <at> scipy.org > > http://projects.scipy.org/mailman/listinfo/scipy-user > >

1 0

sandbox.timeseries missing cseries
by D Hering 09 Nov '07

09 Nov '07

Hi. Just to let you know that sandbox.timeseries __init__.py is missing "import cseries" and in "__all__ = [...,'cseries']". dieter

3 2