Mailman 3 July 2010 - NumPy-Discussion

module compiled against ABI version 2000000 but this version of numpy is 1000009
by Peter Schmidtke 29 Jul '10

29 Jul '10

Hi all, I am trying to install manually the latest releases of scipy and numpy on Mac OSX 10.6 Snow Leopard. I previously used the dmg installer that is available, but the numpy version is too new for some other modules I have and need on my machine, so I went for a manual install of numpy 1.4 and scipy 0.8. I can import numpy without problems, but when I import scipy.cluster for example, I get an error message like : module compiled against ABI version 2000000 but this version of numpy is 1000009 I understand that both scipy and numpy are not compatible...but what numpy version should I use? I already use the newest release and I cannot go for the svn version as pycuda won't work as it should anymore. Thanks in advance for your lights on this. Peter Schmidtke ----------------- PhD Student Department of Physical Chemistry School of Pharmacy University of Barcelona Barcelona, Spain

2 2

1.5 release schedule proposal
by Ralf Gommers 29 Jul '10

29 Jul '10

Hi all, For numpy 1.5.0 no one has yet said they have urgent changes that need to go in. If you do, please reply with the what and why. If nothing big has to go in, I propose the following release schedule: Aug 1 : beta 1 Aug 15: rc 1 Aug 22: rc 2 Aug 29: release Please note that the branch has been created already, so no new features should go in anymore. If you have urgent tickets, please let's hear those as well. My small laundry list is pasted below. 1242 (no tests for matlib) 1343 (OS X build error) 1124 (remove string exceptions) 1416 (frompyfunc segfault) 1088 (add cython/swig examples) 1502 (msvc, isnan patch) 1499 (msvc, ldexp patch) 1438 (trapz patch) 1491 (str comparison patch) 1164 (distutils, fcompiler/pg.py patch) 1448 (distutils, Intel compiler detection patch) Cheers, Ralf

12 24

Problem with importing numpy in Ubuntu
by Robert Faryabi 28 Jul '10

28 Jul '10

I am new to numpy. Hopefully this is a correct forum to post my question. I have Ubuntu Luci system. I installed Python 2.6.5 and Python 3.0 as well as python-numpy using Ubuntu repository. When I import the numpy into python, I get the following error. >> import numpy Traceback (most recent call last): File "<stdin>", line 1, in <module> ImportError: No module named numpy The package cannot be located. Then I tried to point the interpreter to the numpy >>> sys.path.append('/usr/lib/ python2.6/dist-packages') and import it >>> import numpy I get the following error >>> import numpy Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.6/dist-packages/numpy/__init__.py", line 130, in <module> import add_newdocs File "/usr/lib/python2.6/dist-packages/numpy/add_newdocs.py", line 9, in <module> from lib import add_newdoc File "/usr/lib/python2.6/dist-packages/numpy/lib/__init__.py", line 4, in <module> from type_check import * File "/usr/lib/python2.6/dist-packages/numpy/lib/type_check.py", line 8, in <module> import numpy.core.numeric as _nx File "/usr/lib/python2.6/dist-packages/numpy/core/__init__.py", line 5, in <module> import multiarray ImportError: /usr/lib/python2.6/dist-packages/numpy/core/multiarray.so: undefined symbol: _PyUnicodeUCS4_IsWhitespace

6 18

Subclassing ndarray in C: getitem ?
by Pierre GM 28 Jul '10

28 Jul '10

All, I'm teaching myself how to subclass ndarrays in C (not in Cython, just plain C). It's slowly coming together, but I'm now running into a problem: I need to overwrite __getitem__ and I'm not sure how to do it. I was thinking about using my own function instead of array_as_mapping.mp_subscript, but how do I import array_as_mapping in the first place ? And is it really the way to go ? Any pointer appreciated... P.

3 4

ceil returns real ?
by Mark Bakker 28 Jul '10

28 Jul '10

Hello list, I don't understand why ceil and floor return real values, while the doc string says: The ceil of the scalar `x` is the smallest integer `i` Wouldn't an integer make more sense? Numpy version 1.3.0. Thanks, Mark

6 7

Is there anyway to read raw binary file via pytable?
by 脑关(BrainGateway)生命科学仪器 28 Jul '10

28 Jul '10

it seems like pytable only support HDF5. I had some 500GB numerical arrays to process. Pytable claims to have some advance feature to enhance processing speed and largely reduce physical memory requirement. However, I do not wanna touch the raw data I had. Simply because I do not have doubled diskspace to covert all 10TB data into HDF5. Is there any way to let pytable read raw binary files or alternatively to package raw files into HDF5 format without change the files themselves.? Thanks Brain Gateway

5 5

how to add columns
by wheres pythonmonks 28 Jul '10

28 Jul '10

I have a rec array and I want to add an additional column. I've seen at least two solutions to this problem: mlab.rec_append_fields (matplotlib) And append_field from http://mail.scipy.org/pipermail/numpy-discussion/2007-September/029357.html In [19]: def append_field(rec, name, arr, dtype=None): arr = np.asarray(arr) if dtype is None: dtype = arr.dtype newdtype = np.dtype(rec.dtype.descr + [(name, dtype)]) newrec = np.empty(rec.shape, dtype=newdtype) for field in rec.dtype.fields: newrec[field] = rec[field] newrec[name] = arr return newrec Is there a best solution? I don't like the matplotlib solution b/c of the "dll-hell" anti-pattern. But the pure numpy solution looks like it has too many copies. (or are recarrays pointers-of-pointers as opposed to contiguous memory?) Help! W

3 3

First shot at svn->git conversion
by David Cournapeau 28 Jul '10

28 Jul '10

Hi there, I have finally prepared and uploaded a test repository containing numpy code: http://github.com/numpy/numpy_svn Please check in particular that your account information are correctly recorded in the log. A few informations: - the repo was converted from a local svn mirror to git through svn-all-fast-export - svn tags are put into the svntags "namespace", and only maintenance branches have been kept - "out of branch" branches (f2py, vendor, etc...) have not been included. They become separate git repos (not uploaded). DO NOT COMMIT ANYTHING TO THIS REPO - this is only for testing, any change made to this repo WILL be lost, cheers, David

4 8

Don't understand this error
by David Goldsmith 27 Jul '10

27 Jul '10

res = np.fromfunction(make_res, (nx, ny)) File "C:\Python26\lib\site-packages\numpy\core\numeric.py", line 1538, in fromfunction args = indices(shape, dtype=dtype) File "C:\Python26\lib\site-packages\numpy\core\numeric.py", line 1480, in indices tmp.shape = (1,)*i + (dim,)+(1,)*(N-i-1) ValueError: total size of new array must be unchanged Script terminated If it's a new array, how can it already have a size that can't be changed? What does this error really mean? DG

2 2

subtract.reduce behavior
by Johann Hibschman 27 Jul '10

27 Jul '10

I'm trying to understand numpy.subtract.reduce. The documentation doesn't seem to match the behavior. The documentation claims For a one-dimensional array, reduce produces results equivalent to: r = op.identity for i in xrange(len(A)): r = op(r,A[i]) return r However, numpy.subtract.reduce([1,2,3]) gives me 1-2-3==-4, not 0-1-2-3==-6. Now, I'm on an older version (1.3.0), which might be the problem, but which is "correct" here, the code or the docs? Thanks, Johann

5 10