[Numpy-discussion] two previously unresolved issues

Hi, I just went through my mail archive and found these two minor outstanding issues. Thought I'd ask for comments before the new release: """ From: "Charles R Harris" <[EMAIL PROTECTED]> Subject: Re: [Numpy-discussion] Assign NaN, get zero On 11/11/06, Lisandro Dalcin <[EMAIL PROTECTED]> wrote

Re: [Numpy-discussion] matrix indexing question

On 3/27/07, Alan Isaac <[EMAIL PROTECTED]> wrote: > > On 3/27/07, Alan Isaac <[EMAIL PROTECTED]> wrote: > >> May I see a use case where the desired > >> return when iterating through a matrix > >> is rows as matrices? That has never > >> been what I wanted. > > > On Tue, 27 Mar 2007, Bill Baxter w

Re: [Numpy-discussion] matrix indexing question

> On 3/27/07, Alan Isaac <[EMAIL PROTECTED]> wrote: >> May I see a use case where the desired >> return when iterating through a matrix >> is rows as matrices? That has never >> been what I wanted. On Tue, 27 Mar 2007, Bill Baxter wrote: > AllMyPoints = mat(rand(100,2)) # 100 two-d points

Re: [Numpy-discussion] matrix indexing question

On 3/27/07, Alan Isaac <[EMAIL PROTECTED]> wrote: > May I see a use case where the desired > return when iterating through a matrix > is rows as matrices? That has never > been what I wanted. If you use a row vector convention it make plenty of sense. AllMyPoints = mat(rand(100,2)) # 100 two-d p

Re: [Numpy-discussion] matrix indexing question

On Mon, 26 Mar 2007, Travis Oliphant wrote: > It actually has been offered. You just don't accept it. > Matrices are containers of matrices. > If M is an (mxn) matrix then M[0] is a (1xn) matrix. > Viewing this 1xn matrix as a 1-d array loses it's row-vectorness. > This seems perfectly l

Re: [Numpy-discussion] matrix indexing question

On Mon, 26 Mar 2007, Charles R Harris wrote: > What happens is a convention Certainly true. > It isn't going to change now, it would break too much > code. That is a different kind of argument. It might be true. May I see a use case where the desired return when iterating through a matrix is

[Numpy-discussion] ATLAS problems

Hi All, I'm trying to compile NumPy using the ATLAS libraries but am having some difficulty with it. I've managed to build numpy such that lapack_lite.so is linked as follows: linux-gate.so.1 => (0xe000) libatlas.so => /usr/local/atlas/lib/libatlas.so (0xb79e9000) li

Re: [Numpy-discussion] defmatrix.py

On 3/26/07, Travis Oliphant <[EMAIL PROTECTED]> wrote: Charles R Harris wrote: > > > The rule 1-d is always a *row* vector only applies when converting > to a > matrix. > > In this case, the dot operator does not "convert to a matrix" but > uses > rules for operating wit

Re: [Numpy-discussion] defmatrix.py

Charles R Harris wrote: > > > The rule 1-d is always a *row* vector only applies when converting > to a > matrix. > > In this case, the dot operator does not "convert to a matrix" but > uses > rules for operating with mixed 2-d and 1-d arrays inherited from > Numeric. >

Re: [Numpy-discussion] defmatrix.py

On 3/26/07, Travis Oliphant <[EMAIL PROTECTED]> wrote: > I think that might be the simplest thing, dot overrides subtypes. BTW, > here is another ambiguity > > In [6]: dot(array([[1]]),ones(2)) > --- > exceptions.ValueError

Re: [Numpy-discussion] Fixed scalar coercion model in NumPy

On Mon, Mar 26, 2007 at 01:52:28PM -0700, Travis Oliphant wrote: > I really do want to get 1.0.2 out the door soon. What still needs to be > fixed before then? The code in ticket 469 still causes a memory error, so that might be worth fixing. Cheers Stéfan __

Re: [Numpy-discussion] defmatrix.py

> I think that might be the simplest thing, dot overrides subtypes. BTW, > here is another ambiguity > > In [6]: dot(array([[1]]),ones(2)) > --- > exceptions.ValueErrorTraceback (most > recent

Re: [Numpy-discussion] defmatrix.py

On 3/26/07, Travis Oliphant <[EMAIL PROTECTED]> wrote: Charles R Harris wrote: > > > On 3/26/07, *Travis Oliphant* <[EMAIL PROTECTED] > > wrote: > > Charles R Harris wrote: > > > > > > > On 3/26/07, *Nils Wagner* <[EMAIL PROTECTED] >

Re: [Numpy-discussion] Fixed scalar coercion model in NumPy

Perry Greenfield wrote: >Great! > >On Mar 26, 2007, at 4:52 PM, Travis Oliphant wrote: > > > >>I've finally made the changes to fix the scalar coercion model >>problems >>in NumPy 1.0.1 >> >>Now, scalar coercion rules only apply when involved types are of the >>same basic "kind". >> >> Ac

Re: [Numpy-discussion] Fixed scalar coercion model in NumPy

Great! On Mar 26, 2007, at 4:52 PM, Travis Oliphant wrote: > I've finally made the changes to fix the scalar coercion model > problems > in NumPy 1.0.1 > > Now, scalar coercion rules only apply when involved types are of the > same basic "kind". [...] ___

[Numpy-discussion] Fixed scalar coercion model in NumPy

I've finally made the changes to fix the scalar coercion model problems in NumPy 1.0.1 Now, scalar coercion rules only apply when involved types are of the same basic "kind". Thus, array([1,2,3],int8)*10 returns an int8 array but array([1,2,3],int8)*10.0 returns a float64 array. If you wa

Re: [Numpy-discussion] defmatrix.py

Charles R Harris wrote: > > > On 3/26/07, *Travis Oliphant* <[EMAIL PROTECTED] > > wrote: > > Charles R Harris wrote: > > > > > > > On 3/26/07, *Nils Wagner* <[EMAIL PROTECTED] > > > >> wro

Re: [Numpy-discussion] matrix indexing question

On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: >> On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: >>> One would expect the iteration over A to return row >>> vectors, represented by (1, n) matrices. > On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: >> This is again simple a

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac wrote: >>>On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: >>> >>> One would expect the iteration over A to return row vectors, represented by (1, n) matrices. > > > > >>On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: >> >> >>>Th

Re: [Numpy-discussion] matrix indexing question

> Since matrices are an iterable Python object, > we *expect* to iterate over the contained objects. > (Arrays.) I am not sure why this is not evident to all, > but it is surely the sticking point in this discussion. > > A matrix is not a container of matrices. > That it acts like one is surprsing

Re: [Numpy-discussion] defmatrix.py

On 3/26/07, Travis Oliphant <[EMAIL PROTECTED]> wrote: Charles R Harris wrote: > > > On 3/26/07, *Nils Wagner* <[EMAIL PROTECTED] > > wrote: > > Charles R Harris wrote: > > > > > > On 3/25/07, *Nils Wagner* <[EMAIL PROTECTED] >

Re: [Numpy-discussion] matrix indexing question

>> On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: >>> One would expect the iteration over A to return row >>> vectors, represented by (1, n) matrices. > On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: >> This is again simple assertion. >> **Why** would "one" expect this? >>

Re: [Numpy-discussion] defmatrix.py

Charles R Harris wrote: > > > On 3/26/07, *Nils Wagner* <[EMAIL PROTECTED] > > wrote: > > Charles R Harris wrote: > > > > > > On 3/25/07, *Nils Wagner* <[EMAIL PROTECTED] > > > >> wrote: >

Re: [Numpy-discussion] matrix indexing question

On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: > One would expect the iteration over A to return row > vectors, represented by (1, n) matrices. This is again simple assertion. **Why** would "one" expect this? Some people clearly do

Re: [Numpy-discussion] defmatrix.py

On 3/26/07, Nils Wagner <[EMAIL PROTECTED]> wrote: Charles R Harris wrote: > > > On 3/25/07, *Nils Wagner* <[EMAIL PROTECTED] > > wrote: > > Hi, > > Ok, things should be working now. In order to get rid of the scipy problems I just needed to rm the scipy package

Re: [Numpy-discussion] New Operators in Python

On 3/26/07, René Bastian <[EMAIL PROTECTED]> wrote: May be it would be possible to implement a class with user "definissable" (?) signs. Yes, it is possible and is done. See this recipe to define an Infix operator class either: x |op| y or: x <> y http://aspn.activestate.com/ASPN/Cookbook/

Re: [Numpy-discussion] Tuning sparse stuff in NumPy

Ok, I did and the results are: csc * csc: 372.601957083 csc * csc: 3.90811300278 csr * csc: 15.3202679157 csr * csr: 3.84498214722 Mhm, quite insightful. Note, that in an operation X.transpose() * X, where X is csc_matrix, then X.tranpose() is automatically cast to csr_matrix. A re-cast to csc

Re: [Numpy-discussion] fastest way to do multiplication with diagonal matrices from left or right

[EMAIL PROTECTED] wrote: > Dear list > > what is the fastet way to multiply with a diagonal matrix from left or > right and without to build a square matrix from the diagonal. Use broadcasting to do your work for you. from numpy import array, newaxis diags = array([...]) mymatrix = array(

Re: [Numpy-discussion] matrix indexing question

On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: > One would expect the iteration over A to return row > vectors, represented by (1, n) matrices. This is again simple assertion. **Why** would "one" expect this? Some people clearly do not. One person commented that this unexpected beh

Re: [Numpy-discussion] fastest way to do multiplication with diagonal matrices from left or right

This is a small function I use to speed up multiplication with diagonal matrices. I don't know if it's the *fastest* way to do it, but it's pretty fast. def mult_diag(d, mtx, left=True): """Multiply a full matrix by a diagonal matrix. This function should always be faster than dot. In

Re: [Numpy-discussion] New Operators in Python

On Mar 24, 3:48 pm, "Charles R Harris" <[EMAIL PROTECTED]> wrote: > > It is somewhat workable as it stands, but I think it would be nicer if > > we could have some "meta" operator that allowed an alternative > > definition of major operators. Something like @* for example (just > > picking a ch

Re: [Numpy-discussion] nd_image.affine_transform edge effects

Hello folks, Hmm, this is worrisome. There really shouldn't be ringing on continuous-tone images like Lena -- right? (And at no step in an image like that should gaussian filtering be necessary if you're doing spline interpolation -- also right?) That's hard to say. Just because it's mainly a

Re: [Numpy-discussion] Array of Arrays

Alexander Michael kirjoitti: > On 3/23/07, Nadav Horesh <[EMAIL PROTECTED]> wrote: > >> How about >> >> a = empty((5,7,4)) >> c = a[...,-1] >> > > Solely because I want to use the array with code that assumes it is > working with two-dimensional arrays but yet only performs operations > o

[Numpy-discussion] fastest way to do multiplication with diagonal matrices from left or right

Dear list what is the fastet way to multiply with a diagonal matrix from left or right and without to build a square matrix from the diagonal. Here it what I am looking for: import numpy as N def diagmult(X, Y): """ Matrix multiplication X*Y where either X or Y is a diagonal matrix.

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac wrote: > On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: >> Perhaps things would be clearer if we thought of the >> constituent groups of data in a matrix as being themselves >> matrices. > > This "thinking of" is what you have suggested before. > You need to explain wh

Re: [Numpy-discussion] Tuning sparse stuff in NumPy

David Koch wrote: > On 3/26/07, Robert Cimrman <[EMAIL PROTECTED]> wrote: >> >> Could you be more specific on which type of the sparse matrix storage >> did you use? > > > > Hi Robert, > > I used csc_matrix. OK, good. Would you mind measuring csc * csr, csc * csc, csr * csc and csr * csr? I am

Re: [Numpy-discussion] matrix indexing question

On Mon, 26 Mar 2007, Sebastian Haase apparently wrote: > A "matrix" is an object that you expect a certain > (mathematical !) behavior from. If some object behaves > intuitively right -- that's ultimately pythonic! The problem is, as I am not the only one to point out, this particular behavior

Re: [Numpy-discussion] Tuning sparse stuff in NumPy

On 3/26/07, Robert Cimrman <[EMAIL PROTECTED]> wrote: Could you be more specific on which type of the sparse matrix storage did you use? Hi Robert, I used csc_matrix. /David ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://pro

Re: [Numpy-discussion] header file location upon installation

On Mar 25, 2007, at 9:44 PM, Robert Kern wrote: For building extensions, either use numpy.distutils, which will take care of everything for you, or use numpy.get_include() to get the directory with headers. get_include() does the trick. Thanks for your help. -- Daniel Wheeler ___

Re: [Numpy-discussion] matrix indexing question

On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: > > Alan G Isaac schrieb: > >> What feels wrong: iterating over a container does not give > >> access to the contained objects. This is not Pythonic. > > On Mon, 26 Mar 2007, Sven Schreiber apparently wrote: > > If you iterate over the rows of th

Re: [Numpy-discussion] Tuning sparse stuff in NumPy

David Koch wrote: > Hi, > > I ran some tests on the very same matrices in Matlab/Numpy and it seems > that > for sparse matrix multipilcation to be faster than dense multiplication - > the degree of sparsity has to be much higher in Numpy than in Matlab. Is > there anything I can tune in the under

[Numpy-discussion] Tuning sparse stuff in NumPy

Hi, I ran some tests on the very same matrices in Matlab/Numpy and it seems that for sparse matrix multipilcation to be faster than dense multiplication - the degree of sparsity has to be much higher in Numpy than in Matlab. Is there anything I can tune in the underlying routines? I need good per

Re: [Numpy-discussion] matrix indexing question

> Alan G Isaac schrieb: >> What feels wrong: iterating over a container does not give >> access to the contained objects. This is not Pythonic. On Mon, 26 Mar 2007, Sven Schreiber apparently wrote: > If you iterate over the rows of the matrix, it feels > natural to me to get the row vectors

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac schrieb: > > What feels wrong: iterating over a container does not give > access to the contained objects. This is not Pythonic. If you iterate over the rows of the matrix, it feels natural to me to get the row vectors -- and as you know a 1d-array does not contain the information

Re: [Numpy-discussion] matrix indexing question

> Alan G Isaac schrieb: >> >>> X[1] >> array([3,4]) >> >>> X[1,:] >> matrix([[3, 4]]) >> But again the point is: >> indexing for submatrices should produce matrices. >> Normal Python indexing should access the constituent arrays. On Mon, 26 Mar 2007, Sven Schreiber apparen

Re: [Numpy-discussion] Adapting algorithm to accept Scipy sparse matrix

Hi again, I want to select/access several columns from a sparse csc_matrix. The only way I could think of is the following enormously inefficient algorithm which basically initalizes a new lil_matrix (for assigments) and loops over all the specified columns and does sparse -> dense -> sparse. All

Re: [Numpy-discussion] matrix indexing question

On Mon, 26 Mar 2007, "Colin J. Williams" apparently wrote: > Perhaps things would be clearer if we thought of the > constituent groups of data in a matrix as being themselves > matrices. This "thinking of" is what you have suggested before. You need to explain why it is not begging the questio

Re: [Numpy-discussion] New Operators in Python

Zachary Pincus wrote: > > Anyhow, feel free to disagree with me -- I'm no expert here. I'm only > mentioning this as a public service to make it clear that most of > what's being proposed in this thread is, for better or worse, 100% > dead-in-the-water for Python 3, and the rest will have a

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac wrote: >> Alan G Isaac wrote: >>> So this :: >>> >>> x[1] >>> matrix([[1, 0]]) >>> feels wrong. (Similarly when iterating across rows.) >>> Of course I realize that I can just :: >>> >>> x.A[1] >>> array([1, 0]) > > > On Sun, 25 Mar 2007, "Colin J. Williams"

Re: [Numpy-discussion] Modular toolkit for Data Processing 2.1 released!

Tiziano Zito schrieb: > The optional symeig module contains a Python wrapper for the LAPACK > functions to solve the standard and generalized eigenvalue problems > for symmetric (hermitian) positive definite matrices. Those > specialized algorithms give an important speed-up with respect to the >

Re: [Numpy-discussion] defmatrix.py

Charles R Harris wrote: > > > On 3/25/07, *Nils Wagner* <[EMAIL PROTECTED] > > wrote: > > Hi, > > > > > BTW, I can't import scipy.sparse, I get the following error: > > ImportError: cannot import name densetocsr > > What am I doing wrong? > > Chuck > ---

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac schrieb: > Oooops, they should match of course. :: > >>> X[1] > array([3,4]) > >>> X[1,:] > matrix([[3, 4]]) > > But again the point is: > indexing for submatrices should produce matrices. > Normal Python indexing should access the constituent arrays. > I think this

Re: [Numpy-discussion] matrix indexing question

Alan G Isaac wrote: >>> On 3/26/07, Alan G Isaac <[EMAIL PROTECTED]> wrote: finds itself in basic conflict with the idea that I ought to be able to iterate over the objects in an iterable container. I mean really, does this not "feel" wrong? :: > >>> for item in x: prin

Re: [Numpy-discussion] matrix indexing question

Bill Baxter wrote: > On 3/26/07, Colin J. Williams <[EMAIL PROTECTED]> wrote: >> Bill Baxter wrote: >>> This may sound silly, but I really think seeing all those brackets is >>> what makes it feel wrong. Matlab's output doesn't put it in your >>> face that your 4 is really a matrix([[4]]), even

[Numpy-discussion] random state not portable?

Dear Numpy devels, First of all, thank you for the great job done so far! We've been updating our MDP module, and had a chance to appreciate it. We found an issue in the random module: the random state contains an array of int32 or int64 depending on the architecture, making it not portable. For

[Numpy-discussion] Modular toolkit for Data Processing 2.1 released!

MDP version 2.1 and symeig 1.2 have been released! What's new in version 2.1? -- - Fully compatible with NumpPy 1.0, the first stable release of the descendant of the Numeric python extension module - symeig project resumed and updated - For increased speed, scipy and

Re: [Numpy-discussion] New Operators in Python

Hi folks, Sorry to rain on this parade, but unicode variable names and/or other syntactic elements have already been rejected for Python 3: http://www.python.org/dev/peps/pep-3099/ > Python 3000 source code won't use non-ASCII Unicode characters for > anything except string literals or commen

Re: [Numpy-discussion] New Operators in Python

The unicode keyboards sailing everywhere is just a matter of time And python 2-symbol operators soon will look obsolete, this will increase migrating from python to Sun fortress etc. I took a look at their unicode syntax for math formulas http://research.sun.com/projects/plrg/faq/NAS-CG.pdf it lo

Re: [Numpy-discussion] nd_image.affine_transform edge effects

Thanks for the information and the paper link, James. I certainly appreciate the perspective, and now see why the anti-aliasing and reconstruction filtering might best be left to clients of a resampling procedure. Hopefully at least some of the kinks in the spline interpolation (to date: o

[Numpy-discussion] nd_image.affine_transform edge effects

PS... (sorry for all the posts, for anyone who isn't interested...) > Agreed, it looks like aliasing. Nevertheless, any resampling > procedure is supposed to deal with this internally, right? Either by > lowpass filtering (traditional case), or by spline fitting (spline > case as described by Unse

Re: [Numpy-discussion] New Operators in Python

Hello, I am interest both in numarray type multiplication and matrix type multiplication. But I am not shure that I can buy an Unicode keyboard. May be it would be possible to implement a class with user "definissable" (?) signs. My choice : a * b -> numarray type multi a !* b -> matrix --

Re: [Numpy-discussion] dtype confusion

I'm afraid I'm not quite done with this, yet. There's still more I am confused about, even after reading the manual (again). There is something fundamentally weird about record arrays, that doesn't seem to click with me. Please have a look at this piece of code: import numpy as N newtype = N.dtyp