On Tue, Dec 20, 2011 at 9:10 PM, Mark Wiebe wrote:
> On Tue, Dec 20, 2011 at 6:24 PM, Geoffrey Irving wrote:
>>
>> Hello,
>>
>> As a followup to the prior thread on bugs in user defined types in
>> numpy, I converted my rational number class from C++ to C and switched
>> to 32 bits to remove the
On Tue, Dec 20, 2011 at 6:24 PM, Geoffrey Irving wrote:
> Hello,
>
> As a followup to the prior thread on bugs in user defined types in
> numpy, I converted my rational number class from C++ to C and switched
> to 32 bits to remove the need for unportable 128 bit numbers. It
> should be usable a
On Tuesday, December 20, 2011, questions anon
wrote:
> ok thanks, a quick try at using it resulted in:
> IndexError: index out of bounds
> but I may need to do abit more investigating to understand how it works.
> thanks
The assumption is that these arrays are all the same shape. If not, then
ex
Sorry about that. I don't think that terminology is commonly used. This
is what I mean.
Let's say I solve the equations and compute the eigenvalues and
eigenvectors for the given two matrices. I call these results
"non-normalized". Then they can be normalized. Once they are normalized if
I mult
This is really excellent. I would like to take a stab at getting this pulled
in to the code base --- and fixing the GIL issue --- if someone hasn't beat me
to it.
Travis
--
Travis Oliphant
(on a mobile)
512-826-7480
On Dec 20, 2011, at 9:24 PM, Geoffrey Irving wrote:
> Hello,
>
> As a fo
ok thanks, a quick try at using it resulted in:
IndexError: index out of bounds
but I may need to do abit more investigating to understand how it works.
thanks
On Wed, Dec 21, 2011 at 2:20 PM, Olivier Delalleau wrote:
> I'm sorry I don't have time to look closely at your code and this may not
>
Ok well I'm sorry, I have no idea what would be the difference between
"non-normalized" and "un-normalized".
In PCA you may decide to scale your eigenvectors by the inverse of the
square root of their corresponding eigenvalue so that your projected data
has unit variance, but it doesn't seem to be
I think I am interested in the non-normalized eigenvectors not the
un-normalized ones. Once the eig function computes the generalized
eigenvectors I would like to use them as they are.
I would think this would be a common request since the normal-mode
frequency response is used in many different f
I'm sorry I don't have time to look closely at your code and this may not
be helpful, but just in case... I find it suspicious that you *seem* (by
quickly glancing at the code) to be taking TIME[max(temperature)] instead
of TIME[argmax(temperature)].
-=- Olivier
2011/12/20 questions anon
> I ha
What I don't get is that "un-normalized" eigenvectors can be pretty much
anything. If you care about the specific output of Matlab / Octave, it
means you understand the particular "un-normalization" that these programs
use. In that case you should be able to recover it from the normalized
output fr
I have a netcdf file that contains hourly temperature data for a whole
month. I would like to find the maximum temperature within that file and
also the corresponding Latitude and Longitude and Time and then plot this.
Below is the code I have so far. I think everything is working except for
identi
I don't think I can do that. I can go to the normalized results but not
the other way.
On Tue, Dec 20, 2011 at 9:45 PM, Olivier Delalleau wrote:
> Hmm, sorry, I don't see any obvious logic that would explain how Octave
> obtains this result, although of course there is probably some logic...
>
Hmm, sorry, I don't see any obvious logic that would explain how Octave
obtains this result, although of course there is probably some logic...
Anyway, since you seem to know what you want, can't you obtain the same
result by doing whatever un-normalizing operation you are after?
-=- Olivier
201
On Tue, Dec 20, 2011 at 7:24 PM, Geoffrey Irving wrote:
> Hello,
>
> As a followup to the prior thread on bugs in user defined types in
> numpy, I converted my rational number class from C++ to C and switched
> to 32 bits to remove the need for unportable 128 bit numbers. It
> should be usable a
Hello,
As a followup to the prior thread on bugs in user defined types in
numpy, I converted my rational number class from C++ to C and switched
to 32 bits to remove the need for unportable 128 bit numbers. It
should be usable as a fairly thorough test case for user defined types
now. It does ra
On Tue, Dec 20, 2011 at 3:16 PM, Jack Bryan wrote:
> Hi,
> I have set up PYTHONPATH :
>
> >>> sys.path
> ['', '/mypath/numpy/lib/python2.7/site-packages', '/
> mypath/python272/lib/python27.zip',
> '/ mypath/python272/lib/python2.7', '/
> mypath/python272/lib/python2.7/plat-linux2',
> '/ mypat
I should include the scipy response too I guess.
scipy.linalg.eig(STIFM, MASSM)
(array([ 3937.15984097+0.j, 3937.15984097+0.j, 3937.15984097+0.j,
3923.07692308+0.j, 3923.07692308+0.j, 7846.15384615+0.j]),
array([[ 1., 0., 0., 0., 0., 0.],
[ 0., 1., 0., 0., 0., 0.],
If I can get the same response as Matlab I would be all set.
Octave results
>> STIFM
STIFM =
Diagonal Matrix
102000000
0 10200000
00 1020000
00
Hmm... ok ;) (sorry, I can't follow you there)
Anyway, what kind of non-normalization are you after? I looked at the doc
for Matlab and it just says eigenvectors are not normalized, without
additional details... so it looks like it could be anything.
-=- Olivier
2011/12/20 Fahreddın Basegmez
>
I am computing normal-mode frequency response of a mass-spring system. The
algorithm I am using requires it.
On Tue, Dec 20, 2011 at 8:10 PM, Olivier Delalleau wrote:
> I'm probably missing something, but... Why would you want non-normalized
> eigenvectors?
>
> -=- Olivier
>
>
> 2011/12/20 Fahr
I'm probably missing something, but... Why would you want non-normalized
eigenvectors?
-=- Olivier
2011/12/20 Fahreddın Basegmez
> Howdy,
>
> Is it possible to get non-normalized eigenvectors from scipy.linalg.eig(a,
> b)? Preferably just by using numpy.
>
> BTW, Matlab/Octave provides this w
Howdy,
Is it possible to get non-normalized eigenvectors from scipy.linalg.eig(a,
b)? Preferably just by using numpy.
BTW, Matlab/Octave provides this with its eig(a, b) function but I would
like to use numpy for obvious reasons.
Regards,
Fahri
___
Hi, I have set up PYTHONPATH :
>>> sys.path['', '/mypath/numpy/lib/python2.7/site-packages', '/
mypath/python272/lib/python27.zip', '/
mypath/python272/lib/python2.7', '/
mypath/python272/lib/python2.7/plat-linux2', '/
mypath /python272/lib/python2.7/lib-tk', '/
mypath/python272/lib/python2.7/lib-
On Tue, Dec 20, 2011 at 10:52 PM, Russell E. Owen wrote:
> In article ,
> "Russell E. Owen" wrote:
>
> > In article
> > ,
> > Ralf Gommers wrote:
> >
> > > On Fri, Dec 9, 2011 at 8:02 PM, Russell E. Owen wrote:
> > >
> > > > I'm trying to build numpy 1.6.1 on Scientific Linux 5 but the unit
In article ,
"Russell E. Owen" wrote:
> In article
> ,
> Ralf Gommers wrote:
>
> > On Fri, Dec 9, 2011 at 8:02 PM, Russell E. Owen wrote:
> >
> > > I'm trying to build numpy 1.6.1 on Scientific Linux 5 but the unit tests
> > > claim the wrong version of fortran was used. I thought I knew h
In article
,
Olivier Delalleau wrote:
> 2011/12/12 Russell E. Owen
>
> > In article
> > ,
> > Ralf Gommers wrote:
> >
> > > On Fri, Dec 9, 2011 at 8:02 PM, Russell E. Owen wrote:
> > >
> > > > I'm trying to build numpy 1.6.1 on Scientific Linux 5 but the unit
> > tests
> > > > claim the wr
On Tue, Dec 20, 2011 at 3:18 PM, Charles R Harris wrote:
> Hi Ralf,
>
> On Mon, Dec 5, 2011 at 12:43 PM, Ralf Gommers > wrote:
>
>> Hi all,
>>
>> It's been a little over 6 months since the release of 1.6.0 and the NA
>> debate has quieted down, so I'd like to ask your opinion on the timing of
>>
On Tue, Dec 20, 2011 at 20:23, Derek Homeier
wrote:
> On 20.12.2011, at 9:01PM, Jack Bryan wrote:
>
>> customize Gnu95FCompiler using config
>> C compiler: gcc -pthread -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3
>> -Wall -Wstrict-prototypes -fPIC
>>
>> compile options: '-Inumpy/core/src/
On 20.12.2011, at 9:01PM, Jack Bryan wrote:
> customize Gnu95FCompiler using config
> C compiler: gcc -pthread -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3
> -Wall -Wstrict-prototypes -fPIC
>
> compile options: '-Inumpy/core/src/private -Inumpy/core/src -Inumpy/core
> -Inumpy/core/src/np
Hi,I run : python setup.py build
and got:
building library "npymath" sourcescustomize GnuFCompilerCould not locate
executable g77Could not locate executable f77customize IntelFCompilerCould not
locate executable ifortCould not locate executable ifccustomize
LaheyFCompilerCould not locate execu
Hi Jack,
> In order to install scipy, I am trying to install numpy 1.6.1. on GNU/linux
> redhat 2.6.18.
>
> But, I got error about fortran compiler.
>
> I have gfortran. I do not have f77/f90/g77/g90.
>
that's good!
> I run :
> python setup.py build --fcompiler=gfortran
>
> It woks well
Hi,
In order to install scipy, I am trying to install numpy 1.6.1. on GNU/linux
redhat 2.6.18.
But, I got error about fortran compiler.
I have gfortran. I do not have f77/f90/g77/g90.
I run :python setup.py build --fcompiler=gfortran
It woks well and tells me that
customize Gnu95FCompilerFo
On Tue, Dec 20, 2011 at 10:31, Marc POINOT wrote:
>
> Hi all,
>
> I've just changed to cython and old numpy module with a raw C API.
> The C module init is removed, and I've put the import_array in the
> 'pure-cython'
> part of the module init. Usual tutorial examples have these lines:
>
> import
Hi Ralf,
On Mon, Dec 5, 2011 at 12:43 PM, Ralf Gommers
wrote:
> Hi all,
>
> It's been a little over 6 months since the release of 1.6.0 and the NA
> debate has quieted down, so I'd like to ask your opinion on the timing of
> 1.7.0. It looks to me like we have a healthy amount of bug fixes and sma
Hi all,
I've just changed to cython and old numpy module with a raw C API.
The C module init is removed, and I've put the import_array in the 'pure-cython'
part of the module init. Usual tutorial examples have these lines:
import numpy as NPY
cimport numpy as NPY
NPY.import_array()
But this f
35 matches
Mail list logo
