pysparse-users

Re: [Pysparse-users] Pysparse installation problems

[Rob S. <rs...@MI...>]

Huh. Now I'm confused. It looks like NumPy and SciPy documentation
expect matvec to have the behavior it does in PysparseMatrix, but
pysparse.jdsym expects the behavior it has in ll_mat.

-- Rob

On Thu, Mar 25, 2010 at 12:11 PM, Rob Speer <rs...@mi...> wrote:
> Thanks for figuring out the include thing!
>
> You should put setuptools in a try/except block. Remember, I want
> users to automatically pull in pysparse as a dependency, so they
> wouldn't have an option to stop the process between downloading and
> installing pysparse to edit a file. And when pip installs something
> from SVN instead of from PyPI, it insists on 'setup.py develop'.
>
> It's fortunate that pip can force setuptools to be imported, but we
> shouldn't rely on that behavior.
>
> Unrelatedly, I've just noticed a bug in the PysparseMatrix class. It
> has a .matvec method that doesn't meet NumPy's requirements (because
> it returns its result, instead of writing it into another array).
> Currently I work around it by using .matrix.matvec, which does the
> right thing, but it would be nice if this worked better. Would it make
> sense for me to have SVN access so I can fix minor things like this?
>
> -- Rob
>
> On Thu, Mar 25, 2010 at 10:18 AM, Dominique Orban
> <dom...@gm...> wrote:
>> On Wed, Mar 24, 2010 at 5:24 PM, Rob Speer <rs...@mi...> wrote:
>>>>> 2. Pysparse won't install in a virtualenv. It tries to install files
>>>>> in the include/ directory of your Python environment, but virtualenvs
>>>>> don't have their own include directory. They just symlink to the
>>>>> system-wide one.
>>>>>
>>>>> I don't know what the correct way to resolve this one is, because I've
>>>>> never had to manage additional include files in a Python package.
>>>>
>>>> Rob,
>>>>
>>>> Thanks for the feedback. #1 has been on my list for a long time but
>>>> never gets adressed because time is scarce. I really want to have a
>>>> proper setup.py for Pysparse. I'm now getting better at writing them
>>>> so maybe I can whip up something quick.
>>>>
>>>> Regarding #2, could you be more specific? In my experience,
>>>> package-specific include files always go to the include/ directory of
>>>> the local Python environment. However, you can normally install
>>>> packages "locally" (e.g., in your home directory) instead of system
>>>> wide, by using an option such as --prefix or --home with setup.py. I
>>>> might have misunderstood your question...
>>>
>>> Responding to this part now: does your local Python environment
>>> actually have its own include/ directory? Because my virtualenv just
>>> has a symbolic link to the Python global one. This means that
>>> installing pysparse in a virtualenv requires permission to modify the
>>> global Python include directory, and it does so.
>>
>> No my virtual environment doesn't have its own include dir. All it has
>> is a symlink to the system include dir. This looks like a defect of
>> virtualenv, really.
>>
>> The pysparse header files are now installed under pysparse/include in
>> the lib/pythonX.Y directory. So this doesn't cause permission issues
>> with virtual environments anymore. I also added a convenience function
>> for packages that rely on pysparse:
>>
>>>>> import pysparse
>>>>> pysparse.get_include()
>>
>> Let me know how that goes.
>>
>> --
>> Dominique
>>
>

Re: [Pysparse-users] Computing largest eigenvalues?

[Rob S. <rs...@MI...>]

Got any pointers? Up until now, I've been working with an old, clunky
C library called SVDLIBC that implements Lanczos. Pysparse's jdsym is
the first thing I've seen that can find eigenvectors and presents an
interface that can actually work with Python objects.

There's the stuff in scipy.sparse, of course, but that's been stalled
in the development for years now, and it doesn't often compile from
SVN.

-- Rob

On Thu, Mar 25, 2010 at 3:17 AM, Roman Geus <rom...@gm...> wrote:
> Hi Rob
>
> If you set tau to some very large number, but still far away from the
> actual largest eigenvalue, you might experience very slow convergence.
> There are certainly better and simpler algorithms than JDSYM for
> computing a few of the largest eigenvalues.
>
> -- Roman
>
> On Wed, Mar 24, 2010 at 8:05 PM, Rob Speer <rs...@mi...> wrote:
>> It looks like it should be possible to compute the truncated spectral
>> decomposition of a sparse, symmetric matrix using pysparse.jdsym. This
>> is the key step in computing a truncated SVD, which is the next thing
>> to do, and it would be great to be able to do it entirely within
>> Pysparse.
>>
>> There's just one thing I'm unsure about: how do I ask for the
>> *largest* eigenvalues? jdsym is set up to return eigenvalues around
>> some value tau, defaulting to 0, so it seems this is set up for
>> finding the smallest eigenvalues. Do I just set tau to some very large
>> number, or would that cause numerical stability issues? Is this the
>> wrong problem for jdsym to solve?
>>
>> -- Rob
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> Pysparse-users mailing list
>> Pys...@li...
>> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>>
>

Re: [Pysparse-users] Pysparse installation problems

[Rob S. <rs...@MI...>]

Thanks for figuring out the include thing!

You should put setuptools in a try/except block. Remember, I want
users to automatically pull in pysparse as a dependency, so they
wouldn't have an option to stop the process between downloading and
installing pysparse to edit a file. And when pip installs something
from SVN instead of from PyPI, it insists on 'setup.py develop'.

It's fortunate that pip can force setuptools to be imported, but we
shouldn't rely on that behavior.

Unrelatedly, I've just noticed a bug in the PysparseMatrix class. It
has a .matvec method that doesn't meet NumPy's requirements (because
it returns its result, instead of writing it into another array).
Currently I work around it by using .matrix.matvec, which does the
right thing, but it would be nice if this worked better. Would it make
sense for me to have SVN access so I can fix minor things like this?

-- Rob

On Thu, Mar 25, 2010 at 10:18 AM, Dominique Orban
<dom...@gm...> wrote:
> On Wed, Mar 24, 2010 at 5:24 PM, Rob Speer <rs...@mi...> wrote:
>>>> 2. Pysparse won't install in a virtualenv. It tries to install files
>>>> in the include/ directory of your Python environment, but virtualenvs
>>>> don't have their own include directory. They just symlink to the
>>>> system-wide one.
>>>>
>>>> I don't know what the correct way to resolve this one is, because I've
>>>> never had to manage additional include files in a Python package.
>>>
>>> Rob,
>>>
>>> Thanks for the feedback. #1 has been on my list for a long time but
>>> never gets adressed because time is scarce. I really want to have a
>>> proper setup.py for Pysparse. I'm now getting better at writing them
>>> so maybe I can whip up something quick.
>>>
>>> Regarding #2, could you be more specific? In my experience,
>>> package-specific include files always go to the include/ directory of
>>> the local Python environment. However, you can normally install
>>> packages "locally" (e.g., in your home directory) instead of system
>>> wide, by using an option such as --prefix or --home with setup.py. I
>>> might have misunderstood your question...
>>
>> Responding to this part now: does your local Python environment
>> actually have its own include/ directory? Because my virtualenv just
>> has a symbolic link to the Python global one. This means that
>> installing pysparse in a virtualenv requires permission to modify the
>> global Python include directory, and it does so.
>
> No my virtual environment doesn't have its own include dir. All it has
> is a symlink to the system include dir. This looks like a defect of
> virtualenv, really.
>
> The pysparse header files are now installed under pysparse/include in
> the lib/pythonX.Y directory. So this doesn't cause permission issues
> with virtual environments anymore. I also added a convenience function
> for packages that rely on pysparse:
>
>>>> import pysparse
>>>> pysparse.get_include()
>
> Let me know how that goes.
>
> --
> Dominique
>

Re: [Pysparse-users] Pysparse installation problems

[Dominique O. <dom...@gm...>]

On Wed, Mar 24, 2010 at 5:24 PM, Rob Speer <rs...@mi...> wrote:
>>> 2. Pysparse won't install in a virtualenv. It tries to install files
>>> in the include/ directory of your Python environment, but virtualenvs
>>> don't have their own include directory. They just symlink to the
>>> system-wide one.
>>>
>>> I don't know what the correct way to resolve this one is, because I've
>>> never had to manage additional include files in a Python package.
>>
>> Rob,
>>
>> Thanks for the feedback. #1 has been on my list for a long time but
>> never gets adressed because time is scarce. I really want to have a
>> proper setup.py for Pysparse. I'm now getting better at writing them
>> so maybe I can whip up something quick.
>>
>> Regarding #2, could you be more specific? In my experience,
>> package-specific include files always go to the include/ directory of
>> the local Python environment. However, you can normally install
>> packages "locally" (e.g., in your home directory) instead of system
>> wide, by using an option such as --prefix or --home with setup.py. I
>> might have misunderstood your question...
>
> Responding to this part now: does your local Python environment
> actually have its own include/ directory? Because my virtualenv just
> has a symbolic link to the Python global one. This means that
> installing pysparse in a virtualenv requires permission to modify the
> global Python include directory, and it does so.

No my virtual environment doesn't have its own include dir. All it has
is a symlink to the system include dir. This looks like a defect of
virtualenv, really.

The pysparse header files are now installed under pysparse/include in
the lib/pythonX.Y directory. So this doesn't cause permission issues
with virtual environments anymore. I also added a convenience function
for packages that rely on pysparse:

>>> import pysparse
>>> pysparse.get_include()

Let me know how that goes.

-- 
Dominique

Re: [Pysparse-users] Computing largest eigenvalues?

[Roman G. <rom...@gm...>]

Hi Rob

If you set tau to some very large number, but still far away from the
actual largest eigenvalue, you might experience very slow convergence.
There are certainly better and simpler algorithms than JDSYM for
computing a few of the largest eigenvalues.

-- Roman

On Wed, Mar 24, 2010 at 8:05 PM, Rob Speer <rs...@mi...> wrote:
> It looks like it should be possible to compute the truncated spectral
> decomposition of a sparse, symmetric matrix using pysparse.jdsym. This
> is the key step in computing a truncated SVD, which is the next thing
> to do, and it would be great to be able to do it entirely within
> Pysparse.
>
> There's just one thing I'm unsure about: how do I ask for the
> *largest* eigenvalues? jdsym is set up to return eigenvalues around
> some value tau, defaulting to 0, so it seems this is set up for
> finding the smallest eigenvalues. Do I just set tau to some very large
> number, or would that cause numerical stability issues? Is this the
> wrong problem for jdsym to solve?
>
> -- Rob
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>

[Pysparse-users] Computing largest eigenvalues?

[Rob S. <rs...@MI...>]

It looks like it should be possible to compute the truncated spectral
decomposition of a sparse, symmetric matrix using pysparse.jdsym. This
is the key step in computing a truncated SVD, which is the next thing
to do, and it would be great to be able to do it entirely within
Pysparse.

There's just one thing I'm unsure about: how do I ask for the
*largest* eigenvalues? jdsym is set up to return eigenvalues around
some value tau, defaulting to 0, so it seems this is set up for
finding the smallest eigenvalues. Do I just set tau to some very large
number, or would that cause numerical stability issues? Is this the
wrong problem for jdsym to solve?

-- Rob

Re: [Pysparse-users] Pysparse installation problems

[Rob S. <rs...@MI...>]

>> 2. Pysparse won't install in a virtualenv. It tries to install files
>> in the include/ directory of your Python environment, but virtualenvs
>> don't have their own include directory. They just symlink to the
>> system-wide one.
>>
>> I don't know what the correct way to resolve this one is, because I've
>> never had to manage additional include files in a Python package.
>
> Rob,
>
> Thanks for the feedback. #1 has been on my list for a long time but
> never gets adressed because time is scarce. I really want to have a
> proper setup.py for Pysparse. I'm now getting better at writing them
> so maybe I can whip up something quick.
>
> Regarding #2, could you be more specific? In my experience,
> package-specific include files always go to the include/ directory of
> the local Python environment. However, you can normally install
> packages "locally" (e.g., in your home directory) instead of system
> wide, by using an option such as --prefix or --home with setup.py. I
> might have misunderstood your question...

Responding to this part now: does your local Python environment
actually have its own include/ directory? Because my virtualenv just
has a symbolic link to the Python global one. This means that
installing pysparse in a virtualenv requires permission to modify the
global Python include directory, and it does so.

-- Rob

Re: [Pysparse-users] Pysparse installation problems

[Dominique O. <dom...@gm...>]

On Thu, Mar 11, 2010 at 7:39 PM, Rob Speer <rs...@mi...> wrote:
> I'm having two problems that make it difficult to install Pysparse in
> many situations. I'm writing a library that depends on Pysparse, but I
> can't distribute this library if Pysparse itself won't install.
>
>
> 1. Pysparse doesn't support "setup.py develop". This means it cannot
> be installed through Pip.
>
> This seems to be because of the line:
>   package_dir = {package_name: 'Lib'},
>
> The problem is that there is no valid directory to put on the Python
> path. Here's a bug report about this issue (which was closed as
> invalid in Pip because they concluded it's a problem with either
> setuptools or pysparse):
>
> http://bitbucket.org/ianb/pip/issue/72/editable-install-doesnt-respect-package_dir
>
> It sounds like this could be fixed by reorganizing the code structure,
> or at least making a symlink, so that there is an actual "pysparse"
> directory at the top level. Renaming "Lib" to "pysparse" might be the
> right answer.
>
>
> 2. Pysparse won't install in a virtualenv. It tries to install files
> in the include/ directory of your Python environment, but virtualenvs
> don't have their own include directory. They just symlink to the
> system-wide one.
>
> I don't know what the correct way to resolve this one is, because I've
> never had to manage additional include files in a Python package.

Rob,

Thanks for the feedback. #1 has been on my list for a long time but
never gets adressed because time is scarce. I really want to have a
proper setup.py for Pysparse. I'm now getting better at writing them
so maybe I can whip up something quick.

Regarding #2, could you be more specific? In my experience,
package-specific include files always go to the include/ directory of
the local Python environment. However, you can normally install
packages "locally" (e.g., in your home directory) instead of system
wide, by using an option such as --prefix or --home with setup.py. I
might have misunderstood your question...

Cheers,

-- 
Dominique

[Pysparse-users] Pysparse installation problems

[Rob S. <rs...@MI...>]

I'm having two problems that make it difficult to install Pysparse in
many situations. I'm writing a library that depends on Pysparse, but I
can't distribute this library if Pysparse itself won't install.


1. Pysparse doesn't support "setup.py develop". This means it cannot
be installed through Pip.

This seems to be because of the line:
   package_dir = {package_name: 'Lib'},

The problem is that there is no valid directory to put on the Python
path. Here's a bug report about this issue (which was closed as
invalid in Pip because they concluded it's a problem with either
setuptools or pysparse):

http://bitbucket.org/ianb/pip/issue/72/editable-install-doesnt-respect-package_dir

It sounds like this could be fixed by reorganizing the code structure,
or at least making a symlink, so that there is an actual "pysparse"
directory at the top level. Renaming "Lib" to "pysparse" might be the
right answer.


2. Pysparse won't install in a virtualenv. It tries to install files
in the include/ directory of your Python environment, but virtualenvs
don't have their own include directory. They just symlink to the
system-wide one.

I don't know what the correct way to resolve this one is, because I've
never had to manage additional include files in a Python package.

Thanks,
-- Rob

Re: [Pysparse-users] transpose of matrix?

[Dominique O. <dom...@gm...>]

On Mon, Mar 8, 2010 at 7:12 PM, Bill Bruno <wb...@gm...> wrote:
> I used pysparse successfully; very pleased with performance!
>
> I really want to do SVD but I can emulate that by finding the
> spectrum of A^t A  and  A A^t.
>
> The first is easy using dot, but how do I get the second.
> I would have thought there would be a built-in transpose.
> Should I write my own at low level or convert to another
> format that supports transpose?

Hi Bill,

Thanks for using Pysparse! It is true that there is no formal
transposition operation in Pysparse. That is because often times, in
sparse matrix computations, there are ways to get rid of the
transpose. In your case, note that the eigenvalues of A^T * A and
those of A * A^T are the same (a real matrix and its transpose have
the same eigenvalues). So, there's no need to compute the second one.

That being said, it might be good to have a formal transposition
operation. If you would like to have a go at that, it is surely easier
to start with the Python PysparseMatrix class rather than the
low-level ll_mat objects.

Cheers,

-- 
Dominique

[Pysparse-users] transpose of matrix?

[Bill B. <wb...@gm...>]

I used pysparse successfully; very pleased with performance!

I really want to do SVD but I can emulate that by finding the
spectrum of A^t A  and  A A^t.

The first is easy using dot, but how do I get the second.
I would have thought there would be a built-in transpose.
Should I write my own at low level or convert to another
format that supports transpose?

Bill

Re: [Pysparse-users] can't assign from a dense matrix

[Dominique O. <dom...@gm...>]

On Fri, Feb 26, 2010 at 11:30 PM, Rob Speer <rs...@mi...> wrote:
> It's hard to tell from the Web, but it sounds like the right thing to
> do when I encounter unexpected behavior in PySparse is to e-mail this
> list, right?
>
> I'm trying to assign to a slice of an ll_mat from another matrix. If
> this other matrix is also an ll_mat, this works fine. However, it acts
> very strange if the other matrix is a standard NumPy ndarray:
>
> ----------
> In [1]: from pysparse.spmatrix import *
>
> In [2]: sq = ll_mat(5,5)
>
> In [3]: sq
> Out[3]: <ll_mat object at 0x1d26bf0>
>
> In [4]: print sq
> ------> print(sq)
> ll_mat(general, [5,5]):
>  --------  --------  --------  --------  --------
>  --------  --------  --------  --------  --------
>  --------  --------  --------  --------  --------
>  --------  --------  --------  --------  --------
>  --------  --------  --------  --------  --------
>
> In [6]: import numpy as np
>
> In [7]: sq[:,:] = np.arange(25).reshape((5,5))
> In LHS=RHS, RHS has shape (2,27483600), LHS has shape (5,5)
> ---------------------------------------------------------------------------
> ValueError                                Traceback (most recent call last)
>
> /Users/rspeer/python/src/pysparse/<ipython console> in <module>()
>
> ValueError: Matrix shapes are different
> -----------
>
> There seems to be no other straightforward way to convert a small
> dense matrix to a sparse one, either. Am I missing how to do this, or
> is this a bug that should be fixed?

Rob,

Yes, e-mailing this list is the right thing to do. Welcome to Pysparse!

Indeed, Pysparse expects the right-hand side of an assignment to be a
ll_mat object. Currently, you can't assign from a dense matrix or
easily convert a dense matrix to a ll_mat object. I agree that this
functionality would be useful. I might not have a chance to code that
before a few weeks though. If anyone wants to have a stab at it, I can
get them started.

-- 
Dominique

[Pysparse-users] can't assign from a dense matrix

[Rob S. <rs...@MI...>]

It's hard to tell from the Web, but it sounds like the right thing to
do when I encounter unexpected behavior in PySparse is to e-mail this
list, right?

I'm trying to assign to a slice of an ll_mat from another matrix. If
this other matrix is also an ll_mat, this works fine. However, it acts
very strange if the other matrix is a standard NumPy ndarray:

----------
In [1]: from pysparse.spmatrix import *

In [2]: sq = ll_mat(5,5)

In [3]: sq
Out[3]: <ll_mat object at 0x1d26bf0>

In [4]: print sq
------> print(sq)
ll_mat(general, [5,5]):
 --------  --------  --------  --------  --------
 --------  --------  --------  --------  --------
 --------  --------  --------  --------  --------
 --------  --------  --------  --------  --------
 --------  --------  --------  --------  --------

In [6]: import numpy as np

In [7]: sq[:,:] = np.arange(25).reshape((5,5))
In LHS=RHS, RHS has shape (2,27483600), LHS has shape (5,5)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)

/Users/rspeer/python/src/pysparse/<ipython console> in <module>()

ValueError: Matrix shapes are different
-----------

There seems to be no other straightforward way to convert a small
dense matrix to a sparse one, either. Am I missing how to do this, or
is this a bug that should be fixed?

-- Rob

Re: [Pysparse-users] Hello

[Daniel W. <dan...@gm...>]

Hi Alex, There is no pysparse windows installer for python 2.6 at the
moment. I hope to complete this in the near future and will post to
the list when it is available. However, there is a pysparse windows
installer for python 2.5. Which version of python do you have?

On Tue, Feb 2, 2010 at 1:31 PM, Alex Bourdages <blu...@gm...> wrote:
> I'm trying to get FiPy working for windows through Python(x,y). I know I
> need pysparce to get things running. The only problem I'm running into is
> pysparce does not "see" version 2.5 of python when installing, I believe it
> uses 2.6. I'm very excited to use this software and hope you can give me
> some pointers.
>
> Thanks a bunch,
> ~Alex Bourdages
>
> P.S. I'm not a programmer, I'm a physicist, so please try to speak in lamens
> terms.
>
> ------------------------------------------------------------------------------
> The Planet: dedicated and managed hosting, cloud storage, colocation
> Stay online with enterprise data centers and the best network in the
> business
> Choose flexible plans and management services without long-term contracts
> Personal 24x7 support from experience hosting pros just a phone call away.
> http://p.sf.net/sfu/theplanet-com
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>
>



-- 
Daniel Wheeler

[Pysparse-users] Hello

[Alex B. <blu...@gm...>]

I'm trying to get FiPy working for windows through Python(x,y). I know I
need pysparce to get things running. The only problem I'm running into is
pysparce does not "see" version 2.5 of python when installing, I believe it
uses 2.6. I'm very excited to use this software and hope you can give me
some pointers.

Thanks a bunch,
~Alex Bourdages

P.S. I'm not a programmer, I'm a physicist, so please try to speak in lamens
terms.

Re: [Pysparse-users] windows version compatible with Python 2.6

[Daniel W. <dan...@gm...>]

I'm having some issues installing the latest version pf pythonxy. I'm
awaiting some help on their mailing list, but I'll hopefully have this
done when I get the issues sorted. Cheers

On Sat, Jan 16, 2010 at 8:36 PM, Kevin Allred <all...@ya...> wrote:
> I'm stuck.  I want to install FiPy, but the windows version of python(x,y)
> is python version 2.6, and the only windows installable version of PySparse
> is for python 2.5.  I have no windows compilers.  Is there a solution that
> will get me windows installable PySparse for python2.6?
>
>
> Thanks,
> Kevin Allred
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>
>



-- 
Daniel Wheeler

Re: [Pysparse-users] windows version compatible with Python 2.6

[Daniel W. <dan...@gm...>]

The best solution is for me to build and release it, which I'll try
and do in the near future.

On Sat, Jan 16, 2010 at 8:36 PM, Kevin Allred <all...@ya...> wrote:
> I'm stuck.  I want to install FiPy, but the windows version of python(x,y)
> is python version 2.6, and the only windows installable version of PySparse
> is for python 2.5.  I have no windows compilers.  Is there a solution that
> will get me windows installable PySparse for python2.6?
>
>
> Thanks,
> Kevin Allred
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>
>



-- 
Daniel Wheeler

Re: [Pysparse-users] windows version compatible with Python 2.6

[Dominique O. <dom...@gm...>]

On 16-Jan-10, at 8:36 PM, Kevin Allred wrote:

> I'm stuck.  I want to install FiPy, but the windows version of  
> python(x,y) is python version 2.6, and the only windows installable  
> version of PySparse is for python 2.5.  I have no windows  
> compilers.  Is there a solution that will get me windows installable  
> PySparse for python2.6?

Hi Kevin,

One way is to install MinGW (www.mingw.org) which is a port of the gcc  
to Windows. It also helps to install MSYS, from the same website,  
which gives you a nice Linux-like terminal. I've compiled Pysparse  
with Python 2.6 so I expect this should work, although I have never  
tried on Windows.

Good luck,
Dominique

[Pysparse-users] windows version compatible with Python 2.6

[Kevin A. <all...@ya...>]

I'm stuck.  I want to install FiPy, but the windows version of python(x,y) is python version 2.6, and the only windows installable version of PySparse is for python 2.5.  I have no windows compilers.  Is there a solution that will get me windows installable PySparse for python2.6?


    
Thanks,
Kevin Allred



      

Re: [Pysparse-users] examples not working

[Dominique O. <dom...@gm...>]

On Tue, Dec 8, 2009 at 10:11 AM, Ranjit Chacko <rjc...@gm...> wrote:
> I just tried running the example code for the vectorized poisson2d
> functions from the documentation and it doesn't seem to work.
>
> This line from poisson2d_vec fails:
> L.put(4.0, d)
> with this error:
> ValueError: object of too small depth for desired array
>
> Any suggestions?

Strange. The example works for me. What is your setup (platform,
compiler, python version, etc.) ?

-- 
Dominique

[Pysparse-users] examples not working

[Ranjit C. <rjc...@gm...>]

I just tried running the example code for the vectorized poisson2d  
functions from the documentation and it doesn't seem to work.

This line from poisson2d_vec fails:
L.put(4.0, d)
with this error:
ValueError: object of too small depth for desired array

Any suggestions?

Re: [Pysparse-users] Full matrix conversion to spmatrix.ll_mat

[Dominique O. <dom...@gm...>]

Hi Rodrigo,

If you look at the jdsym documentation you will see that the input
matrix (and preconditioners) need not be in linked-list format, or any
other Pysparse format for that matter. The matrix object need only
have the shape, matvec and matvec_transp attributes. If you can
generate irow, jcol and val arrays efficiently, then it is easy to
wrap those into a simple sparse matrix class that provides the
required attributes. What kind of preconditioner are you using?

Regards,
Dominique

On Thu, Nov 19, 2009 at 3:44 AM, iza...@t-...
<iza...@t-...> wrote:
>
> Hello Dominique,
>
> Thank you very much for answering. Conversion to scipy sparse formats is not
> as expensive as using put. I can generate the three vectors irow, icol, val
> in a short time. The problem is to put the values in the pysparse ll format.
> I can even generate a scipy linked list format efficiently, but the problem
> is this is not the same as the pysparse linked list format and I can not use
> it with jdsym.
>
> The option to use put is unfortunately very expensive ( I have tried it).  I
> will try to generate the pysparse matrix directly but I am afraid all the
> operations I used are not supported. For example I do a very expensive
> double loop operation in C with Pyrex and I am able to generate a n*n vector
> that I reshape inside python to do further processing in matrix form. Can I
> reshape a vector into pysparse matrix format?
>
> Best regards
>
> Rodrigo
>
> --
> Hi Rodrigo,
>
> Thanks for using PySparse! All comments can help us improve the
> library. A (rough) design decision in PySparse is that matrix
> operations should be "cheap", i.e., O(nnz), for matrices that are
> indeed sparse. If your matrix is dense, I'm afraid any constructor of
> the form spmatrix.ll_mat(K) will require O(n^2) operations. Please
> correct me if I'm wrong but the Matrix Market format lists all the
> nonzero elements of your matrix, and we *do* need all that information
> to construct the matrix. I don't see how to lower this cost. Using a
> different sparse format won't help either.
>
> What could help is detect any exploitable structure in your matrix.
> Yes it is dense but it probably isn't random and may have some pattern
> to it. For instance, it could be (anti-)symmetric, (block-)circulant,
> (block-)Toeplitz, or whatever. I feel that is where savings might be
> found.
>
> Alternatively, is there any chance to bypass the 2-dimensional array
> that you process before building the PySparse matrix, construct the
> PySparse matrix directly and operate on it instead? Or operate on
> arrays of the form (irow,jcol,val) and then use put()?
>
> I hope this helps. Good luck.
>
> --
> Dominique
>
> ---Original-Nachricht-----
> Subject: Re: [Pysparse-users] Full matrix conversion to spmatrix.ll_mat
> Date: Wed, 18 Nov 2009 23:30:24 +0100
> From: Dominique Orban <dom...@gm...>
> To: "iza...@t-..." <iza...@t-...>
>
> On Wed, Nov 18, 2009 at 11:47 AM, iza...@t-...
> <iza...@t-...> wrote:
>> Dear Sirs,
>>
>> Im using pysparse with the eigenvalue solver jdsym. It works great !!.
>> The
>> only problem I have at the moment is the conversion time from the original
>> format of my matrix. Here is a description of my process:
>>
>> - Read a file wit around 20000 points
>> - Process this points and get a very big array (20000*20000) (this part is
>> done with pyrex)
>> - I reshape the array into a matrix (20000,20000)
>> - I do some math ( matrix transpose, matrix scaling, dot product,
>> transpose,
>> etc.)
>> - I get a matrix from which I would like to obtain a couple of
>> eigenvectors
>>
>> In order to use jdsm I need the matrix in linked list format. I have tried
>> several things:
>>
>> - write the matrix in matrix market format (using scipy) and the read with
>> spmatrix.ll_mat_from_mtx
>>   (it takes very very long)
>> - convert directly to sparse coo format and then into  ll format
>>  (I get an error that this format is not supported for conversion)
>> - use the internal routine put(V,index1,index2)
>> (this is faster but still takes a long time compared with the eigenvalue
>> problem)
>>  - I also use
>>    for ii in range(N):
>>     for jj in range(N):
>>         Al[ii,jj]=K[ii,jj]
>> (this is almost the same as using put)
>>
>>
>> Is there any way to have it faster in the right format? This will improve
>> a
>> lot the all process of obtaining the eigenvalues. If would be great  to
>> have
>> something like:
>>
>> spmatrix.ll_mat(K)
>>
>> where K is a full matrix.
>>
>> I appreciate any help with this problem.
>
>



-- 
Dominique

Re: [Pysparse-users] pysparse with ATLAS : modified setup.py needed

[morovia m. <jal...@go...>]

Dear Dominique,


> If libf77blas.a, libcblas.a, libatlas.a are in $HOME/bin/lib, it seems
> that you would need instead:
>
> library_dirs_list= ['/home/viswanath/bin/lib','/usr/lib']
> libraries_list = ['f77blas', 'cblas', 'atlas', 'lapack']
>
> Let us know if that works.
>

Your suggestion has worked and I could import precon !  However, when I
tried to run
the Examples, poisson_test_eig.py,

ImportError occurred,
undefined symbol : _gfortran_st_write in jdsym.so

Any suggestions !

Thanks
Best regards
Viswanath.

Re: [Pysparse-users] pysparse with ATLAS : modified setup.py needed

[Dominique O. <dom...@gm...>]

On Fri, Nov 27, 2009 at 12:10 AM, morovia morovia
<jal...@go...> wrote:

> Thanks for your reply.  I have my atlas libraries installed in $HOME/bin/lib
> where the libraries, libf77blas.a, libcblas.a, libatlas.a are present.
>
> My .bashrc contains :
> export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HOME/bin/lib:$HOME/bin/lib/ATLAS/
>
> I have modified the following lines in the setup.py of pysparse,
>
> library_dirs_list= ['','/home/viswanath/bin/lib/ATLAS','/usr/lib']
> libraries_list = ['atlas','lapack', 'blas']
> superlu_defs = [('USE_VENDOR_BLAS',1)],('USE_VENDOR_ATLAS',1)]
> umfpack_defs = [('DINT', 1), ('CBLAS', 1), ('ATLAS',1)] # with atlas c-blas
> (http://math-atlas.sf.net)
>
> else:
>     umfpack_include_dirs = ['/home/viswanath/bin/lib/ATLAS','amd',
> 'umfpack']
>     umfpack_library_dirs = ['','/home/viswanath/bin/lib/ATLAS']
>     umfpack_libraries = ['atlas', 'cblas', 'm']

If libf77blas.a, libcblas.a, libatlas.a are in $HOME/bin/lib, it seems
that you would need instead:

library_dirs_list= ['/home/viswanath/bin/lib','/usr/lib']
libraries_list = ['f77blas', 'cblas', 'atlas', 'lapack']

Let us know if that works.

-- 
Dominique

Re: [Pysparse-users] pysparse with ATLAS : modified setup.py needed

[morovia m. <jal...@go...>]

Dear Dominique,


> Could you show us the variables you changed in setup.py? Basically for
> ATLAS support you need to link with libatlas, libf77blas and libcblas.
> Those libraries must also be in your LD_LIBRARY_PATH.
>
> We really need to be using the Numpy Distutils in Pysparse !
>
>
Thanks for your reply.  I have my atlas libraries installed in $HOME/bin/lib
where the libraries, libf77blas.a, libcblas.a, libatlas.a are present.

My .bashrc contains :
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HOME/bin/lib:$HOME/bin/lib/ATLAS/

I have modified the following lines in the setup.py of pysparse,

library_dirs_list= ['','/home/viswanath/bin/lib/ATLAS','/usr/lib']
libraries_list = ['atlas','lapack', 'blas']
superlu_defs = [('USE_VENDOR_BLAS',1)],('USE_VENDOR_ATLAS',1)]
umfpack_defs = [('DINT', 1), ('CBLAS', 1), ('ATLAS',1)] # with atlas c-blas
(http://math-atlas.sf.net)

else:
    umfpack_include_dirs = ['/home/viswanath/bin/lib/ATLAS','amd',
'umfpack']
    umfpack_library_dirs = ['','/home/viswanath/bin/lib/ATLAS']
    umfpack_libraries = ['atlas', 'cblas', 'm']

Thanks
Best regards
Viswanath.