Mailman 3 February 2002 - SciPy-Dev

log plots
by Heiko Henkelmann 03 Mar '02

03 Mar '02

Hello There, are there any plans to include log plots in any of the plot modules in the future? Or did I miss anything in the current version? Heiko

3 6

scipy import failures, what can we do about it?
by Pearu Peterson 28 Feb '02

28 Feb '02

Hi, I have several times noticed that if scipy import fails then it shows 1) unrelated error messages and 2) the corresponding message of the true reason of the failure is either hidden or presented in an unnoticable way. The following message illustrates a particular case of all that: ---------- Forwarded message ---------- Date: Tue, 26 Feb 2002 15:43:04 +0100 From: Nils Wagner <nwagner(a)mecha.uni-stuttgart.de> Reply-To: scipy-user(a)scipy.org To: "scipy-user(a)scipy.net" <scipy-user(a)scipy.net> Subject: [SciPy-user] Latest CVS - import scipy fails >>> import scipy /usr/local/lib/python2.1/site-packages/scipy/linalg/clapack.so: undefined symbol: clapack_sgetrf ^^^^^^^^^^^^^^^ - this is the true reason for import failure Warning: FFT package not found. Some names will not be available ^^^^^^^^^^^^^^^ - this warning is completely wrong. As a result, the FFT package gets a bad name that it does not serve (I a bit suspect that this is what happend to fftw). Traceback (most recent call last): <snip> ImportError: cannot import name eig >>> ^^^^^^^^^^^^^ - the exception raised here is not related to the true import error above and one may start looking for errors in wrong places. My suggesion is to review how scipy imports its submodules. Though the current one may work perfectly if nothing is wrong but it behaves unpredictable if something *is* wrong. Scipy should be able to fail at least with a sensible message indicating the right direction to look for errors. The current error messages are just too misleading. If we will not fix this, we are going to get continuously scipy import failure reports that have very little to do with scipy bugs, the failures are more likely to be of incorrectly installed third-party libraries like atlas, fftw, etc. Currently, I don't have a good solution to suggest that would transparently fix all the importing issues. However, I would like to learn what others feel about this issue so that I could know if it makes sense to work something out or not (may be some of you already have a good solution?). Thanks, Pearu

3 5

Missing functions --> Matrix functions, generalized eigenvalue problem, iterative solvers, CWT (Continuos wavelet transform)
by Nils Wagner 27 Feb '02

27 Feb '02

Hi all, The current output of >>> help (linalg) is Linear algebra routines. Solving Linear Systems: inv --- Find the inverse of a square matrix solve --- Solve a linear system of equations det --- determinant of a matrix pinv --- Moore-Penrose pseudo inverse (using least-squares) pinv2 --- Moore-Penrose pseudo inverse (using SVD) lstsq --- Least-squares solve Matrix Factorizations: lu --- LU decomposition cholesky --- Cholesky factorization qr --- QR factorization schur --- Schur decomposition rsf2csf --- Real to complex schur form. norm --- vector and matrix norm eig --- eigenvectors and eigenvalues eigvals --- only eigenvalues svd --- singular value decomposition Matrix Functions expm --- exponential (using Pade approximation) cosm --- cosine sinm --- sine tanm --- tangent coshm --- hyperbolic cosine sinhm --- hyperbolic sine tanhm --- hyperbolic tangent funm --- arbitrary function >>> Is there any progress in implementing the matrix logarithm logm ? Has anyone written some functions concerning the generalized eigenvalue problem (or even more general polynomial eigenvalue problems), that is A x = \lambda B x (\lambda^n A_n + \dots + \lambda A_1 + A_0) x = 0 Are there any iterative solvers (CG, GMRES, ...) ? Nils I am also looking for discrete and continuous wavelet transforms.

1 0

New in fastumath ~ (means conjugate on floats and complex numbers)
by Travis Oliphant 27 Feb '02

27 Feb '02

I finally realized that with a simple change we can use the unary operator on floats and complex numbers to mean complex conjugation. I've made the simple change in the CVS version of fastumath. So, in scipy complex-conjugation is as simple as ~a if a is a complex number (or a float). The only problem is that if a is an integer it still means bitwise inversion. Is the added convenience worth the possible confusion? The problem is that complex conjugation happens all the time, but bitwise inversion rarely. -Travis

8 11

special/cephes: undefined symbol: psi_
by Pearu Peterson 25 Feb '02

25 Feb '02

Hi, When I try the latest CVS build then I get ImportError: scipy/special/cephes.so: undefined symbol: psi_ It seems that cdflib misses a file psi.f since cdflib/apser.f uses EXTERNAL psi and psi_ is defined nowhere. Do we need a C wrapper containing double psi_( double a ) { return psi(a); } ? Pearu

2 1

About stats/distributions.xls
by Pearu Peterson 25 Feb '02

25 Feb '02

Hi again, stats/distributions.xls is Excel, hmm, it most contain very useful information but how I can see it if I don't have Excel? :( Would it be possible to convert stats/distributions.xls to something more readable? PDF, TXT, TEX, may be. Thanks, Pearu

2 1

Problem with CVS co
by Janko 25 Feb '02

25 Feb '02

Hi, I thought that the problem with the open file limits were solved? Tried to checkout this morning all the great work you had done, I get: cvs server: cannot open .cvsignore: Too many open files cvs server: cannot open .cvswrappers: Too many open files cvs [server aborted]: cannot open CVS/Repository: Too many open files The co command is cvs -z7 -q up -P -d Is this a problem with my cvs setup? Should I try a completly new co? __Janko

4 4

RE: CVS Build Instructions
by Travis N. Vaught 25 Feb '02

25 Feb '02

F- I completely agree. I think the 'edit' feature was a casualty of some unrelated edition to the SciPy site dealing with menuing or something. The behaviour should definitely be for an 'Owner' of a document (its creator) to be able to edit--even after publication. This should be fixed in the next generation of the site along with other improvements and fixes. We will hopefully migrate to it later next week. BTW if you have any other suggestions about the site, they are welcome. Here are some thoughts that are being bandied about: o Roundup bug-tracker instead of the currently ignored sourceforge tracker o Better facility for community comments to docs and the site in general--perhaps just the mailing lists o get rid of duplicate functionality (Wikis are similar to comments (now disabled) are similar to archived mailing lists--we should probably use one for discussion (mailing lists) one for co-generated docs (Probably not wikis but I'm not sure--could be some CMF document in users' folders) I've posted this to the scipy-dev list to invite (gingerly worded) comments about how the scipy.org site would ideally work. Thanks, TV > -----Original Message----- > From: Fernando Perez [mailto:fperez@pizero.colorado.edu] > Sent: Wednesday, February 20, 2002 12:46 PM > To: Travis N. Vaught > Subject: Re: CVS Build Instructions > > > Hi Travis, > > a quick question/comment: > > > http://www.scipy.org/site_content/tutorials/user_folder_instructions > > seems to give no way of *editing* an existing document. For > example if I now > want to fix some typos in the CVS instructions doc, if I > understand correctly > the only option is to upload a full new one, delete the old and > go through the > editorial 'quarantine'. It seems a bit cumbersome both for you > guys and for > users, especially when one only wants to do minor editing. > > I wonder if it would be possible to add an 'edit' button that > would open the > document in a simple html text field, nothing fancy. I understand > this creates > a run around your editorial oversight, but I'd be willing to say > that scipy is > a small and focused enough group that abuse shouldn't be an > issue, at least > for quite a while. > > Just some thoughts, please correct me if any of my above > assumptions are wrong > or I missed something in the instructions. > > Best regards, > > Fernando.

5 5

Re: Tests in Scipy
by Tom Loredo 25 Feb '02

25 Feb '02

Travis wrote: > Adding from fastumath import * > > after the from Numeric import * > > fixes the problem. Where did you do this fix, Travis? scipy.test() with my build of 0.1 on Solaris has 14 failures, all related to infinitude/NaN checking. Thanks, Tom

2 1

NAN in special/cephes breaks CVS build
by Pearu Peterson 25 Feb '02

25 Feb '02

Hi, The latest CVS does not build on my debian box and it fails with the following message: /home/users/pearu/src_cvs/scipy/special/cephes/iv.c:85: `NAN' undeclared (first use in this function) NAN is defined in math.h that includes bits/nan.h. So I tried to fix it by adding #include <math.h> to scipy/special/cephes/mconf.h. But it has no effect :(. I am confused. Does it work for you? Pearu

3 7