Mailman 3 January 2009 - NumPy-Discussion

coding style: citations
by Alan G Isaac 23 Jan '09

23 Jan '09

The docstring standard at http://projects.scipy.org/scipy/numpy/wiki/CodingStyleGuidelines#docstring-… suggests a citation reference format that is not compatible with reStructuredText. Quoting from http://docutils.sourceforge.net/docs/ref/rst/restructuredtext.html#citations: Citations are identical to footnotes except that they use only non-numeric labels such as [note] or [GVR2001]. In other words, the standard currently promotes a conflation of notes and citations. Another reason for avoiding the style currently in the docstring standard is that it renders impossible the generation of citations using a citation database. (See e.g. the bibtext-like facilities offered by bibstuff's bib4txt (http://code.google.com/p/bibstuff/). Alan Isaac

4 13

ANN: Numexpr 1.1.1 available
by Francesc Alted 23 Jan '09

23 Jan '09

========================== Announcing Numexpr 1.1.1 ========================== Numexpr is a fast numerical expression evaluator for NumPy. With it, expressions that operate on arrays (like "3*a+4*b") are accelerated and use less memory than doing the same calculation in Python. This is a bug-fix release, where the next issues have been addressed: - The case for multidimensional array operands is properly accelerated now. Added a new benchmark (based on a script provided by Andrew Collette, thanks!) for easily testing this case in the future. Closes #12. - Added a fix to avoid the caches in numexpr to grow too much. The dictionary caches are kept now always with less than 256 entries. Closes #11. - The VERSION file is correctly copied now (support for it was not present in the 1.1 tar file, don't know exactly why). Closes #8. - The package is accessible now from PyPI, so that people from 'evil' countries can access it too. In case you want to know more in detail what has changed in this version, have a look at ``RELEASE_NOTES.txt`` in the tarball. Where I can find Numexpr? ========================= The project is hosted at Google code in: http://code.google.com/p/numexpr/ And you can get the packages from PyPI as well: http://pypi.python.org/pypi How it works? ============= See: http://code.google.com/p/numexpr/wiki/Overview for a detailed description by the original author of Numexpr (David M. Cooke). Share your experience ===================== Let us know of any bugs, suggestions, gripes, kudos, etc. you may have. Enjoy! -- Francesc Alted

1 0

python numpy code many times slower than c++
by Neal Becker 23 Jan '09

23 Jan '09

I tried a little experiment, implementing some code in numpy (usually I build modules in c++ to interface to python). Since these operations are all large vectors, I hoped it would be reasonably efficient. The code in question is simple. It is a model of an amplifier, modeled by it's AM/AM and AM/PM characteristics. The function in question is the __call__ operator. The test program plots a spectrum, calling this operator 1024 times each time with a vector of 4096. Any ideas? The code is not too big, so I'll try to attach it.

8 31

Re: [Numpy-discussion] Numpy unexpected (for me) behaviour
by V. Armando Sole 23 Jan '09

23 Jan '09

At 01:44 23/01/2009 -0600, Robert Kern wrote: >It is an inevitable consequence of several features interacting >together. Basically, Python expands "a[b] += 1" into this: > > c = a[b] > d = c.__iadd__(1) > a[b] = d > >Basically, the array c doesn't know that it was created by indexing a, >so it can't do the accumulation you want. Well inevitable would not be the word I would use. I would have expected a different behaviour between a[b] = a[b] + 1 and a[b] += 1 In the second case python (or numpy) does not need to generate an intermediate array and could be doing in-place operations. > > Is there a way I can achieve the first result > > without a for loop? In my application the difference is a factor 10 in > > execution time (1000 secons instead of 100 ...) > >In [6]: bincount? >Type: builtin_function_or_method >Base Class: <type 'builtin_function_or_method'> >String Form: <built-in function bincount> >Namespace: Interactive >Docstring: > bincount(x,weights=None) > > Return the number of occurrences of each value in x. > > x must be a list of non-negative integers. The output, b[i], > represents the number of times that i is found in x. If weights > is specified, every occurrence of i at a position p contributes > weights[p] instead of 1. > > See also: histogram, digitize, unique. Indeed what I am doing is very close to histogramming. Unfortunaly what I have to add is not just one but a value. I have a set of scattered points (x,y,z) and values corresponding to those points. My goal is to get a regular grid in which in each voxel I sum the values of the points falling in them. I guess I will have to write a tiny C extension. I had expected the += sintax would trigger the in-place operation. Best regards, Armando

2 1

Numpy unexpected (for me) behaviour
by V. Armando Sole 23 Jan '09

23 Jan '09

Hello, In an effort to suppress for loops, I have arrived to the following situation. Through vectorial logical operations I generate a set of indices for which the contents of an array have to be incremented. My problem can be reduced to the following: #This works import numpy a=numpy.zeros(10) b=numpy.ones(4, numpy.int) for i in b: a[i] += 1 #a[1] contains 4 at the end #This does not work import numpy a=numpy.zeros(10) b=numpy.ones(4, numpy.int) a[b] += 1 #a[1] contains 1 at the end Is that a bug or a feature? Is there a way I can achieve the first result without a for loop? In my application the difference is a factor 10 in execution time (1000 secons instead of 100 ...) Thanks, Armando

3 4

numpy.array and subok kwarg
by Darren Dale 23 Jan '09

23 Jan '09

Hello, I have a test script: import numpy as np class MyArray(np.ndarray): __array_priority__ = 20 def __new__(cls): return np.asarray(1).view(cls).copy() def __repr__(self): return 'my_array' __str__ = __repr__ def __mul__(self, other): return super(MyArray, self).__mul__(other) def __rmul__(self, other): return super(MyArray, self).__rmul__(other) mine = MyArray() print type(np.array(mine,dtype='f')) The type returned by np.array is ndarray, unless I specifically set subok=True, in which case I get a MyArray. The default value of subok is True, so I dont understand why I have to specify subok unless I want it to be False. Is my subclass missing something important? Thanks, Darren

3 2

range in numpy where
by Ross Williamson 23 Jan '09

23 Jan '09

Hi All I want to get out the index of values in an array. Normally WHERE works fine for one conditional statement but it does not work for two - i.e. a = array([0,1,2,3,4,5,6,7,8,9]) ind, = where(a > 5) Works fine but if I wanted: ind = where((a > 5) and (a<8)) Then it bugs out with the following message: "The truth value of an array with more than one message is ambiguous" Any ideas? Cheers Ross

2 1

have to use matlab: what is the equiv. to pdb.pm()
by Sebastian Haase 22 Jan '09

22 Jan '09

Hi, I have to debug a matlab program -- is there a way to do a "post mortem" debugging ? (I don't even know if this term even exists outside of python ;-) ((I want to know some local variable values and find out why our program crashes when "there are no particles found"...)) Thanks, Sebastian Haase PS: This is obviously somewhat OT

2 1

Failures in test_recfunctions.py
by Nils Wagner 22 Jan '09

22 Jan '09

>>> numpy.__version__ '1.3.0.dev6331' ====================================================================== FAIL: Test the ignoremask option of find_duplicates ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/lib/tests/test_recfunctions.py", line 186, in test_find_duplicates_ignoremask assert_equal(test[-1], control) File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/ma/testutils.py", line 121, in assert_equal return assert_array_equal(actual, desired, err_msg) File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/ma/testutils.py", line 193, in assert_array_equal header='Arrays are not equal') File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/ma/testutils.py", line 186, in assert_array_compare verbose=verbose, header=header) File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/testing/utils.py", line 295, in assert_array_compare raise AssertionError(msg) AssertionError: Arrays are not equal (mismatch 33.3333333333%) x: array([0, 1, 3, 4, 2, 6]) y: array([0, 1, 3, 4, 6, 2]) ====================================================================== FAIL: Test zip_descr ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/lib/tests/test_recfunctions.py", line 34, in test_zip_descr np.dtype([('', '<i4'), ('', '<i4')])) File "/home/nwagner/local/lib64/python2.6/site-packages/numpy/ma/testutils.py", line 98, in assert_equal raise AssertionError(msg) AssertionError: Items are not equal: ACTUAL: [('f0', '<i8'), ('f1', '<i8')] DESIRED: dtype([('f0', '<i4'), ('f1', '<i4')]) ---------------------------------------------------------------------- Ran 1881 tests in 15.223s FAILED (KNOWNFAIL=9, failures=2) <nose.result.TextTestResult run=1881 errors=0 failures=2>

2 3

Tutorial on working with Excel files in Python (without COM and cross platform!) at PyConUS 2009
by Chris Withers 22 Jan '09

22 Jan '09

Hi All, Too many people in the Python community think the only way to work with Excel files in Python is using COM on Windows. To try and correct this, I'm giving a tutorial at this year's PyCon in Chicago on Wednesday, 25th March that will cover working with Excel files in Python using the pure-python libraries xlrd, xlwt and xlutils. I'll be looking to cover: - Reading Excel Files Including formatting, unicode dates and formulae. - Writing Excel Files Including formatting with easyxf and things like freeze pains, print areas, etc - Filtering Excel Files A run through on the structure of xlutils.filter and some examples to show you how it works. - Workshop for your problems I'm hoping anyone who attends will get a lot out of this! If you're planning on attending and have a particular problem you'd like to work on in this part of the tutorial, please drop me an email and I'll try and make sure I come prepared! All you need for the tutorial is a working knowledge of Excel and Python, with a laptop as an added benefit, and to be at PyCon this year: http://us.pycon.org I look forward to seeing you all there! Chris -- Simplistix - Content Management, Zope & Python Consulting - http://www.simplistix.co.uk

1 0