Mailman 3 February 2013 - NumPy-Discussion

example reading binary Fortran file
by David Froger 13 Feb '13

13 Feb '13

Hy, My question is about reading Fortran binary file (oh no this question again...) Until now, I was using the unpack module like that : def lread(f,fourBeginning,fourEnd,*tuple): from struct import unpack """Reading a Fortran binary file in litte-endian""" if fourBeginning: f.seek(4,1) for array in tuple: for elt in xrange(array.size): transpose(array).flat[elt] = unpack(array.dtype.char,f.read(array.itemsize))[0] if fourEnd: f.seek(4,1) After googling, I read that fopen and npfille was deprecated, and we should use numpy.fromfile and ndarray.tofile, but despite of the documentaion, the cookbook, the mailling list and google, I don't succed in making a simple example. Considering the simple Fortran code below what is the Python script to read the four arrrays? What about if my pc is little endian and the file big endian? I think it will be a good idea to put the Fortran writting-arrays code and the Python reading-array script in the cookbook and maybe a page to help people comming from Fortran to start with Python ? Best, David Froger program makeArray implicit none integer,parameter:: nx=10,ny=20 real(4),dimension(nx,ny):: ux,uy,p integer :: i,j open(11,file='uxuyp.bin',form='unformatted') do i = 1,nx do j = 1,ny ux(i,j) = real(i*j) uy(i,j) = real(i)/real(j) p (i,j) = real(i) + real(j) enddo enddo write(11) ux,uy write(11) p close(11) end program makeArray

8 22

View on sliced array without copy
by Nicolas Rougier 13 Feb '13

13 Feb '13

Hi, I'm trying to get a view on a sliced array without copy but I'm not able to do it as illustrated below: dtype = np.dtype( [('color', 'f4', 4)] ) Z = np.zeros(100, dtype=dtype) print Z.view('f4').base is Z # True print Z[:50].base is Z # True print (Z.view('f4'))[:50].base is Z # False print Z[:50].view('f4').base is Z # False Did I do something wrong or is it expected behavior ? Nicolas

2 2

Help with python in C code
by santhu kumar 13 Feb '13

13 Feb '13

Hello all, I was able to successfully embed python code into C code. The basic skeleton is something like this : In C code : PyObject *pName, *pModule, *pFunc,*pArgs,*pReturn; PyArrayObject *cusForce; Py_Initialize(); import_array(); // required while using numpy arrays in C // Call some python code // Py_Decref all the used values Py_Finalize() Py_Finalize caused segmentation faults when numpy was one of import modules [node1:05762] Failing at address: 0x1 [node1:05762] [ 0] /lib64/libc.so.6 [0x3afa630280] [node1:05762] [ 1] /usr/lib64/python2.4/site-packages/numpy/core/multiarray.so [0x2b4e950aaf37] [node1:05762] [ 2] /usr/lib64/python2.4/site-packages/numpy/core/multiarray.so [0x2b4e950dd1d0] [node1:05762] [ 3] /usr/lib64/python2.4/site-packages/numpy/core/multiarray.so [0x2b4e950e22fa] [node1:05762] [ 4] /usr/lib64/libpython2.4.so.1.0(PyObject_IsTrue+0x37) [0x2b4e8a683c77] Which could be avoided by not calling Py_Finalize. But without calling Py_Finalize(), I am having huge memory leaks as the process keeps on increasing its memory usage as I see in top. Is there any clean way of calling Py_Finalize() when using numpy. Thanks santhosh

1 0

numpy.test('full') errors and failures
by Dominic Steinitz 13 Feb '13

13 Feb '13

Apologies if this is the wrong mailing list. I have installed bumpy using the excellent script here: http://fonnesbeck.github.com/ScipySuperpack/ I ran numpy.test('full') and got several errors. Should I be worried? Thanks, Dominic. PS I can send the full errors if that would be helpful. >>> np.test('full') Running unit tests for numpy NumPy version 1.8.0.dev-4600b2f NumPy is installed in /Library/Python/2.7/site-packages/numpy-1.8.0.dev_4600b2f_20130131-py2.7-macosx-10.8-intel.egg/numpy Python version 2.7.3 (v2.7.3:70274d53c1dd, Apr 9 2012, 20:52:43) [GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] nose version 1.2.1 … Ran 4471 tests in 13.065s FAILED (KNOWNFAIL=5, SKIP=26, errors=2, failures=3) <nose.result.TextTestResult run=4471 errors=2 failures=3>

3 2

number of tests
by Davide Del Vento 13 Feb '13

13 Feb '13

I compiled numpy 1.6.2 (right before 1.7 came out) with the intel compiler and MKL library. I'm trying to assess whether or not everything has been build fine. Since my machine is actually a cluster, I'm running the tests in different configurations (login node and batch script). However, I'm confused by the number of tests which ran. On the login nodes (either interactively or without tty) I get: Ran 3587 tests in 22.211s FAILED (KNOWNFAIL=5, SKIP=11, failures=2) Whereas in a remote batch node (with a script) I get: Ran 3229 tests in 15.642s OK (KNOWNFAIL=5, SKIP=19) Where did the 358 "missing" tests go in the batch run? The handful difference in SKIPped and FAILed (which I am investigating) cannot be the reason. What is it happening? PS: a similar thing happened with scipy, which I'm asking on the scipy mailing list. Thanks and Regards, Davide

2 2

pip install numpy throwing a lot of output.
by Daπid 12 Feb '13

12 Feb '13

I have just upgraded numpy with pip on Linux 64 bits with Python 2.7, and I got *a lot* of output, so much it doesn't fit in the terminal. Most of it are gcc commands, but there are many different errors thrown by the compiler. Is this expected? I am not too worried as the test suite passes, but pip is supposed to give only meaningful output (or at least, this is what the creators intended). David.

2 3

nditer gurus: is there a more efficient way to do this?
by Dave Hirschfeld 10 Feb '13

10 Feb '13

I have two NxMx3 arrays and I want to reduce over the last dimension of the first array by selecting those elements corresponding to the index of the maximum value of each 3-vector of the second array to give an NxM result. Hopefully that makes sense? If not hopefully the example below will shed some light. Can anyone think of a more efficient way to do this than my 1st attempt? I thought that nditer might be the solution; I got it to work (mostly by trial & error) but found that it's ~50x slower for this task! Is this not a good usecase for nditer or am I doing something wrong? In [42]: value_array = np.outer(ones(1000), arange(3)).reshape(20,50,3) ...: index_array = randn(*value_array.shape) In [43]: indices = index_array.reshape(-1,3).argmax(axis=1) ...: result = value_array.reshape(-1,3)[np.arange(indices.size), indices] ...: result = result.reshape(value_array.shape[0:-1]) In [44]: it = np.nditer([value_array, index_array, None], ...: flags=['reduce_ok', 'external_loop','buffered', 'delay_bufalloc'], ...: op_flags=[['readonly'],['readonly'],['readwrite', 'allocate']], ...: op_axes=[[0,1,2], [0,1,2], [0,1,-1]]) ...: it.reset() ...: for values, index_values, out in it: ...: out[...] = values[index_values.argmax()] ...: # In [45]: allclose(result, it.operands[2]) Out[45]: True In [46]: %%timeit ...: indices = index_array.reshape(-1,3).argmax(axis=1) ...: result = value_array.reshape(-1,3)[np.arange(indices.size), indices] ...: result = result.reshape(value_array.shape[0:-1]) ...: 10000 loops, best of 3: 113 µs per loop In [47]: %%timeit ...: it = np.nditer([value_array, index_array, None], ...: flags=['reduce_ok', 'external_loop','buffered', 'delay_bufalloc'], ...: op_flags=[['readonly'],['readonly'],['readwrite', 'allocate']], ...: op_axes=[[0,1,2], [0,1,2], [0,1,-1]]) ...: it.reset() ...: for values, index_values, out in it: ...: out[...] = values[index_values.argmax()] ...: # ...: 100 loops, best of 3: 5.26 ms per loop In [48]: Thanks, Dave

1 0

ANN: NumPy 1.7.0 release
by Ondřej Čertík 10 Feb '13

10 Feb '13

Hi, I'm pleased to announce the availability of the final release of NumPy 1.7.0. Sources and binary installers can be found at https://sourceforge.net/projects/numpy/files/NumPy/1.7.0/ This release is equivalent to the 1.7.0rc2 release, since no more problems were found. For release notes see below. I would like to thank everybody who contributed to this release. Cheers, Ondrej ========================= NumPy 1.7.0 Release Notes ========================= This release includes several new features as well as numerous bug fixes and refactorings. It supports Python 2.4 - 2.7 and 3.1 - 3.3 and is the last release that supports Python 2.4 - 2.5. Highlights ========== * ``where=`` parameter to ufuncs (allows the use of boolean arrays to choose where a computation should be done) * ``vectorize`` improvements (added 'excluded' and 'cache' keyword, general cleanup and bug fixes) * ``numpy.random.choice`` (random sample generating function) Compatibility notes =================== In a future version of numpy, the functions np.diag, np.diagonal, and the diagonal method of ndarrays will return a view onto the original array, instead of producing a copy as they do now. This makes a difference if you write to the array returned by any of these functions. To facilitate this transition, numpy 1.7 produces a FutureWarning if it detects that you may be attempting to write to such an array. See the documentation for np.diagonal for details. Similar to np.diagonal above, in a future version of numpy, indexing a record array by a list of field names will return a view onto the original array, instead of producing a copy as they do now. As with np.diagonal, numpy 1.7 produces a FutureWarning if it detects that you may be attempting to write to such an array. See the documentation for array indexing for details. In a future version of numpy, the default casting rule for UFunc out= parameters will be changed from 'unsafe' to 'same_kind'. (This also applies to in-place operations like a += b, which is equivalent to np.add(a, b, out=a).) Most usages which violate the 'same_kind' rule are likely bugs, so this change may expose previously undetected errors in projects that depend on NumPy. In this version of numpy, such usages will continue to succeed, but will raise a DeprecationWarning. Full-array boolean indexing has been optimized to use a different, optimized code path. This code path should produce the same results, but any feedback about changes to your code would be appreciated. Attempting to write to a read-only array (one with ``arr.flags.writeable`` set to ``False``) used to raise either a RuntimeError, ValueError, or TypeError inconsistently, depending on which code path was taken. It now consistently raises a ValueError. The <ufunc>.reduce functions evaluate some reductions in a different order than in previous versions of NumPy, generally providing higher performance. Because of the nature of floating-point arithmetic, this may subtly change some results, just as linking NumPy to a different BLAS implementations such as MKL can. If upgrading from 1.5, then generally in 1.6 and 1.7 there have been substantial code added and some code paths altered, particularly in the areas of type resolution and buffered iteration over universal functions. This might have an impact on your code particularly if you relied on accidental behavior in the past. New features ============ Reduction UFuncs Generalize axis= Parameter ------------------------------------------- Any ufunc.reduce function call, as well as other reductions like sum, prod, any, all, max and min support the ability to choose a subset of the axes to reduce over. Previously, one could say axis=None to mean all the axes or axis=# to pick a single axis. Now, one can also say axis=(#,#) to pick a list of axes for reduction. Reduction UFuncs New keepdims= Parameter ---------------------------------------- There is a new keepdims= parameter, which if set to True, doesn't throw away the reduction axes but instead sets them to have size one. When this option is set, the reduction result will broadcast correctly to the original operand which was reduced. Datetime support ---------------- .. note:: The datetime API is *experimental* in 1.7.0, and may undergo changes in future versions of NumPy. There have been a lot of fixes and enhancements to datetime64 compared to NumPy 1.6: * the parser is quite strict about only accepting ISO 8601 dates, with a few convenience extensions * converts between units correctly * datetime arithmetic works correctly * business day functionality (allows the datetime to be used in contexts where only certain days of the week are valid) The notes in `doc/source/reference/arrays.datetime.rst <https://github.com/numpy/numpy/blob/maintenance/1.7.x/doc/source/reference/…>`_ (also available in the online docs at `arrays.datetime.html <http://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_) should be consulted for more details. Custom formatter for printing arrays ------------------------------------ See the new ``formatter`` parameter of the ``numpy.set_printoptions`` function. New function numpy.random.choice --------------------------------- A generic sampling function has been added which will generate samples from a given array-like. The samples can be with or without replacement, and with uniform or given non-uniform probabilities. New function isclose -------------------- Returns a boolean array where two arrays are element-wise equal within a tolerance. Both relative and absolute tolerance can be specified. Preliminary multi-dimensional support in the polynomial package --------------------------------------------------------------- Axis keywords have been added to the integration and differentiation functions and a tensor keyword was added to the evaluation functions. These additions allow multi-dimensional coefficient arrays to be used in those functions. New functions for evaluating 2-D and 3-D coefficient arrays on grids or sets of points were added together with 2-D and 3-D pseudo-Vandermonde matrices that can be used for fitting. Ability to pad rank-n arrays ---------------------------- A pad module containing functions for padding n-dimensional arrays has been added. The various private padding functions are exposed as options to a public 'pad' function. Example:: pad(a, 5, mode='mean') Current modes are ``constant``, ``edge``, ``linear_ramp``, ``maximum``, ``mean``, ``median``, ``minimum``, ``reflect``, ``symmetric``, ``wrap``, and ``<function>``. New argument to searchsorted ---------------------------- The function searchsorted now accepts a 'sorter' argument that is a permutation array that sorts the array to search. Build system ------------ Added experimental support for the AArch64 architecture. C API ----- New function ``PyArray_RequireWriteable`` provides a consistent interface for checking array writeability -- any C code which works with arrays whose WRITEABLE flag is not known to be True a priori, should make sure to call this function before writing. NumPy C Style Guide added (``doc/C_STYLE_GUIDE.rst.txt``). Changes ======= General ------- The function np.concatenate tries to match the layout of its input arrays. Previously, the layout did not follow any particular reason, and depended in an undesirable way on the particular axis chosen for concatenation. A bug was also fixed which silently allowed out of bounds axis arguments. The ufuncs logical_or, logical_and, and logical_not now follow Python's behavior with object arrays, instead of trying to call methods on the objects. For example the expression (3 and 'test') produces the string 'test', and now np.logical_and(np.array(3, 'O'), np.array('test', 'O')) produces 'test' as well. The ``.base`` attribute on ndarrays, which is used on views to ensure that the underlying array owning the memory is not deallocated prematurely, now collapses out references when you have a view-of-a-view. For example:: a = np.arange(10) b = a[1:] c = b[1:] In numpy 1.6, ``c.base`` is ``b``, and ``c.base.base`` is ``a``. In numpy 1.7, ``c.base`` is ``a``. To increase backwards compatibility for software which relies on the old behaviour of ``.base``, we only 'skip over' objects which have exactly the same type as the newly created view. This makes a difference if you use ``ndarray`` subclasses. For example, if we have a mix of ``ndarray`` and ``matrix`` objects which are all views on the same original ``ndarray``:: a = np.arange(10) b = np.asmatrix(a) c = b[0, 1:] d = c[0, 1:] then ``d.base`` will be ``b``. This is because ``d`` is a ``matrix`` object, and so the collapsing process only continues so long as it encounters other ``matrix`` objects. It considers ``c``, ``b``, and ``a`` in that order, and ``b`` is the last entry in that list which is a ``matrix`` object. Casting Rules ------------- Casting rules have undergone some changes in corner cases, due to the NA-related work. In particular for combinations of scalar+scalar: * the `longlong` type (`q`) now stays `longlong` for operations with any other number (`? b h i l q p B H I`), previously it was cast as `int_` (`l`). The `ulonglong` type (`Q`) now stays as `ulonglong` instead of `uint` (`L`). * the `timedelta64` type (`m`) can now be mixed with any integer type (`b h i l q p B H I L Q P`), previously it raised `TypeError`. For array + scalar, the above rules just broadcast except the case when the array and scalars are unsigned/signed integers, then the result gets converted to the array type (of possibly larger size) as illustrated by the following examples:: >>> (np.zeros((2,), dtype=np.uint8) + np.int16(257)).dtype dtype('uint16') >>> (np.zeros((2,), dtype=np.int8) + np.uint16(257)).dtype dtype('int16') >>> (np.zeros((2,), dtype=np.int16) + np.uint32(2**17)).dtype dtype('int32') Whether the size gets increased depends on the size of the scalar, for example:: >>> (np.zeros((2,), dtype=np.uint8) + np.int16(255)).dtype dtype('uint8') >>> (np.zeros((2,), dtype=np.uint8) + np.int16(256)).dtype dtype('uint16') Also a ``complex128`` scalar + ``float32`` array is cast to ``complex64``. In NumPy 1.7 the `datetime64` type (`M`) must be constructed by explicitly specifying the type as the second argument (e.g. ``np.datetime64(2000, 'Y')``). Deprecations ============ General ------- Specifying a custom string formatter with a `_format` array attribute is deprecated. The new `formatter` keyword in ``numpy.set_printoptions`` or ``numpy.array2string`` can be used instead. The deprecated imports in the polynomial package have been removed. ``concatenate`` now raises DepractionWarning for 1D arrays if ``axis != 0``. Versions of numpy < 1.7.0 ignored axis argument value for 1D arrays. We allow this for now, but in due course we will raise an error. C-API ----- Direct access to the fields of PyArrayObject* has been deprecated. Direct access has been recommended against for many releases. Expect similar deprecations for PyArray_Descr* and other core objects in the future as preparation for NumPy 2.0. The macros in old_defines.h are deprecated and will be removed in the next major release (>= 2.0). The sed script tools/replace_old_macros.sed can be used to replace these macros with the newer versions. You can test your code against the deprecated C API by #defining NPY_NO_DEPRECATED_API to the target version number, for example NPY_1_7_API_VERSION, before including any NumPy headers. The ``NPY_CHAR`` member of the ``NPY_TYPES`` enum is deprecated and will be removed in NumPy 1.8. See the discussion at `gh-2801 <https://github.com/numpy/numpy/issues/2801>`_ for more details. Checksums ========= 7b72cc17b6a9043f6d46af4e71cd3dbe release/installers/numpy-1.7.0-win32-superpack-python3.3.exe 4fa54e40b6a243416f0248123b6ec332 release/installers/numpy-1.7.0.tar.gz 9ef1688bb9f8deb058a8022b4788686c release/installers/numpy-1.7.0-win32-superpack-python2.7.exe 909fe47da05d2a35edd6909ba0152213 release/installers/numpy-1.7.0-win32-superpack-python3.2.exe 5d4318b722d0098f78b49c0030d47026 release/installers/numpy-1.7.0-win32-superpack-python2.6.exe 92b61d6f278a81cf9a5033b0c8e7b53e release/installers/numpy-1.7.0-win32-superpack-python3.1.exe 51d6f4f854cdca224fa56a327ad7c620 release/installers/numpy-1.7.0-win32-superpack-python2.5.exe ca27913c59393940e880fab420f985b4 release/installers/numpy-1.7.0.zip 3f20becbb80da09412d94815ad3b586b release/installers/numpy-1.7.0-py2.5-python.org-macosx10.3.dmg 600dfa4dab31db5dc2ed9655521cfa9e release/installers/numpy-1.7.0-py2.6-python.org-macosx10.3.dmg a907a37416163b3245a30cfd160506ab release/installers/numpy-1.7.0-py2.7-python.org-macosx10.3.dmg

5 5

Modern Fortran vs NumPy syntax
by Ondřej Čertík 10 Feb '13

10 Feb '13

Hi, I have recently setup a page about modern Fortran: http://fortran90.org/ and in particular, it has a long section with side by side syntax examples of Python/NumPy vs Fortran: http://fortran90.org/src/rosetta.html I would be very interested if some NumPy gurus would provide me feedback. I personally knew NumPy long before I learned Fortran, and I was amazed that the modern Fortran pretty much allows 1:1 syntax with NumPy, including most of all the fancy indexing etc. Is there some NumPy feature that is not covered there? I would like it to be a nice resource for people who know NumPy to feel like at home with Fortran, and vice versa. I personally use both every day (Fortran a bit more than NumPy). Or of you have any other comments or tips for the site, please let me know. Eventually I'd like to also put there C++ way of doing the same things, but at the moment I want to get Fortran and Python/NumPy done first. Ondrej

5 5

ANN: NumPy 1.7.0rc2 release
by Ondřej Čertík 10 Feb '13

10 Feb '13

Hi, I'm pleased to announce the availability of the second release candidate of NumPy 1.7.0rc2. Sources and binary installers can be found at https://sourceforge.net/projects/numpy/files/NumPy/1.7.0rc2/ We have fixed all issues known to us since the 1.7.0rc1 release. Please test this release and report any issues on the numpy-discussion mailing list. If there are no further problems, I plan to release the final version in a few days. I would like to thank Sandro Tosi, Sebastian Berg, Charles Harris, Marcin Juszkiewicz, Mark Wiebe, Ralf Gommers and Nathaniel J. Smith for sending patches, fixes and helping with reviews for this release since 1.7.0rc1, and Vincent Davis for providing the Mac build machine. Cheers, Ondrej ========================= NumPy 1.7.0 Release Notes ========================= This release includes several new features as well as numerous bug fixes and refactorings. It supports Python 2.4 - 2.7 and 3.1 - 3.3 and is the last release that supports Python 2.4 - 2.5. Highlights ========== * ``where=`` parameter to ufuncs (allows the use of boolean arrays to choose where a computation should be done) * ``vectorize`` improvements (added 'excluded' and 'cache' keyword, general cleanup and bug fixes) * ``numpy.random.choice`` (random sample generating function) Compatibility notes =================== In a future version of numpy, the functions np.diag, np.diagonal, and the diagonal method of ndarrays will return a view onto the original array, instead of producing a copy as they do now. This makes a difference if you write to the array returned by any of these functions. To facilitate this transition, numpy 1.7 produces a FutureWarning if it detects that you may be attempting to write to such an array. See the documentation for np.diagonal for details. Similar to np.diagonal above, in a future version of numpy, indexing a record array by a list of field names will return a view onto the original array, instead of producing a copy as they do now. As with np.diagonal, numpy 1.7 produces a FutureWarning if it detects that you may be attempting to write to such an array. See the documentation for array indexing for details. In a future version of numpy, the default casting rule for UFunc out= parameters will be changed from 'unsafe' to 'same_kind'. (This also applies to in-place operations like a += b, which is equivalent to np.add(a, b, out=a).) Most usages which violate the 'same_kind' rule are likely bugs, so this change may expose previously undetected errors in projects that depend on NumPy. In this version of numpy, such usages will continue to succeed, but will raise a DeprecationWarning. Full-array boolean indexing has been optimized to use a different, optimized code path. This code path should produce the same results, but any feedback about changes to your code would be appreciated. Attempting to write to a read-only array (one with ``arr.flags.writeable`` set to ``False``) used to raise either a RuntimeError, ValueError, or TypeError inconsistently, depending on which code path was taken. It now consistently raises a ValueError. The <ufunc>.reduce functions evaluate some reductions in a different order than in previous versions of NumPy, generally providing higher performance. Because of the nature of floating-point arithmetic, this may subtly change some results, just as linking NumPy to a different BLAS implementations such as MKL can. If upgrading from 1.5, then generally in 1.6 and 1.7 there have been substantial code added and some code paths altered, particularly in the areas of type resolution and buffered iteration over universal functions. This might have an impact on your code particularly if you relied on accidental behavior in the past. New features ============ Reduction UFuncs Generalize axis= Parameter ------------------------------------------- Any ufunc.reduce function call, as well as other reductions like sum, prod, any, all, max and min support the ability to choose a subset of the axes to reduce over. Previously, one could say axis=None to mean all the axes or axis=# to pick a single axis. Now, one can also say axis=(#,#) to pick a list of axes for reduction. Reduction UFuncs New keepdims= Parameter ---------------------------------------- There is a new keepdims= parameter, which if set to True, doesn't throw away the reduction axes but instead sets them to have size one. When this option is set, the reduction result will broadcast correctly to the original operand which was reduced. Datetime support ---------------- .. note:: The datetime API is *experimental* in 1.7.0, and may undergo changes in future versions of NumPy. There have been a lot of fixes and enhancements to datetime64 compared to NumPy 1.6: * the parser is quite strict about only accepting ISO 8601 dates, with a few convenience extensions * converts between units correctly * datetime arithmetic works correctly * business day functionality (allows the datetime to be used in contexts where only certain days of the week are valid) The notes in `doc/source/reference/arrays.datetime.rst <https://github.com/numpy/numpy/blob/maintenance/1.7.x/doc/source/reference/…>`_ (also available in the online docs at `arrays.datetime.html <http://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_) should be consulted for more details. Custom formatter for printing arrays ------------------------------------ See the new ``formatter`` parameter of the ``numpy.set_printoptions`` function. New function numpy.random.choice --------------------------------- A generic sampling function has been added which will generate samples from a given array-like. The samples can be with or without replacement, and with uniform or given non-uniform probabilities. New function isclose -------------------- Returns a boolean array where two arrays are element-wise equal within a tolerance. Both relative and absolute tolerance can be specified. Preliminary multi-dimensional support in the polynomial package --------------------------------------------------------------- Axis keywords have been added to the integration and differentiation functions and a tensor keyword was added to the evaluation functions. These additions allow multi-dimensional coefficient arrays to be used in those functions. New functions for evaluating 2-D and 3-D coefficient arrays on grids or sets of points were added together with 2-D and 3-D pseudo-Vandermonde matrices that can be used for fitting. Ability to pad rank-n arrays ---------------------------- A pad module containing functions for padding n-dimensional arrays has been added. The various private padding functions are exposed as options to a public 'pad' function. Example:: pad(a, 5, mode='mean') Current modes are ``constant``, ``edge``, ``linear_ramp``, ``maximum``, ``mean``, ``median``, ``minimum``, ``reflect``, ``symmetric``, ``wrap``, and ``<function>``. New argument to searchsorted ---------------------------- The function searchsorted now accepts a 'sorter' argument that is a permutation array that sorts the array to search. Build system ------------ Added experimental support for the AArch64 architecture. C API ----- New function ``PyArray_RequireWriteable`` provides a consistent interface for checking array writeability -- any C code which works with arrays whose WRITEABLE flag is not known to be True a priori, should make sure to call this function before writing. NumPy C Style Guide added (``doc/C_STYLE_GUIDE.rst.txt``). Changes ======= General ------- The function np.concatenate tries to match the layout of its input arrays. Previously, the layout did not follow any particular reason, and depended in an undesirable way on the particular axis chosen for concatenation. A bug was also fixed which silently allowed out of bounds axis arguments. The ufuncs logical_or, logical_and, and logical_not now follow Python's behavior with object arrays, instead of trying to call methods on the objects. For example the expression (3 and 'test') produces the string 'test', and now np.logical_and(np.array(3, 'O'), np.array('test', 'O')) produces 'test' as well. The ``.base`` attribute on ndarrays, which is used on views to ensure that the underlying array owning the memory is not deallocated prematurely, now collapses out references when you have a view-of-a-view. For example:: a = np.arange(10) b = a[1:] c = b[1:] In numpy 1.6, ``c.base`` is ``b``, and ``c.base.base`` is ``a``. In numpy 1.7, ``c.base`` is ``a``. To increase backwards compatibility for software which relies on the old behaviour of ``.base``, we only 'skip over' objects which have exactly the same type as the newly created view. This makes a difference if you use ``ndarray`` subclasses. For example, if we have a mix of ``ndarray`` and ``matrix`` objects which are all views on the same original ``ndarray``:: a = np.arange(10) b = np.asmatrix(a) c = b[0, 1:] d = c[0, 1:] then ``d.base`` will be ``b``. This is because ``d`` is a ``matrix`` object, and so the collapsing process only continues so long as it encounters other ``matrix`` objects. It considers ``c``, ``b``, and ``a`` in that order, and ``b`` is the last entry in that list which is a ``matrix`` object. Casting Rules ------------- Casting rules have undergone some changes in corner cases, due to the NA-related work. In particular for combinations of scalar+scalar: * the `longlong` type (`q`) now stays `longlong` for operations with any other number (`? b h i l q p B H I`), previously it was cast as `int_` (`l`). The `ulonglong` type (`Q`) now stays as `ulonglong` instead of `uint` (`L`). * the `timedelta64` type (`m`) can now be mixed with any integer type (`b h i l q p B H I L Q P`), previously it raised `TypeError`. For array + scalar, the above rules just broadcast except the case when the array and scalars are unsigned/signed integers, then the result gets converted to the array type (of possibly larger size) as illustrated by the following examples:: >>> (np.zeros((2,), dtype=np.uint8) + np.int16(257)).dtype dtype('uint16') >>> (np.zeros((2,), dtype=np.int8) + np.uint16(257)).dtype dtype('int16') >>> (np.zeros((2,), dtype=np.int16) + np.uint32(2**17)).dtype dtype('int32') Whether the size gets increased depends on the size of the scalar, for example:: >>> (np.zeros((2,), dtype=np.uint8) + np.int16(255)).dtype dtype('uint8') >>> (np.zeros((2,), dtype=np.uint8) + np.int16(256)).dtype dtype('uint16') Also a ``complex128`` scalar + ``float32`` array is cast to ``complex64``. In NumPy 1.7 the `datetime64` type (`M`) must be constructed by explicitly specifying the type as the second argument (e.g. ``np.datetime64(2000, 'Y')``). Deprecations ============ General ------- Specifying a custom string formatter with a `_format` array attribute is deprecated. The new `formatter` keyword in ``numpy.set_printoptions`` or ``numpy.array2string`` can be used instead. The deprecated imports in the polynomial package have been removed. ``concatenate`` now raises DepractionWarning for 1D arrays if ``axis != 0``. Versions of numpy < 1.7.0 ignored axis argument value for 1D arrays. We allow this for now, but in due course we will raise an error. C-API ----- Direct access to the fields of PyArrayObject* has been deprecated. Direct access has been recommended against for many releases. Expect similar deprecations for PyArray_Descr* and other core objects in the future as preparation for NumPy 2.0. The macros in old_defines.h are deprecated and will be removed in the next major release (>= 2.0). The sed script tools/replace_old_macros.sed can be used to replace these macros with the newer versions. You can test your code against the deprecated C API by #defining NPY_NO_DEPRECATED_API to the target version number, for example NPY_1_7_API_VERSION, before including any NumPy headers. The ``NPY_CHAR`` member of the ``NPY_TYPES`` enum is deprecated and will be removed in NumPy 1.8. See the discussion at `gh-2801 <https://github.com/numpy/numpy/issues/2801>`_ for more details.

4 4

NumPy-Discussion February 2013