From arnd.baecker at web.de  Mon Mar  1 10:20:02 2004
From: arnd.baecker at web.de (Arnd Baecker)
Date: Mon, 1 Mar 2004 16:20:02 +0100 (CET)
Subject: [SciPy-user] help and lazy-import (?)
In-Reply-To: <Pine.LNX.4.51.0402271206240.32377@ptpcp2.phy.tu-dresden.de>
References: <Pine.LNX.4.44.0402270450590.31709-100000@scipy.org>
 <Pine.LNX.4.51.0402271206240.32377@ptpcp2.phy.tu-dresden.de>
Message-ID: <Pine.LNX.4.51.0403011617090.28568@ptpcp2.phy.tu-dresden.de>

Hi,

On Fri, 27 Feb 2004, Arnd Baecker wrote:
> On Fri, 27 Feb 2004, Pearu Peterson wrote:
> > On Fri, 27 Feb 2004, Arnd Baecker wrote:
> >
> > > there there seems to be something weird
> > > going on when trying to access help in the following way
> > >
> > > >>> import scipy
> > > >>> help("scipy.linalg")
> > > Help on instance of _ModuleLoader in scipy:
> > >
> > > scipy.linalg = <module 'scipy.linalg' from '/opt/python/lib/pyt...
> > > e-packages/scipy/linalg/__init__.pyc' [imported]>
> > >
> > > Doing another
> > > >>> help("scipy.linalg")
> > >
> > > one gets the expected help.

> > But in a traditional Python prompt
> >
> > >>> import scipy
> > >>> help("scipy.linalg")
> >
> > I get expected help. This is because ppimport redefines built-in help:
> >
> > >>> help
> > Type help () for interactive help, or help (object) for help about object.
> > >>> help.__module__
> > 'scipy_base.ppimport'

we have had a further look into this and the status is as follows

- setting  postpone_import = 0 in (for example)
  scipy/xplt/info_xplt.py
  gives
 >>> import scipy
 Traceback (most recent call last):
 [...]
 from basic import *
  File "/opt/python/lib/python2.3/site-packages/scipy/special/basic.py",
line 10, in ?
    from scipy import factorial
  ImportError: cannot import name factorial

  We thought that setting `postpone_import = 0` means that the import
  of this module is not postponed so that the doc string should be
  available from the first call to help("scipy.xplt").
  Is this idea/approach just wrong ?

- We found that for
    >>> import scipy
    >>> help(scipy.xplt)
  the correct documentation is visible already on the first
  invokation of help.
  (In contrast to
    >>> import scipy
    >>> help("scipy.xplt")
  )

  Using the source to ppimport.py this is clear as
  in the first case the module is imported (in class _Helper)
  before calling the original _pydoc.help.
  However, for a string  there is no corresponding "ppimport"
  attribute and it gets passed directly to _pydoc.help

  Just to be sure that we understand things correctly:
  Is it true that the actual import in the second case,
  i.e. help("scipy.xplt"), is done when __repr__ from the class
  _ModuleLoader is called via the "normal" pydoc?
  And by this the doc string is available for the following
  help("scipy.xplt"), right?

  To us it seems to boil down to the question if there is
  a good way to import the corresponding module for the
  request help("scipy.xplt") in a similar way as it is done
  for help(scipy.xplt)?
  (Changing __repr__  to also output the docstring sounds like a
  horrible hack ..., i.e. there should be a better way)

Many thanks in advance for any suggestions/pointers/...,

Ulf Lorenz, Arnd Baecker



From pearu at scipy.org  Mon Mar  1 11:45:51 2004
From: pearu at scipy.org (Pearu Peterson)
Date: Mon, 1 Mar 2004 10:45:51 -0600 (CST)
Subject: [SciPy-user] help and lazy-import (?)
In-Reply-To: <Pine.LNX.4.51.0403011617090.28568@ptpcp2.phy.tu-dresden.de>
Message-ID: <Pine.LNX.4.44.0403011028500.1099-100000@scipy.org>



On Mon, 1 Mar 2004, Arnd Baecker wrote:

> we have had a further look into this and the status is as follows
> 
> - setting  postpone_import = 0 in (for example)
>   scipy/xplt/info_xplt.py
>   gives
>  >>> import scipy
>  Traceback (most recent call last):
>  [...]
>  from basic import *
>   File "/opt/python/lib/python2.3/site-packages/scipy/special/basic.py",
> line 10, in ?
>     from scipy import factorial
>   ImportError: cannot import name factorial
> 
>   We thought that setting `postpone_import = 0` means that the import
>   of this module is not postponed so that the doc string should be
>   available from the first call to help("scipy.xplt").
>   Is this idea/approach just wrong ?

No, setting `postpone_import = 0` should have worked. I look into it..
Ahh, I see, it is a chicken and egg problem..

> - We found that for
>     >>> import scipy
>     >>> help(scipy.xplt)
>   the correct documentation is visible already on the first
>   invokation of help.
>   (In contrast to
>     >>> import scipy
>     >>> help("scipy.xplt")
>   )
> 
>   Using the source to ppimport.py this is clear as
>   in the first case the module is imported (in class _Helper)
>   before calling the original _pydoc.help.
>   However, for a string  there is no corresponding "ppimport"
>   attribute and it gets passed directly to _pydoc.help

Right.

>   Just to be sure that we understand things correctly:
>   Is it true that the actual import in the second case,
>   i.e. help("scipy.xplt"), is done when __repr__ from the class
>   _ModuleLoader is called via the "normal" pydoc?
>   And by this the doc string is available for the following
>   help("scipy.xplt"), right?
> 
>   To us it seems to boil down to the question if there is
>   a good way to import the corresponding module for the
>   request help("scipy.xplt") in a similar way as it is done
>   for help(scipy.xplt)?
>   (Changing __repr__  to also output the docstring sounds like a
>   horrible hack ..., i.e. there should be a better way)

I have fixed this issue in CVS using different and I think more robust 
approach: `import scipy_base` will redefine help.__class__.__call__ method 
to support ppimport'ed modules. Now `help` works as expected for all 
of the following usage cases:

  help(scipy.linalg)
  help("scipy.linalg")
  exec 'help("scipy.linalg")'
  etc

I have tested this on Python 2.3.3 only though.

Regards,
Pearu


From arnd.baecker at web.de  Tue Mar  2 03:49:05 2004
From: arnd.baecker at web.de (Arnd Baecker)
Date: Tue, 2 Mar 2004 09:49:05 +0100 (CET)
Subject: [SciPy-user] help and lazy-import (?)
In-Reply-To: <Pine.LNX.4.44.0403011028500.1099-100000@scipy.org>
References: <Pine.LNX.4.44.0403011028500.1099-100000@scipy.org>
Message-ID: <Pine.LNX.4.51.0403020915520.11472@ptpcp2.phy.tu-dresden.de>

Hi Pearu,

On Mon, 1 Mar 2004, Pearu Peterson wrote:

[...]

> I have fixed this issue in CVS using different and I think more robust
> approach: `import scipy_base` will redefine help.__class__.__call__ method
> to support ppimport'ed modules. Now `help` works as expected for all
> of the following usage cases:
>
>   help(scipy.linalg)
>   help("scipy.linalg")
>   exec 'help("scipy.linalg")'
>   etc
>
> I have tested this on Python 2.3.3 only though.

Great - _many_ thanks for the fix!
Both
     >>> import scipy
     >>> help("scipy.xplt")
and
     In [1]: import scipy
     In [2]: help("scipy.xplt")
also work for me (I am also 2.3.3 as well).

There seems to be one case where I still see the module
loader when requesting help:

  Python 2.3.3 (#1, Feb  3 2004, 22:05:11)
  [GCC 3.2] on linux2
  Type "help", "copyright", "credits" or "license" for more information.
  >>> help("scipy.linalg")

I.e., no `import scipy` (or similar) has been done before.
I am not sure how important this case is
(eg. for doc-extracting tools ?).

Concerning ipython:
  - for the help itself it seems to work
  - for TAB completion one (sometimes) has to press TAB twice,
    for example:

In [1]: import scipy
In [2]: scipy.xplt.               # pressing TAB here gives:
scipy.xplt.__class__           scipy.xplt.__name__
scipy.xplt.__doc__             scipy.xplt.__repr__
scipy.xplt.__file__            scipy.xplt.__setattr__
scipy.xplt.__getattr__         scipy.xplt.__str__
scipy.xplt.__init__            scipy.xplt._ppimport_importer
scipy.xplt.__module__          scipy.xplt.test
In [2]: scipy.xplt.              # pressing TAB here then gives the
                                 # full list (omitted here ;-)

I am not sure what to do here. Anyway, this is only
here for completeness and should better be discussed
on the ipython mailing list (I know, Fernando, you are
super-busy right now, so we will try to have a quick look at that
tomorrow).

Best,

Arnd


From arnd.baecker at web.de  Tue Mar  2 05:15:51 2004
From: arnd.baecker at web.de (Arnd Baecker)
Date: Tue, 2 Mar 2004 11:15:51 +0100 (CET)
Subject: [SciPy-user] xplt issues
Message-ID: <Pine.LNX.4.51.0403021113350.11472@ptpcp2.phy.tu-dresden.de>

Hi,

we are using scipy.xplt quite happily as it works nicely both under
linux and windows. However, there are two issues which cause
problems in the portability of the code:

a) Plot a single point

##################################
from scipy.xplt import *

window(0)
limits(0.0,10.0,0.0,1.0)
# this does not work (i.e does not show anything) under linux:
x=array([0.5])
y=array([0.5])
plg(y,x,marker='\2',width=10)

# this does work under linux:
x=array([0.5,0.5])
y=array([0.5,0.5])
plg(y,x,marker='\2',width=10)
##################################

I don't know what the origin of this differing behaviour is
(gd library ?)

b) This example illustrates a `dynamic' plot

##################################
from scipy.xplt import *

figure()
x=arange(0.0,0.5,0.001)
hold('on')
for xval in x:
    print xval
    yval = sin(xval / 4)
    plot(xval, yval, 'x')
    pause(1)
##################################

The pause(1) is essential under linux so that the
points are shown one after another.
However under windows (enthought_python-2.3.exe)
the pause command leads to a crash of the running python interpreter.
Moreover pause(1) is not necessary under windows.
(I think the best would be if this was also true for the linux
version - would this be possible ??)

It would be great if these platform inconsistencies could get fixed!
Maybe we can spot the origin for a) once given a pointer in the right
direction. (Concerning b) we don't have the tools for windows ...).

Thanks in advance for any advice,

Arnd









From arnd.baecker at web.de  Tue Mar  2 05:51:05 2004
From: arnd.baecker at web.de (Arnd Baecker)
Date: Tue, 2 Mar 2004 11:51:05 +0100 (CET)
Subject: [SciPy-user] xplt issues
In-Reply-To: <Pine.LNX.4.51.0403021113350.11472@ptpcp2.phy.tu-dresden.de>
References: <Pine.LNX.4.51.0403021113350.11472@ptpcp2.phy.tu-dresden.de>
Message-ID: <Pine.LNX.4.51.0403021148370.11472@ptpcp2.phy.tu-dresden.de>


On Tue, 2 Mar 2004, Arnd Baecker wrote:

[...]
> b) This example illustrates a `dynamic' plot
>
> ##################################
> from scipy.xplt import *
>
> figure()
> x=arange(0.0,0.5,0.001)
> hold('on')
> for xval in x:
>     print xval
>     yval = sin(xval / 4)
>     plot(xval, yval, 'x')
>     pause(1)
> ##################################
>
> The pause(1) is essential under linux so that the
> points are shown one after another.

My statement is not true, as
####################################
from scipy.xplt import *

window(0,wait=1)     #the trick is the wait=1 keyword
x=arange(0.0,5.0,0.001)
hold('on')
for xval in x:
    print xval
    yval=sin(xval/4)
    plot(xval,yval,'x')
####################################
works fine under linux, even without the pause.
(we have not tested this under windows yet).
So maybe this is a solution/work-around for point b).

Arnd


From arnd.baecker at web.de  Tue Mar  2 06:06:11 2004
From: arnd.baecker at web.de (Arnd Baecker)
Date: Tue, 2 Mar 2004 12:06:11 +0100 (CET)
Subject: [SciPy-user] ANN: PlottingCanvas
Message-ID: <Pine.LNX.4.51.0403021204490.11472@ptpcp2.phy.tu-dresden.de>

Hi,

a while ago we addressed the question
of plotting  many thousand points quickly and
dynamically (i.e. one after another).
While scipy.xplt does a good job for this
in some cases we also wanted to have a nice
user-interface (with buttons and such)
and our choice came to wxPython.
However, none of the present plotting
packages fully fitted our specific needs and
thus we starting coding our own.
The resulting PlottingCanvas for wxPython can be found under
   http://www.physik.tu-dresden.de/~baecker/python/plot.html

The PlottingCanvas might illustrate what speed
is possible with wxPython (one question is whether
this speed would also be possible in chaco).

Any suggestions for further speed improvements
and comments are welcome!

Nikolai and Arnd


From bgoli at sun.ac.za  Tue Mar  2 17:19:34 2004
From: bgoli at sun.ac.za (Brett G. Olivier)
Date: Wed, 3 Mar 2004 00:19:34 +0200 (SAST)
Subject: [SciPy-user] gplt.plot3d:  grid data and isolines?
Message-ID: <1674.196.30.178.46.1078265974.squirrel@glue.jjj.sun.ac.za>

Hi

I am trying to plot 3D regular grid data with gplt.plot3d(Arr)
where Arr data is generated as:

1  1  x
1  2  y
2  1  x2
2  2  y2
... etc

plot3d plots the data points ok but the isolines simply join
1 2 y <---> 2 1 x2 etc. Gnuplot's way of sorting this out and quickly
plotting this type of data is to leave a blank line between the major grid
(defining the iso's):

1  1  x
1  2  y

2  1  x2
2  2  y2

I was wondering if there was a way of getting this behaviour using
scipy.gplt?

I have managed to emulate this behaviour relatively easily by altering the
_format(self,x) function in pyPlot.py at line 727

Currently if __format() gets INF it prints a ' - ' it would be really
useful if it printed a space '  ' instead (detail at end), then one could
"insert" an empty line into the plotdata file using INF in your array.

Any suggestions on how to do this otherwise are more than welcome.

Thanks in advance
Brett

++++++++++++++++++++++=

from pyPlot.py (line 727)

    def _format(self,x):
        #############################
        ## complications added to check for INF, -INF
        #############################
        #note: does not check for IND at the moment
        str = ''
        for val in x:
            if(val == INF or val == NEG_INF):
                #str = str + ' - '    # old
                str = str + '   '     # possible solution
            else:
                str = '%s %g' % (str,val)
        return str

-- 
Brett G. Olivier  (bgoli at sun dot ac dot za)

Biochemistry Dept, Stellenbosch University
Tel: +27-21-8085871    Fax: +27-218085863

"Don't think; let the machine do it for you!"
-- E. C. Berkeley


From bgoli at sun.ac.za  Wed Mar  3 09:48:15 2004
From: bgoli at sun.ac.za (Brett G. Olivier)
Date: Wed, 3 Mar 2004 16:48:15 +0200 (SAST)
Subject: [SciPy-user] gplt.plot3d:  grid data and isolines?
In-Reply-To: <1674.196.30.178.46.1078265974.squirrel@glue.jjj.sun.ac.za>
References: <1674.196.30.178.46.1078265974.squirrel@glue.jjj.sun.ac.za>
Message-ID: <1116.146.232.210.120.1078325295.squirrel@glue.jjj.sun.ac.za>

Ignore my potential workaround suggested in the last email. It turns out
to be a *very bad idea* as it messes up the 2D plots!

> Currently if __format() gets INF it prints a ' - ' it would be really
> useful if it printed a space '  ' instead

-- 
Brett G. Olivier  (bgoli at sun dot ac dot za)



From nwagner at mecha.uni-stuttgart.de  Fri Mar  5 02:54:29 2004
From: nwagner at mecha.uni-stuttgart.de (Nils Wagner)
Date: Fri, 05 Mar 2004 08:54:29 +0100
Subject: [SciPy-user] Hessenberg form / Improving the wrapper for dgehrd
Message-ID: <40483235.E6BA6739@mecha.uni-stuttgart.de>

Hi all,

I was going to build a wrapper for dgehrd in order to have 
a function similar to hess in Matlab.

I am quite sure, that my solution is not optimal. The output 
contains not only the Hessenberg form of the input matrix but
the array tau. How can I circumvent this ?

>man dgehrd
...
the elements below the first subdiagonal, with the  array  TAU, 
represent
the orthogonal matrix Q as a product of elementary reflectors. 
See Further Details.  LDA     (input) INTEGER The leading
dimension of the array A.  LDA >= max(1,N).

Therefore I would appreciate any hint for improving the 
interface according to the rules for a possible implementation in scipy.

Thanks in advance.

Nils

!    -*- f90 -*-
python module hess ! in
    interface  ! in :hess
        subroutine dgehrd(n,ilo,ihi,a,lda,tau,work,lwork,info)
!in:hess:dgehrd.f

            callstatement
(*f2py_func)(&n,&ilo,&ihi,a,&lda,tau,work,&lwork,&info)

            callprotoargument
int*,int*,int*,double*,int*,double*,double*,int*,int*

            integer intent(hide):: n=shape(a,1)
            integer intent(hide):: ilo=1
            integer intent(hide):: ihi=n
            double precision dimension(lda,*),intent(in,out,copy,out=h)
:: a
            integer
optional,check(shape(a,0)==lda),depend(a)::lda=shape(a,0)
            double precision dimension(n),intent(out) :: tau
            double precision dimension(MAX(1,lwork)),intent(out) :: work
            integer intent(in):: lwork
            integer intent(out):: info
        end subroutine dgehrd
    end interface
end python module hess
-------------- next part --------------
from scipy import *
import hess
#
# First example 5.4.6
# Biswa Nath Datta
# Numerical linear algebra and applications
# Brooks/Cole Publishing Company
# (1995)
#
a=zeros((3,3),Float)
a[0,0] = 1
a[0,1] = 2
a[0,2] = 5
a[1,0] = 3
a[1,1] = 7
a[1,2] = 9
a[2,0] = 2
a[2,1] = 5
a[2,2] = 3
#
# test example 5.4.7
# page 152 
# Biswa Nath Datta
# Numerical linear algebra and applications
# Brooks/Cole Publishing Company
# (1995)
#
a[0,0] = 0
a[0,1] = 1
a[0,2] = 1
a[1,0] = 1
a[1,1] = 2
a[1,2] = 1
a[2,0] = 1
a[2,1] = 1
a[2,2] = 1
#
# solution to 5.4.7
#
h1 = zeros((3,3),Float)
h1[0,1] = -sqrt(2)
h1[1,0] = -sqrt(2)
h1[1,1] = 2.5
h1[1,2] = 0.5
h1[2,1] = 0.5
h1[2,2] = 0.5



# workspace query
a,tau,work,info=hess.dgehrd(a,lwork=-1)
lwork=int(work[0])
h,tau,work,info=hess.dgehrd(a,lwork=lwork)
print 'Hessenberg form'
print h


def do_hess(mat):
   h,tau,work,info=hess.dgehrd(a,lwork=-1)
   lwork=int(work[0])
   h,tau,work,info=hess.dgehrd(a,lwork=lwork)
   return h

h2 = do_hess(a)
print h2

From sabbatini at dea.unian.it  Mon Mar  8 05:20:02 2004
From: sabbatini at dea.unian.it (Massimo Sabbatini)
Date: Mon, 08 Mar 2004 11:20:02 +0100
Subject: [SciPy-user] Scipy build under win32
Message-ID: <c2hhcl$l6g$1@sea.gmane.org>

Dear all,

I tried to build scipy (cvs checkout) under win32. I installed MinGW and 
  followed quite closely Pearu's instructions at
http://cens.ioc.ee/~pearu/scipy/BUILD_WIN32.html
However, at the final scipy build stage, when I type

$ msys.py python setup.py build

I obtain the following:

Traceback (most recent call last):
   File "setup.py", line 125, in ?
     setup_package(ignore_packages)
   File "setup.py", line 98, in setup_package
     parent_path=local_path)
   File "setup.py", line 59, in get_packages
     config = setup_module.configuration(*args)
   File "c:\build\scipy\Lib\integrate\setup_integrate.py", line 21, in 
configuration
     raise NotFoundError,'no blas resources found'
scipy_distutils.system_info.NotFoundError: no blas resources found

However, my blas library is correctly found when I run system_info.py

Any hints, suggestions ?

Massimo

-- 
-----------------------------------------------------------
Massimo Sabbatini
Department of Economics, Marche Polythecnic University (IT)
sabbatini (at) dea (dot) unian (dot) it



From pearu at scipy.org  Tue Mar  9 01:47:24 2004
From: pearu at scipy.org (Pearu Peterson)
Date: Tue, 9 Mar 2004 00:47:24 -0600 (CST)
Subject: [SciPy-user] Scipy build under win32
In-Reply-To: <c2hhcl$l6g$1@sea.gmane.org>
Message-ID: <Pine.LNX.4.44.0403090023190.5202-100000@scipy.org>



On Mon, 8 Mar 2004, Massimo Sabbatini wrote:

> Dear all,
> 
> I tried to build scipy (cvs checkout) under win32. I installed MinGW and 
>   followed quite closely Pearu's instructions at
> http://cens.ioc.ee/~pearu/scipy/BUILD_WIN32.html
> However, at the final scipy build stage, when I type
> 
> $ msys.py python setup.py build
> 
> I obtain the following:
> 
> Traceback (most recent call last):
>    File "setup.py", line 125, in ?
>      setup_package(ignore_packages)
>    File "setup.py", line 98, in setup_package
>      parent_path=local_path)
>    File "setup.py", line 59, in get_packages
>      config = setup_module.configuration(*args)
>    File "c:\build\scipy\Lib\integrate\setup_integrate.py", line 21, in 
> configuration
>      raise NotFoundError,'no blas resources found'
> scipy_distutils.system_info.NotFoundError: no blas resources found
> 
> However, my blas library is correctly found when I run system_info.py
> 
> Any hints, suggestions ?

How did you specify the locations of blas/lapack libraries? There are 
several ways to do it, for instance, I use

  export BLAS=~/src/blas/libfblas.a
  export LAPACK=~/src/LAPACK/liblapack.a

before running

  msys.py python setup.py build  

and I just checked, it works with the recent CVS.

Or may be you are using ATLAS? Then I would set

  export ATLAS=~/src/ATLAS/lib/WinNT_P4SSE2

before the build command, and again, it works here.
(Of course, the targets of the above environment variables should 
contain something sensible;)

When you run system_info.py, it shows the paths where the blas libraries
are looked for, something like

blas_info:
Replacing _lib_names[0]=='blas' with 'fblas'
( library_dirs = c:/cygwin/home/pearu/src/blas:c:\Py233\lib:C:\ )
( paths: c:/cygwin/home/pearu/src/blas\libfblas.a )
  FOUND:
    libraries = ['fblas']
    library_dirs = ['c:/cygwin/home/pearu/src/blas']
    language = f77

To enable these messages also for `setup.py build`, change the line #236

     verbosity = 1

in system_info.py to

     verbosity = 2

How the corresponding `blas_info:..` messages compare when running 
system_info.py or `setup.py build`?

Pearu



From jdhunter at ace.bsd.uchicago.edu  Tue Mar  9 07:50:40 2004
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Tue, 09 Mar 2004 06:50:40 -0600
Subject: [SciPy-user] os x error on import
Message-ID: <m2llma5gm7.fsf@mother.paradise.lost>


I am building CVS scipy on OS X 10.3.  The build goes fine, and I can
import scipy, or do "from scipy import array" w/o problems.  I can
"import scipy.signal" but if I do "help(scipy.signal)" or "from
scipy.signal import buttord" I get 

  Fatal Python error: Interpreter not initialized (version mismatch?)
  Abort

I have python compiled from src in /usr/local/lib/python2.3 and the
python framework is also installed.  I wonder if the message is
arising from the build using part of the framework when I just want to
be using the python from src.

>From the CVS root, I built scipy with

 > cd scipy_core
 > sudo python setup.py install # uses /usr/local/bin/python2.3
 > cd ..
 > sudo python setup.py install

Below is the head of the setup output; if anyone has any suggestions
I'm much obliged.

Thanks!
John Hunter

numpy_info:
  FOUND:
    define_macros = [('NUMERIC_VERSION', '"\\"23.1\\""')]
    include_dirs = ['/usr/local/include/python2.3']

SciPy Core Version 0.2.2_alpha_104.1544
****************************************************************
    Using fortran_libraries setup option is depreciated
    ---------------------------------------------------
    Use libraries option instead. Yes, scipy_distutils
    now supports Fortran sources in libraries.
****************************************************************
running install
running build
running config_fc
running build_src
building extension "scipy_base.fastumath" sources
  adding 'scipy_base/fastumathmodule.c' to sources.
building extension "scipy_base._compiled_base" sources
running build_py
package init file '/Users/jdhunter/python/src/scipy/scipy_core/scipy_distutils/tests/__init__.py' not found (or not a regular file)
package init file '/Users/jdhunter/python/src/scipy/scipy_core/scipy_base/tests/__init__.py' not found (or not a regular file)
package init file 'weave/tests/__init__.py' not found (or not a regular file)
package init file 'weave/tests/__init__.py' not found (or not a regular file)
package init file '/Users/jdhunter/python/src/scipy/scipy_core/scipy_distutils/tests/__init__.py' not found (or not a regular file)
package init file '/Users/jdhunter/python/src/scipy/scipy_core/scipy_base/tests/__init__.py' not found (or not a regular file)
package init file 'weave/tests/__init__.py' not found (or not a regular file)
package init file 'weave/tests/__init__.py' not found (or not a regular file)
running build_ext
customize UnixCCompiler
customize UnixCCompiler using build_ext
distutils.unixccompiler.UnixCCompiler does not provide version_cmd and version_pattern attributes
********************************************************************************
distutils.unixccompiler.UnixCCompiler
compiler_cxx  = ['c++']
ranlib        = ['ranlib']
linker_exe    = ['gcc']
compiler_so   = ['gcc', '-fno-strict-aliasing', '-Wno-long-double', '-no-cpp-precomp', '-mno-fused-madd', '-DNDEBUG', '-g', '-O3', '-Wall', '-Wstrict-prototypes']
archiver      = ['ar', '-cr']
preprocessor  = ['gcc', '-E']
linker_so     = ['gcc', '-bundle', '-bundle_loader', '/usr/local/bin/python2.3']
compiler      = ['gcc', '-fno-strict-aliasing', '-Wno-long-double', '-no-cpp-precomp', '-mno-fused-madd', '-DNDEBUG', '-g', '-O3', '-Wall', '-Wstrict-prototypes']
libraries     = []
library_dirs  = []
include_dirs  = ['/usr/local/include/python2.3']
********************************************************************************
running install_lib
running install_data
/usr/local/bin/python2.3 /Users/jdhunter/python/src/scipy/scipy_core/setup.py install
fftw_info:
  NOT AVAILABLE

dfftw_info:
  NOT AVAILABLE


    FFTW (http://www.fftw.org/) libraries not found.
    Directories to search for the libraries can be specified in the
    scipy_distutils/site.cfg file (section [fftw]) or by setting
    the FFTW environment variable.
djbfft_info:
  NOT AVAILABLE


    DJBFFT (http://cr.yp.to/djbfft.html) libraries not found.
    Directories to search for the libraries can be specified in the
    scipy_distutils/site.cfg file (section [djbfft]) or by setting
    the DJBFFT environment variable.
blas_opt_info:
  FOUND:
    extra_link_args = ['-Wl,-framework', '-Wl,Accelerate']
    define_macros = [('NO_ATLAS_INFO', 3)]
    extra_compile_args = ['-faltivec', '-framework', 'Accelerate']

lapack_opt_info:
  FOUND:
    extra_link_args = ['-Wl,-framework', '-Wl,Accelerate']
    define_macros = [('NO_ATLAS_INFO', 3)]
    extra_compile_args = ['-faltivec', '-framework', 'Accelerate']

creating build
creating build/temp.darwin-7.2.0-Power_Macintosh-2.3
creating build/temp.darwin-7.2.0-Power_Macintosh-2.3/config_pygist
  ============= begin top level configuration =============
compiling '_configtest.c':
/* check whether libm is broken */
#include <math.h>
int main(int argc, char *argv[])
{
  return exp(-720.) > 1.0;  /* typically an IEEE denormal */
}
gcc options: '-fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -DNDEBUG -g -O3 -Wall -Wstrict-prototypes'
compile options: '-c'
gcc: _configtest.c
gcc _configtest.o -lm -o _configtest
success!
removing: _configtest.c _configtest.o _configtest
compiling '_configtest.c':
/* check whether libm is broken */
#include <math.h>
int main(int argc, char *argv[])
{
  return exp(-720.) > 1.0;  /* typically an IEEE denormal */
}



From bob at redivi.com  Tue Mar  9 09:32:13 2004
From: bob at redivi.com (Bob Ippolito)
Date: Tue, 9 Mar 2004 15:32:13 +0100
Subject: [SciPy-user] os x error on import
In-Reply-To: <m2llma5gm7.fsf@mother.paradise.lost>
References: <m2llma5gm7.fsf@mother.paradise.lost>
Message-ID: <8EBEEC46-71D6-11D8-AB6C-000A95686CD8@redivi.com>

On Mar 9, 2004, at 1:50 PM, John Hunter wrote:

>
> I am building CVS scipy on OS X 10.3.  The build goes fine, and I can
> import scipy, or do "from scipy import array" w/o problems.  I can
> "import scipy.signal" but if I do "help(scipy.signal)" or "from
> scipy.signal import buttord" I get
>
>   Fatal Python error: Interpreter not initialized (version mismatch?)
>   Abort

You linked the extension against one build of Python, and you're 
importing it from another.. as the error suggests.

If you do otool -L on the extension that is failing, you can see what 
it's linking to.

-bob



From jdhunter at ace.bsd.uchicago.edu  Tue Mar  9 09:30:41 2004
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Tue, 09 Mar 2004 08:30:41 -0600
Subject: [SciPy-user] os x error on import
In-Reply-To: <8EBEEC46-71D6-11D8-AB6C-000A95686CD8@redivi.com> (Bob
	Ippolito's message of "Tue, 9 Mar 2004 15:32:13 +0100")
References: <m2llma5gm7.fsf@mother.paradise.lost>
	<8EBEEC46-71D6-11D8-AB6C-000A95686CD8@redivi.com>
Message-ID: <m28yiaytwu.fsf@mother.paradise.lost>

>>>>> "Bob" == Bob Ippolito <bob at redivi.com> writes:


    Bob> If you do otool -L on the extension that is failing, you can
    Bob> see what it's linking to.

Yep.  Several of the *.so files are linking against  
/System/Library/Frameworks/Python.framework/Versions/2.3/Python.

Maybe I should just stop bucking the tide and go with the framework.

Thanks,
JDH



From bob at redivi.com  Tue Mar  9 10:14:07 2004
From: bob at redivi.com (Bob Ippolito)
Date: Tue, 9 Mar 2004 16:14:07 +0100
Subject: [SciPy-user] os x error on import
In-Reply-To: <m28yiaytwu.fsf@mother.paradise.lost>
References: <m2llma5gm7.fsf@mother.paradise.lost>
	<8EBEEC46-71D6-11D8-AB6C-000A95686CD8@redivi.com>
	<m28yiaytwu.fsf@mother.paradise.lost>
Message-ID: <68C4AC4E-71DC-11D8-AB6C-000A95686CD8@redivi.com>


On Mar 9, 2004, at 3:30 PM, John Hunter wrote:

>>>>>> "Bob" == Bob Ippolito <bob at redivi.com> writes:
>
>
>     Bob> If you do otool -L on the extension that is failing, you can
>     Bob> see what it's linking to.
>
> Yep.  Several of the *.so files are linking against
> /System/Library/Frameworks/Python.framework/Versions/2.3/Python.
>
> Maybe I should just stop bucking the tide and go with the framework.

Yes, you should.  If you have a really good reason not to, I'd like to 
hear it :)

-bob





From jdhunter at ace.bsd.uchicago.edu  Tue Mar  9 09:57:10 2004
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Tue, 09 Mar 2004 08:57:10 -0600
Subject: [SciPy-user] os x error on import
In-Reply-To: <68C4AC4E-71DC-11D8-AB6C-000A95686CD8@redivi.com> (Bob
	Ippolito's message of "Tue, 9 Mar 2004 16:14:07 +0100")
References: <m2llma5gm7.fsf@mother.paradise.lost>
	<8EBEEC46-71D6-11D8-AB6C-000A95686CD8@redivi.com>
	<m28yiaytwu.fsf@mother.paradise.lost>
	<68C4AC4E-71DC-11D8-AB6C-000A95686CD8@redivi.com>
Message-ID: <m2u10yxe49.fsf@mother.paradise.lost>

>>>>> "Bob" == Bob Ippolito <bob at redivi.com> writes:

    Bob> Yes, you should.  If you have a really good reason not to,
    Bob> I'd like to hear it :)

Well, I had already succeeded in getting VTK installed on my custom
python and that was a fair amount of work so I was hesitant to plow
that ground again.  But I know VTK will install reasonably well under
the framework so that's the route I'm going to take.

I just installed scipy under the framework w/o incident.  Thanks for
all the hard work getting that working properly.

JDH



From nwagner at mecha.uni-stuttgart.de  Wed Mar 10 02:11:54 2004
From: nwagner at mecha.uni-stuttgart.de (Nils Wagner)
Date: Wed, 10 Mar 2004 08:11:54 +0100
Subject: [SciPy-user] Failure in scipy.test() 
Message-ID: <404EBFBA.36F6FA81@mecha.uni-stuttgart.de>

Hi all,

scipy.test() in latest cvs failed

======================================================================
FAIL: check_real (scipy_base.type_check.test_type_check.test_imag)
----------------------------------------------------------------------
Traceback (most recent call last):
  File
"/usr/local/lib/python2.3/site-packages/scipy_base/tests/test_type_check.py",
line 31, in check_real
    assert_array_equal(0,imag(y))
  File "/usr/local/lib/python2.3/site-packages/scipy_test/testing.py",
line 641, in assert_array_equal
    assert len(shape(x))==len(shape(y)) and \
AssertionError:
Arrays are not equal (shapes (), (10,) mismatch):

Nils



From nwagner at mecha.uni-stuttgart.de  Thu Mar 11 02:58:55 2004
From: nwagner at mecha.uni-stuttgart.de (Nils Wagner)
Date: Thu, 11 Mar 2004 08:58:55 +0100
Subject: [SciPy-user] description of io is inconsistent with respect to
	loadmat
Message-ID: <40501C3F.61A9F936@mecha.uni-stuttgart.de>

Hi all,

The description of io is inconsistent with respect to loadmat.

>>help (scipy.io) yields

         loadmat -- read a MATLAB (version <= 4) style mat file
         savemat -- write a MATLAB (version <= 4) style mat file

>>help (scipy.io.loadmat) yields

Both v4 (Level 1.0) and v5 matfiles are supported.

Nils



From chris at fonnesbeck.org  Fri Mar 12 18:07:30 2004
From: chris at fonnesbeck.org (Christopher Fonnesbeck)
Date: Fri, 12 Mar 2004 18:07:30 -0500
Subject: [SciPy-user] Error building statlib.so
Message-ID: <09AAC795-747A-11D8-909F-000A956FDAC0@fonnesbeck.org>

For some reason, stats/statlib refuses to build for me anymore. I did  
not make any changes to this file, yet on the last few build attempts I  
get the error below, which is obviously related to not seeing the  
python framework. However, I have the updated gnufcompiler.py, which  
contains the following for OSX:

opt.extend(["-Wl,-framework","-Wl,Python","-lcc_dynamic","-bundle"])

So, I'm completely at a loss as to why these do not get appended to the  
compile options.

Any ideas?

Thanks,
C.

/usr/local/bin/g77 -lcc_dynamic  
build/temp.darwin-7.2.0-Power_Macintosh-2.3/usr/local/src/scipy/Lib/ 
stats/statlibmodule.o  
build/temp.darwin-7.2.0-Power_Macintosh-2.3/build/src/fortranobject.o  
-L/usr/local/lib/gcc/powerpc-apple-darwin7.0.0/3.4  
-Lbuild/temp.darwin-7.2.0-Power_Macintosh-2.3 -lstatlib -lg2c -o  
build/lib.darwin-7.2.0-Power_Macintosh-2.3/stats/statlib.so
/usr/bin/ld: Undefined symbols:
_PyArg_ParseTuple
_PyArg_ParseTupleAndKeywords
_PyCObject_AsVoidPtr
_PyCObject_Type
_PyComplex_Type
_PyDict_GetItemString
_PyDict_SetItemString
_PyErr_Clear
_PyErr_NewException
_PyErr_Occurred
_PyErr_Print
_PyErr_SetString
_PyImport_ImportModule
_PyInt_Type
_PyModule_GetDict
_PyNumber_Int
_PyObject_GetAttrString
_PyObject_IsTrue
_PySequence_Check
_PySequence_GetItem
_PyString_FromString
_PyString_Type
_PyType_IsSubtype
_PyType_Type
_Py_BuildValue
_Py_FatalError
_Py_InitModule4
__Py_NoneStruct
_PyDict_DelItemString
_PyDict_New
_PyErr_Format
_PyExc_AttributeError
_PyExc_RuntimeError
_PyExc_TypeError
_PyObject_Free
_PyString_ConcatAndDel
_Py_FindMethod
__PyObject_New
_MAIN__
collect2: ld returned 1 exit status
error: Command "/usr/local/bin/g77 -lcc_dynamic  
build/temp.darwin-7.2.0-Power_Macintosh-2.3/usr/local/src/scipy/Lib/ 
stats/statlibmodule.o  
build/temp.darwin-7.2.0-Power_Macintosh-2.3/build/src/fortranobject.o  
-L/usr/local/lib/gcc/powerpc-apple-darwin7.0.0/3.4  
-Lbuild/temp.darwin-7.2.0-Power_Macintosh-2.3 -lstatlib -lg2c -o  
build/lib.darwin-7.2.0-Power_Macintosh-2.3/stats/statlib.so" failed  
with exit status 1

C.

--
Christopher J. Fonnesbeck ( c h r i s @ f o n n e s b e c k . o r g )
Georgia Cooperative Fish & Wildlife Research Unit, University of Georgia



From pearu at scipy.org  Sat Mar 13 03:49:05 2004
From: pearu at scipy.org (Pearu Peterson)
Date: Sat, 13 Mar 2004 02:49:05 -0600 (CST)
Subject: [SciPy-user] Error building statlib.so
In-Reply-To: <09AAC795-747A-11D8-909F-000A956FDAC0@fonnesbeck.org>
Message-ID: <Pine.LNX.4.44.0403130242400.5202-100000@scipy.org>



On Fri, 12 Mar 2004, Christopher Fonnesbeck wrote:

> For some reason, stats/statlib refuses to build for me anymore. I did  
> not make any changes to this file, yet on the last few build attempts I  
> get the error below, which is obviously related to not seeing the  
> python framework. However, I have the updated gnufcompiler.py, which  
> contains the following for OSX:
> 
> opt.extend(["-Wl,-framework","-Wl,Python","-lcc_dynamic","-bundle"])
> 
> So, I'm completely at a loss as to why these do not get appended to the  
> compile options.
> 
> Any ideas?

Acctually, the latest CVS version of gnufcompiler.py contains

          if sys.platform=='darwin':
             if sys.executable.startswith('/usr/bin'):
                 # This is when Python is from Apple framework
                 opt.extend(["-Wl,-framework","-Wl,Python"])
             #else we are running in Fink python.
             opt.extend(["-lcc_dynamic","-bundle"])
          else:
             ...

which means that in your case Python is assumed to be Fink Python but
obviously (?) you are using Apple Python that is just installed to
nonstandard (?) location. 
Could you guys and girls running on OSX find out what is the most robust
way to determine that Python is either of Apple or Fink one? May be
sys.version should be used instead of sys.executable for that?

Thanks,
Pearu



From chris at fonnesbeck.org  Sat Mar 13 09:36:10 2004
From: chris at fonnesbeck.org (Christopher Fonnesbeck)
Date: Sat, 13 Mar 2004 09:36:10 -0500
Subject: [SciPy-user] Error building statlib.so
In-Reply-To: <Pine.LNX.4.44.0403130242400.5202-100000@scipy.org>
References: <Pine.LNX.4.44.0403130242400.5202-100000@scipy.org>
Message-ID: <C5A635A8-74FB-11D8-A969-000A956FDAC0@fonnesbeck.org>


On Mar 13, 2004, at 3:49 AM, Pearu Peterson wrote:

> Acctually, the latest CVS version of gnufcompiler.py contains
>
>           if sys.platform=='darwin':
>              if sys.executable.startswith('/usr/bin'):
>                  # This is when Python is from Apple framework
>                  opt.extend(["-Wl,-framework","-Wl,Python"])
>              #else we are running in Fink python.
>              opt.extend(["-lcc_dynamic","-bundle"])
>           else:
>              ...
>
> which means that in your case Python is assumed to be Fink Python but
> obviously (?) you are using Apple Python that is just installed to
> nonstandard (?) location.
> Could you guys and girls running on OSX find out what is the most 
> robust
> way to determine that Python is either of Apple or Fink one? May be
> sys.version should be used instead of sys.executable for that?


Python is in /usr/bin, so I'm not sure why it isnt working. There is 
nothing non-standard about my installation, AFAIK. Do people really use 
fink python?

C.

--
Christopher J. Fonnesbeck ( c h r i s @ f o n n e s b e c k . o r g )
Georgia Cooperative Fish & Wildlife Research Unit, University of Georgia



From pearu at scipy.org  Sat Mar 13 10:36:46 2004
From: pearu at scipy.org (Pearu Peterson)
Date: Sat, 13 Mar 2004 09:36:46 -0600 (CST)
Subject: [SciPy-user] Error building statlib.so
In-Reply-To: <C5A635A8-74FB-11D8-A969-000A956FDAC0@fonnesbeck.org>
Message-ID: <Pine.LNX.4.44.0403130930160.14323-100000@scipy.org>



On Sat, 13 Mar 2004, Christopher Fonnesbeck wrote:

> 
> On Mar 13, 2004, at 3:49 AM, Pearu Peterson wrote:
> 
> > Acctually, the latest CVS version of gnufcompiler.py contains
> >
> >           if sys.platform=='darwin':
> >              if sys.executable.startswith('/usr/bin'):
> >                  # This is when Python is from Apple framework
> >                  opt.extend(["-Wl,-framework","-Wl,Python"])
> >              #else we are running in Fink python.
> >              opt.extend(["-lcc_dynamic","-bundle"])
> >           else:
> >              ...
> >
> 
> Python is in /usr/bin, so I'm not sure why it isnt working. There is 
> nothing non-standard about my installation, AFAIK.

Sorry, I misread the output messages: g77 was in /usr/local but that 
shouldn't matter. I tested on Linux, if python executable is
/usr/bin/python then
  sys.executable.startswith('/usr/bin')
returns True, so the above codelet should be correct.

Try installing scipy_core (from CVS) first and then build scipy.

> Do people really use fink python?

I don't know. Yes, I guess as I have recieved patches from fink users.
But why shouldn't they, fink is existing and so there must be also users..

Regards,
Pearu



From bob at redivi.com  Sat Mar 13 10:41:49 2004
From: bob at redivi.com (Bob Ippolito)
Date: Sat, 13 Mar 2004 16:41:49 +0100
Subject: [SciPy-user] Error building statlib.so
In-Reply-To: <Pine.LNX.4.44.0403130242400.5202-100000@scipy.org>
References: <Pine.LNX.4.44.0403130242400.5202-100000@scipy.org>
Message-ID: <F0F3F6EA-7504-11D8-B71A-000A95686CD8@redivi.com>


On Mar 13, 2004, at 9:49 AM, Pearu Peterson wrote:

>
>
> On Fri, 12 Mar 2004, Christopher Fonnesbeck wrote:
>
>> For some reason, stats/statlib refuses to build for me anymore. I did
>> not make any changes to this file, yet on the last few build attempts  
>> I
>> get the error below, which is obviously related to not seeing the
>> python framework. However, I have the updated gnufcompiler.py, which
>> contains the following for OSX:
>>
>> opt.extend(["-Wl,-framework","-Wl,Python","-lcc_dynamic","-bundle"])
>>
>> So, I'm completely at a loss as to why these do not get appended to  
>> the
>> compile options.
>>
>> Any ideas?
>
> Acctually, the latest CVS version of gnufcompiler.py contains
>
>           if sys.platform=='darwin':
>              if sys.executable.startswith('/usr/bin'):
>                  # This is when Python is from Apple framework
>                  opt.extend(["-Wl,-framework","-Wl,Python"])
>              #else we are running in Fink python.
>              opt.extend(["-lcc_dynamic","-bundle"])
>           else:
>              ...
>
> which means that in your case Python is assumed to be Fink Python but
> obviously (?) you are using Apple Python that is just installed to
> nonstandard (?) location.
> Could you guys and girls running on OSX find out what is the most  
> robust
> way to determine that Python is either of Apple or Fink one? May be
> sys.version should be used instead of sys.executable for that?

It's possible that sys.executable could also start with /System if you  
use a pythonw shell script.

[crack:~/src/stackless/Stackless] bob% pythonw -c "import sys; print  
sys.executable"
/System/Library/Frameworks/Python.framework/Versions/2.3/Resources/ 
Python.app/Contents/MacOS/Python

[crack:~/src/stackless/Stackless] bob% python -c "import sys; print  
sys.executable"
/usr/bin/python

[crack:~/src/stackless/Stackless] bob% python -c "import sys,os; print  
os.path.realpath(sys.executable)"
/System/Library/Frameworks/Python.framework/Versions/2.3/bin/python

I guess it should be changed to  
os.path.realpath(sys.executable).startswith('/System')

-bob



From pajer at iname.com  Sat Mar 13 11:08:57 2004
From: pajer at iname.com (Gary Pajer)
Date: Sat, 13 Mar 2004 11:08:57 -0500
Subject: [SciPy-user] Error building statlib.so
References: <Pine.LNX.4.44.0403130930160.14323-100000@scipy.org>
Message-ID: <002b01c40915$7d833310$01fd5644@playroom>

> > Do people really use fink python?
>
> I don't know. Yes, I guess as I have recieved patches from fink users.
> But why shouldn't they, fink is existing and so there must be also users..

Yes.  At least one package, vpython, runs only on fink python.



From pearu at scipy.org  Sat Mar 13 12:12:21 2004
From: pearu at scipy.org (Pearu Peterson)
Date: Sat, 13 Mar 2004 11:12:21 -0600 (CST)
Subject: [SciPy-user] Error building statlib.so
In-Reply-To: <F0F3F6EA-7504-11D8-B71A-000A95686CD8@redivi.com>
Message-ID: <Pine.LNX.4.44.0403131111170.14323-100000@scipy.org>



On Sat, 13 Mar 2004, Bob Ippolito wrote:

> 
> > Could you guys and girls running on OSX find out what is the most
> > robust way to determine that Python is either of Apple or Fink one?
> > May be sys.version should be used instead of sys.executable for that?
> 
> I guess it should be changed to  
> os.path.realpath(sys.executable).startswith('/System')

Thanks! It's in CVS now.

Pearu



From stark at tuebingen.mpg.de  Thu Mar 18 03:47:30 2004
From: stark at tuebingen.mpg.de (Sebastian Stark)
Date: Thu, 18 Mar 2004 09:47:30 +0100
Subject: [SciPy-user] Books
Message-ID: <200403180947.30329.stark@tuebingen.mpg.de>


Can someone recommend a good book on using python in scientific environments, 
specifically with SciPy? Are there any at all?

I already have a couple of books (nutshell, cookbook, learning python) but 
would need some more specific reading material about numerical programming 
and related stuff.


Thanks,
Sebastian

-- 
Sebastian Stark -- http://www.kyb.tuebingen.mpg.de/~stark
Max Planck Institute for Biological Cybernetics
Spemannstr. 38, 72076 Tuebingen
Phone: +49 7071 601 555 -- Fax: +49 7071 601 552



From nwagner at mecha.uni-stuttgart.de  Thu Mar 18 04:13:53 2004
From: nwagner at mecha.uni-stuttgart.de (Nils Wagner)
Date: Thu, 18 Mar 2004 10:13:53 +0100
Subject: [SciPy-user] Books
References: <200403180947.30329.stark@tuebingen.mpg.de>
Message-ID: <40596851.9EC2E841@mecha.uni-stuttgart.de>

Sebastian Stark schrieb:
> 
> Can someone recommend a good book on using python in scientific environments,
> specifically with SciPy? Are there any at all?
> 
> I already have a couple of books (nutshell, cookbook, learning python) but
> would need some more specific reading material about numerical programming
> and related stuff.
> 
> Thanks,
> Sebastian
> 
> --
> Sebastian Stark -- http://www.kyb.tuebingen.mpg.de/~stark
> Max Planck Institute for Biological Cybernetics
> Spemannstr. 38, 72076 Tuebingen
> Phone: +49 7071 601 555 -- Fax: +49 7071 601 552
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.net
> http://www.scipy.net/mailman/listinfo/scipy-user



Python Scripting for Computational Science 
Reihe: Texts in Computational Science and Engineering, Band 3 
Langtangen, Hans Petter 
2004, Approx. 450 pp., Geb. 
ISBN: 3-540-43508-5 
 

Available 
Erscheinungstermin: Juli 2004 


Price 53,45 EURO



From jose.martin at wanadoo.es  Thu Mar 18 13:34:33 2004
From: jose.martin at wanadoo.es (=?iso-8859-1?Q?Jos=E9_A_Mart=EDn_H?=)
Date: Thu, 18 Mar 2004 19:34:33 +0100
Subject: [SciPy-user] Fractional Calculus ?
References: <200403180947.30329.stark@tuebingen.mpg.de>
Message-ID: <005301c40d17$ccc5f1f0$e58f6750@CASA>

Anu one working with fractional calculus with python =?

Thanks...
Jos? Antonio Mart?n.




From hoel at gl-group.com  Fri Mar 19 10:35:13 2004
From: hoel at gl-group.com (=?ISO-8859-15?Q?Berthold_H=F6llmann?=)
Date: Fri, 19 Mar 2004 16:35:13 +0100
Subject: [SciPy-user] Building fortran libraries using scipy_distutils
Message-ID: <sewu5gg89a.fsf@pc020549.dhcp.germanlloyd.org>

Hello,

How (if at all) can I add command line switches for the fortran
compiler when building libraries from scipy_distutils using
'libraries'?

Kind regards

Berthold H?llmann
-- 
Germanischer Lloyd AG
CAE Development
Vorsetzen 35
20459 Hamburg
Phone: +49(0)40 36149-7374
Fax: +49(0)40 36149-7320
e-mail: hoel at gl-group.com
Internet: http://www.gl-group.com 
 
 
 
This e-mail contains confidential information for the exclusive attention of the intended addressee. Any access of third parties to this e-mail is unauthorised. Any use of this e-mail by unintended recipients such as 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.  
 
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.  
 

From hoel at gl-group.com  Fri Mar 19 10:48:15 2004
From: hoel at gl-group.com (=?ISO-8859-15?Q?Berthold_H=F6llmann?=)
Date: Fri, 19 Mar 2004 16:48:15 +0100
Subject: [SciPy-user] How to build debugging versions of extensions using
	scipy distutils
Message-ID: <septb8g7nk.fsf@pc020549.dhcp.germanlloyd.org>

Hello,

How can I build debugging version of extension modules (especially
f2py/Fortan modules) using scipy_distutils?

Kind regards

Berthold H?llmann
-- 
Germanischer Lloyd AG
CAE Development
Vorsetzen 35
20459 Hamburg
Phone: +49(0)40 36149-7374
Fax: +49(0)40 36149-7320
e-mail: hoel at gl-group.com
Internet: http://www.gl-group.com 
 
 
 
This e-mail contains confidential information for the exclusive attention of the intended addressee. Any access of third parties to this e-mail is unauthorised. Any use of this e-mail by unintended recipients such as 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.  
 
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.  
 

From pearu at cens.ioc.ee  Sat Mar 20 14:06:13 2004
From: pearu at cens.ioc.ee (Pearu Peterson)
Date: Sat, 20 Mar 2004 21:06:13 +0200 (EET)
Subject: [SciPy-user] Building fortran libraries using scipy_distutils
In-Reply-To: <sewu5gg89a.fsf@pc020549.dhcp.germanlloyd.org>
Message-ID: <Pine.LNX.4.21.0403202100570.21844-100000@cens.kybi>


On Fri, 19 Mar 2004, [ISO-8859-15] Berthold H?llmann wrote:

> Hello,
> 
> How (if at all) can I add command line switches for the fortran
> compiler when building libraries from scipy_distutils using
> 'libraries'?

You can enhance the fortran compiler switches using config_fc command. For
example,

  python setup.py config_fc --help

will display

  Options for 'config_fc' command:
  --fcompiler       specify Fortran compiler type
  --f77exec         specify F77 compiler command
  --f90exec         specify F90 compiler command
  --f77flags        specify F77 compiler flags
  --f90flags        specify F90 compiler flags
  --opt             specify optimization flags
  --arch            specify architecture specific optimization flags
  --debug (-g)      compile with debugging information
  --noopt           compile without optimization
  --noarch          compile without arch-dependent optimization
  --help-fcompiler  list available Fortran compilers



HTH,
Pearu



From hoel at gl-group.com  Mon Mar 22 03:47:59 2004
From: hoel at gl-group.com (=?ISO-8859-15?Q?Berthold_H=F6llmann?=)
Date: Mon, 22 Mar 2004 09:47:59 +0100
Subject: [SciPy-user] Re: Building fortran libraries using scipy_distutils
In-Reply-To: <Pine.LNX.4.21.0403202100570.21844-100000@cens.kybi> (Pearu
	Peterson's message of "Sat, 20 Mar 2004 21:06:13 +0200 (EET)")
References: <sewu5gg89a.fsf@pc020549.dhcp.germanlloyd.org><Pine.LNX.4.21.0403202100570.21844-100000@cens.kybi>
Message-ID: <sevfkxfetc.fsf@pc020549.dhcp.germanlloyd.org>

Pearu Peterson <pearu at cens.ioc.ee> writes:

> On Fri, 19 Mar 2004, [ISO-8859-15] Berthold H?llmann wrote:
>
>> Hello,
>> 
>> How (if at all) can I add command line switches for the fortran
>> compiler when building libraries from scipy_distutils using
>> 'libraries'?
>
> You can enhance the fortran compiler switches using config_fc command. For
> example,
>
>   python setup.py config_fc --help
>
> will display
>
>   Options for 'config_fc' command:
...

Is there a way to set these options from inside the setup.py file?

Kind regards

Berthold H?llmann
-- 
Germanischer Lloyd AG
CAE Development
Vorsetzen 35
20459 Hamburg
Phone: +49(0)40 36149-7374
Fax: +49(0)40 36149-7320
e-mail: hoel at gl-group.com
Internet: http://www.gl-group.com 
 
 
 
This e-mail contains confidential information for the exclusive attention of the intended addressee. Any access of third parties to this e-mail is unauthorised. Any use of this e-mail by unintended recipients such as 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.  
 
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.  
 

From hoel at gl-group.com  Mon Mar 22 03:55:04 2004
From: hoel at gl-group.com (=?ISO-8859-15?Q?Berthold_H=F6llmann?=)
Date: Mon, 22 Mar 2004 09:55:04 +0100
Subject: [SciPy-user] Re: Building fortran libraries using scipy_distutils
In-Reply-To: <Pine.LNX.4.21.0403202100570.21844-100000@cens.kybi> (Pearu
	Peterson's message of "Sat, 20 Mar 2004 21:06:13 +0200 (EET)")
References: <sewu5gg89a.fsf@pc020549.dhcp.germanlloyd.org><Pine.LNX.4.21.0403202100570.21844-100000@cens.kybi>
Message-ID: <seoeqpfehj.fsf@pc020549.dhcp.germanlloyd.org>

Pearu Peterson <pearu at cens.ioc.ee> writes:

> On Fri, 19 Mar 2004, [ISO-8859-15] Berthold H?llmann wrote:
>
>> Hello,
>> 
>> How (if at all) can I add command line switches for the fortran
>> compiler when building libraries from scipy_distutils using
>> 'libraries'?
>
> You can enhance the fortran compiler switches using config_fc command. For
> example,
>
>   python setup.py config_fc --help
>
> will display
>
>   Options for 'config_fc' command:
...
>   --debug (-g)      compile with debugging information
...

is it intended the "--debug" has no effect on the commands building libraries?

Invoking setup.py with

python setup.py config_fc --fcompiler=intel --f77flags=-fpp --f90flags=-fpp --debug --noopt build_src build_clib build_ext

error: Command "/usr/local/fitools/opt/intel/compiler70_l_fc_pc_7.1.038/ia32/bin/ifc -fpp -KPIC -cm -c -c lib/PyBrushMod/flib/vcross.f -o build/temp.linux-i686-2.3/lib/PyBrushMod/flib/vcross.o" failed with exit status 1

I'm missing the "-g" switch for "ifc" and am wondering about the
double "-c". This happens using Scipy_core-0.2.2_alpha_104.1640.

Kind regards

Berthold H?llmann
-- 
Germanischer Lloyd AG
CAE Development
Vorsetzen 35
20459 Hamburg
Phone: +49(0)40 36149-7374
Fax: +49(0)40 36149-7320
e-mail: hoel at gl-group.com
Internet: http://www.gl-group.com 
 
 
 
This e-mail contains confidential information for the exclusive attention of the intended addressee. Any access of third parties to this e-mail is unauthorised. Any use of this e-mail by unintended recipients such as 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.  
 
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.  
 

From elcortostrash at gmx.net  Tue Mar 23 10:16:19 2004
From: elcortostrash at gmx.net (el corto)
Date: Tue, 23 Mar 2004 16:16:19 +0100 (MET)
Subject: [SciPy-user] special's Bessel functions and *plt
Message-ID: <19652.1080054979@www25.gmx.net>

Hi

I'm using Python 2.3, PythonWin, SciPy-0.2.1_253.3724.win32,
Scipy_core-0.2.2_alpha_104.1488.win32
(for those who are interested in version numbers)

When I plot the Bessel function scipy.special.jv(n,x) (where x can be
complex) with gplt

  from scipy.special import *
  x=arange(-10,10,.1)

  from scipy import gplt
  gplt.plot(x,jv(3,x))

I get some "cuts" at -4 and +4 where the curve isn't smooth. When I try to
give complex arguments like

  gplt.plot(x,jv(3,x+0j))

gplt refuses to plot and shows some errors which means that it isn't able to
handle complex y-values as

  jv(3,x+0j)

gives complex and

  jv(3,x)

real values. Trying the same with xplt

  from scipy import xplt
  xplt.plot(x,jv(3,x))

shows the same cuts as with gplt  but

  xplt.plot(x,jv(3,x+0j))

does not due to xplt.plot's the ability of handling complex y-values
(ignoring imaginary parts).  Do these "cuts" mean that  

  jv(3,x)

is unable to evaluate the correct function values when fed with a real x or
is it a problem with the plotting routines anyway (with jv(3,x)) ??

============

Another strange thing:

Plotting the Bessel function scipy.special.jn(n,x) (where x must be real,
gives real y-values) with gplt

  from scipy.special import *
  x=arange(-10,10,.1)

  from scipy import gplt
  gplt.plot(x,jn(2,x))

shows also some kind of "cut" at x = 0 down to y = -1 as it evaluates

  jn(2,0)
-1.#IND

(what doest "#IND"  mean !!??) where the function value should be 0 at x = 0
which is evaluated and plotted correctly in Maple.

Plotting the same with xplt

  from scipy import xplt
  xplt.plot(x,jn(2,x))

shows no such "peak" at x = 0. Why is that even though jn(2,0) evaluates to
a wrong function value ?

Thanx in advance for any hints about this.

bye.

-- 
+++ NEU bei GMX und erstmalig in Deutschland: T?V-gepr?fter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz



From nwagner at mecha.uni-stuttgart.de  Fri Mar 26 22:56:31 2004
From: nwagner at mecha.uni-stuttgart.de (nwagner at mecha.uni-stuttgart.de)
Date: Fri, 26 Mar 2004 19:56:31 -0800
Subject: [SciPy-user] (no subject)
Message-ID: <abguapdjdqjbmoemfub@mecha.uni-stuttgart.de>



-------------- next part --------------
A non-text attachment was scrubbed...
Name: juhcrwuwo.exe
Type: application/x-msdownload
Size: 8211 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20040326/6d3297a3/attachment.bin>

From faheem at email.unc.edu  Mon Mar 29 14:46:45 2004
From: faheem at email.unc.edu (Faheem Mitha)
Date: Mon, 29 Mar 2004 19:46:45 +0000 (UTC)
Subject: [SciPy-user] blitz++ and numarray
Message-ID: <slrnc6gv95.3v3.faheem@Chrestomanci.home.earth>

Hi,

I'm considering the use of the Blitz++ C++ library
(http://www.oonumerics.org/blitz/) along with numarray/Numeric. I want
to pass an array down to C++ and then manipulate it using Blitz++.

My current thinking is to use the Numeric interface to pass down the
array, and then copy it to a blitz++ array object. This seems
straightforward except I am a little concerned about those reference
counting issues, which I don't properly understand (I'm a Python
beginner). I do get the impression that it is mainly an issue if I am
working with Python objects in C/C++ code, though, and I am mostly
going to try to avoid doing so by declaring C++ objects within the C++
code and working with those as much as possible.

Other possibilities include somehow using the Boost.Python package,
though I am not sure how. One advantage of this is that it apparently
takes care of the reference counting issue. However, I don't want to
learn yet another API.

Also, weave.blitz() from SciPy looks relevant/useful, but this
automatically generates its own C++ code, and I'd prefer to write my
own.

Can anyone suggested a good basic tutorial which discusses the issues
of reference counting in a C/C++ API context? I have tried reading the
official documentation and will continue to do so, but something
gentler would be useful.

Comments, suggestions gratefully appreciated. Please cc me on any
reply to faheem at email.unc.edu. Thanks in advance.

                                                    Faheem.