Re: [Numpy-discussion] import scipy.linalg is hanging on Marvell armada 370

2015-05-05 Thread Charles R Harris
On Tue, May 5, 2015 at 1:52 PM, Alexander Brezinov < alexander.brezi...@mmbresearch.com> wrote: > Hello > > The import of scipy.linalg is hanging in DOUBLE_mutiply function > (BINARY_LOOP) in umath.so. After attaching the gdb and dumping the local > varibles the args are empty strings. Could you p

Re: [Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Allan Haldane
That makes sense. I think it's the way to go, thanks. The downside is using **kwargs instead of an explicit keepdims arg gives a more obscure signature, but using the new __signature__ attribute this could be hidden for python 3 users and python2 using ipython3+. On 05/05/2015 01:55 PM, Nathaniel

Re: [Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Nathaniel Smith
On May 5, 2015 11:05 AM, "Neil Girdhar" wrote: > > Maybe they should have written their code with **kwargs that consumes all keyword arguments rather than assuming that no keyword arguments would be added? The problem with this approach in general is that it makes writing code unnecessarily convo

[Numpy-discussion] import scipy.linalg is hanging on Marvell armada 370

2015-05-05 Thread Alexander Brezinov
Hello The import of scipy.linalg is hanging in DOUBLE_mutiply function (BINARY_LOOP) in umath.so. After attaching the gdb and dumping the local varibles the args are empty strings. Could you please advise if this is known issue? I just search the mailing list and could not find any solution for th

Re: [Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Sebastian Berg
On Di, 2015-05-05 at 11:13 -0400, Allan Haldane wrote: > Hello all, > > A question: > > Many ndarray methods (eg sum, mean, any, min) have a "keepdims" keyword > argument, but ndarray subclass methods sometimes don't. The 'matrix' > subclass doesn't, and numpy functions like 'np.sum' intentionall

Re: [Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Neil Girdhar
Maybe they should have written their code with **kwargs that consumes all keyword arguments rather than assuming that no keyword arguments would be added? The problem with this approach in general is that it makes writing code unnecessarily convoluted. On Tue, May 5, 2015 at 1:55 PM, Nathaniel Sm

Re: [Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Nathaniel Smith
AFAICT the only real solution here is for np.sum and friends to propagate the keepdims argument if and only if it was explicitly passed to them (or maybe the slightly different, if and only if it has a non-default value). If we just started requiring code to handle it and passing it unconditionally

[Numpy-discussion] Should ndarray subclasses support the keepdims arg?

2015-05-05 Thread Allan Haldane
Hello all, A question: Many ndarray methods (eg sum, mean, any, min) have a "keepdims" keyword argument, but ndarray subclass methods sometimes don't. The 'matrix' subclass doesn't, and numpy functions like 'np.sum' intentionally drop/ignore the keepdims argument when called with an ndarray subcl

Re: [Numpy-discussion] read not byte aligned records

2015-05-05 Thread Benjamin Root
I have been very happy with the bitarray package. I don't know if it is faster than bitstring, but it is worth a mention. Just watch out for any hashing operations on its objects, it doesn't seem to do them right (set(), dict(), etc...), but comparison operations work just fine. Ben Root _

Re: [Numpy-discussion] read not byte aligned records

2015-05-05 Thread aymeric . rateau
Hi, To answer Jerome (I hope), data is sometime spread on bytes shared by other data in the whole record. 10 bits was an example, sometimes, 24, 2, 8, 7 etc. all combined including some padding between them. I am not sure to have understood... To Nathaniel, yes indeed I could read the records i