Mailman 3 March 2009 - NumPy-Discussion

Build Failure on WIndows Vista
by Patrick Marsh March 7, 2009

March 7, 2009

Greetings, I am running Windows Vista Ultimate and trying to build numpy from the SVN branch using MSVC 2003. I have been able to build previously, but with my latest SVN update I am no longer able to build. My CPU is an Intel Core2 T7600 @2.33GHz. The error is below. e:\svn\numpy\numpy\core\include\numpy\npy_cpu.h(44) : fatal error C1189: #error: Unknown CPU, please report this to numpy maintainers with information about your platform (OS, CPU and compiler) error: Command "D:\Program Files\Microsoft Visual Studio 2003\bin\cl.exe /c /nologo /Ox /MD /W3 /GX /DNDEBUG -Inumpy\core\include -Ibuild\src.win32-2.5\numpy\core\include/numpy -Inumpy\core\src -Inumpy\core\include -ID:\Python25\include -ID:\Python25\PC /Tcbuild\src.win32-2.5\numpy\core\src\_sortmodule.c /Fobuild\temp.win32-2.5\Release\build\src.win32-2.5\numpy\core\src\_sortmodule.obj" failed with exit status 2 -Patrick -- Patrick Marsh Graduate Research Assistant School of Meteorology University of Oklahoma http://www.patricktmarsh.com

2 1

Re: [Numpy-discussion] can't take FFT of ndarray
by charles reid March 6, 2009

March 6, 2009

Hi there - I've imported some data from a file, and it's in a list called mixfrac. I'd like to take the Fourier transform of the data, but when I try to take the FFT of the list, I get this error: --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /Users/charles/apriori/read.py in <module>() ----> 1 2 3 4 5 /Library/Python/2.5/site-packages/numpy-1.3.0.dev5825-py2.5-macosx-10.3-i386.egg/numpy/fft/fftpack.pyc in fft(a, n, axis) 105 """ 106 --> 107 return _raw_fft(a, n, axis, fftpack.cffti, fftpack.cfftf, _fft_cache) 108 109 /Library/Python/2.5/site-packages/numpy-1.3.0.dev5825-py2.5-macosx-10.3-i386.egg/numpy/fft/fftpack.pyc in _raw_fft(a, n, axis, init_function, work_function, fft_cache) 64 if axis != -1: 65 a = swapaxes(a, axis, -1) ---> 66 r = work_function(a, wsave) 67 if axis != -1: 68 r = swapaxes(r, axis, -1) TypeError: array cannot be safely cast to required type so I convert to an array and run fft(mixfracarray). mixfracarray = array(mixfrac) fft(mixfracarray) whereupon I recieve the error --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /Users/charles/apriori/read.py in <module>() ----> 1 2 3 4 5 /Library/Python/2.5/site-packages/numpy-1.3.0.dev5825-py2.5-macosx-10.3-i386.egg/numpy/fft/fftpack.pyc in fft(a, n, axis) 105 """ 106 --> 107 return _raw_fft(a, n, axis, fftpack.cffti, fftpack.cfftf, _fft_cache) 108 109 /Library/Python/2.5/site-packages/numpy-1.3.0.dev5825-py2.5-macosx-10.3-i386.egg/numpy/fft/fftpack.pyc in _raw_fft(a, n, axis, init_function, work_function, fft_cache) 64 if axis != -1: 65 a = swapaxes(a, axis, -1) ---> 66 r = work_function(a, wsave) 67 if axis != -1: 68 r = swapaxes(r, axis, -1) TypeError: array cannot be safely cast to required type This is strange, because I can run fft(array([0,0,0,1,1,1])), or fft([0,0,0,1,1,1]), perfectly fine. This is passing an array and a list, respectively. type(mixfrac) is list and size(mixfrac) is 100; type(mixfracarray) is ndarray, and mixfracarray.shape is (100,). I've also tried taking the FFT of the transpose of mixfracarray, but that doesn't work either. I'm stumped - why can't I run an FFT on either mixfrac or mixfracarray? Charles

3 7

numpy-svn mails
by Ryan May March 6, 2009

March 6, 2009

Hi, Is anyone getting mails of the SVN commits? I've gotten 1 spam message from that list, but no commits. Ryan -- Ryan May Graduate Research Assistant School of Meteorology University of Oklahoma

3 2

Cython numerical syntax revisited
by Dag Sverre Seljebotn March 6, 2009

March 6, 2009

This is NOT yet discussed on the Cython list; I wanted to check with more numerical users to see if the issue should even be brought up there. The idea behind the current syntax was to keep things as close as possible to Python/NumPy, and only provide some "hints" to Cython for optimization. My problem with this now is that a) it's too easy to get non-optimized code without a warning by letting in untyped indices, b) I think the whole thing is a bit too "magic" and that it is too unclear what is going on to newcomers (though I'm guessing there). My proposal: Introduce an explicit "buffer syntax": arr = np.zeros(..) cdef int[:,:] buf = arr # 2D buffer Here, buf would be something else than arr; it is a seperate view to the array for low-level purposes. This has certain disadvantages; consider: a1 = np.zeros(...) + 1; a2 = np.zeros(...) + 2 cdef int[:] b1 = a1, b2 = a2 Here, one would need to use b1 and b2 for for-loop arithmetic, but a1 and a2 for vectorized operations and slicing. "b1 + b2" would mean something else and not-NumPy-related (at first disallowed, but see below). "print b1" would likely coerce b1 to a Python memoryview and print "<memoryview ...>" (at least on newer Python versions); one would need to use some function to get b1 back to a NumPy array. Advantages: - More explicit - Leaves a path open in the syntax for introducing low-level slicing and arithmetic as seperate operations in Cython independent of NumPy (think Numexpr compile-time folded into Cython code). - Possible to have some creative syntax like "int[0:]" for disallowing negative wraparound and perhaps even "int[-3:]" for non-zero-based indexing. More details: http://wiki.cython.org/enhancements/buffersyntax (Of course, compatability with existing code will be taken seriously no matter how this plays out!) -- Dag Sverre

8 22

Interpolation via Fourier transform
by Nadav Horesh March 6, 2009

March 6, 2009

I apology for this off topic question: I have a 2D FT of size N x N, and I would like to reconstruct the original signal with a lower sampling frequency directly (without using an interpolation procedure): Given M < N the goal is to compute a M x M "time domain" signal. In the case of 1D signal the trick is simple --- given a length N freq. domain Sig: sig = np.fft.ifft(Sig, M) This trick does not work in 2D: sig = np.fft.ifft2(Sig, (M,M)) is far from being the right answer. Any ideas?

4 5

Ticket mailing down?
by Charles R Harris March 6, 2009

March 6, 2009

I'm not receiving notifications of new/modified tickets. Anyone else having this problem? ... Chuck

2 3

indexing question
by Robin March 5, 2009

March 5, 2009

Hi, I have an indexing problem, and I know it's a bit lazy to ask the list, sometime when people do interesting tricks come up so I hope no one minds! I have a 2D array X.shape = (a,b) and I want to change it into new array which is shape (2,(a*b)) which has the following form: [ X[0,0], X[0,1] X[1,0], X[1,1] X[2,0], X[2,1] .... X[a,0], X[a,1] X[0,1], X[0,2] X[1,1], X[1,2] ... ] The first access is trials and the second axis is the different outputs, which I am trying to represent as an overlapping sliding window of 2 samples (but in general n). Because of the repeats I'm not sure if I can do it without loops. Thanks for any help Cheers Robin

2 5

Missing svn mailings?
by Charles R Harris March 5, 2009

March 5, 2009

It looks like the new system is failing to mail svn commit notifications... Chuck

2 1

Numpy array in iterable
by Kim Hansen March 5, 2009

March 5, 2009

Hi Numpy discussions Quite often I find myself wanting to generate a boolean mask for fancy slicing of some array, where the mask itself is generated by checking if its value has one of several relevant values (corresponding to states) So at the the element level thsi corresponds to checking if element in iterable But I can't use the in operator on a numpy array: In [1]: test = arange(5) In [2]: states = [0, 2] In [3]: mask = test in states --------------------------------------------------------------------------- ValueError Traceback (most recent call last) C:\Documents and Settings\kha\<ipython console> in <module>() ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all() I can however make my own utility function which works effectively the same way by iterating through the states In [4]: for i, state in enumerate(states): ...: if i == 0: ...: result = test == state ...: else: ...: result |= test == state ...: ...: In [5]: result Out[5]: array([ True, False, True, False, False], dtype=bool) However, I would have thought such an "array.is_in()" utility function was already available in the numpy package? But I can't find it, and I am curious to hear if it is there or if it just available in another form which I have simply overlooked. If it is not there I think it could be a nice extra utility funtion for the ndarray object. --Slaunger

3 10

calling _import_array() twice crashes python
by Soeren Sonnenburg March 5, 2009

March 5, 2009

Dear all, I've written a wrapper enabling to run python code from within octave (and vice versa). To this end I am embedding python in octave. So I am calling Py_Initialize(); _import_array(); <do something with numpy> Py_Finalize(); multiple times. While things work nicely on the first run, I am getting a crash on _import_array() on the second run... Is there any cleanup function I should potentially call that would prevent this? Soeren

2 1