Re: [Numpy-discussion] very large matrices.
They are very large numbers indeed. Thanks for giving me a wake up call. Currently my data is represented as vectors in a vectorset, a typical sparse representation. I reduced the problem significantly by removing lots of noise. I'm basically recording traces of a terms occurrence throughout a corpus and doing an analysis of the eigenvectors. I reduced my matrix to 4863 x 4863 by filtering the original corpus. Now when I attempt svd, I'm finding a memory error in the svd routine. Is there a hard upper limit of the size of a matrix for these calculations? File /usr/lib/python2.4/site-packages/numpy/linalg/linalg.py, line 575, in svd vt = zeros((n, nvt), t) MemoryError Cheers Dave On 5/13/07, Anne Archibald [EMAIL PROTECTED] wrote: On 12/05/07, Dave P. Novakovic [EMAIL PROTECTED] wrote: core 2 duo with 4gb RAM. I've heard about iterative svd functions. I actually need a complete svd, with all eigenvalues (not LSI). I'm actually more interested in the individual eigenvectors. As an example, a single row could probably have about 3000 non zero elements. I think you need to think hard about whether your problem can be done in another way. First of all, the singular values (as returned from the svd) are not eigenvalues - eigenvalue decomposition is a much harder problem, numerically. Second, your full non-sparse matrix will be 8*75000*75000 bytes, or about 42 gibibytes. Put another way, the representation of your data alone is ten times the size of the RAM on the machine you're using. Third, your matrix has 225 000 000 nonzero entries; assuming a perfect sparse representation with no extra bytes (at least two bytes per entry is typical, usually more), that's 1.7 GiB. Recall that basically any matrix operation is at least O(N^3), so you can expect order 10^14 floating-point operations to be required. This is actually the *least* significant constraint; pushing stuff into and out of disk caches will be taking most of your time. Even if you can represent your matrix sparsely (using only a couple of gibibytes), you've said you want the full set of eigenvectors, which is not likely to be a sparse matrix - so your result is back up to 42 GiB. And you should expect an eigenvalue algorithm, if it even survives massive roundoff problems, to require something like that much working space; thus your problem probably has a working size of something like 84 GiB. SVD is a little easier, if that's what you want, but the full solution is twice as large, though if you discard entries corresponding to small values it might be quite reasonable. You'll still need some fairly specialized code, though. Which form are you looking for? Solving your problem in a reasonable amount of time, as described and on the hardware you specify, is going to require some very specialized algorithms; you could try looking for an out-of-core eigenvalue package, but I'd first look to see if there's any way you can simplify your problem - getting just one eigenvector, maybe. Anne ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] problems with calculating numpy.float64
Hi out there, this is the code segment if m maxN and n maxN and self.activeWide[m+1, n+1]: try: deltaX = x[m+1] - x[m] except TypeError: print '-' * 40 print type(x) type_a, type_b = map(type, (x[m + 1], x[m])) print type_a, type_b, type_a is type_b print '-' * 40 raise the if-conclusion is TRUE! I got an error at the line deltaX = x[m+1] - x[m], the values of these types: x : type 'numpy.ndarray' x[m] : type 'numpy.float64' x[m+1]: type 'numpy.float64' m : counting variable, i guess an integer? I just try to subtract x[m] from x[m+1], but I got an error: Inappropriate argument type. unsupported operand type(s) for -: 'numpy.float64' and 'numpy.float64' For more code snippets you can follow these URL: http://www.python-forum.de/topic-10580.html It is in german language, but there are more code snippets and some tests from me and other users. The problem is that both variables are definitely from the same type, but i cant substract theme. When I go to my python interpreter and reproduce the code there is no problem. I think the imports from numpy are OK, i cant see the error? Thanks for any help! Michael -- ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
[Numpy-discussion] howto make from flat array (1-dim) 2-dimensional?
i.e. for example from flat array [1, 2, 3] obtain array([[ 1.], [ 2.], [ 3.]]) I have numpy v 1.0.1 Thx, D. ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] howto make from flat array (1-dim) 2-dimensional?
On Sun, May 13, 2007 at 02:36:39PM +0300, dmitrey wrote: i.e. for example from flat array [1, 2, 3] obtain array([[ 1.], [ 2.], [ 3.]]) I have numpy v 1.0.1 Thx, D. Use newaxis: In [1]: a = array([1., 2., 3.]) In [2]: a Out[2]: array([ 1., 2., 3.]) In [3]: a[:,newaxis] Out[3]: array([[ 1.], [ 2.], [ 3.]]) In [4]: a[newaxis,:] Out[4]: array([[ 1., 2., 3.]]) When newaxis is used as an index, a new axis of dimension 1 is added. -- ||\/| /--\ |David M. Cooke http://arbutus.physics.mcmaster.ca/dmc/ |[EMAIL PROTECTED] ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] howto make from flat array (1-dim) 2-dimensional?
On Sunday 13 May 2007 7:36:39 am dmitrey wrote: i.e. for example from flat array [1, 2, 3] obtain array([[ 1.], [ 2.], [ 3.]]) a=array([1,2,3]) a.shape=(len(a),1) ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] NumPy 1.0.3 release next week
Hello all On Sat, 12 May 2007, Charles R Harris wrote: On 5/12/07, Albert Strasheim [EMAIL PROTECTED] wrote: I've more or less finished my quick triage effort. Thanks, Albert. The tickets look much better organized now. My pleasure. Stefan van der Walt has also gotten in on the act and we're now down to 19 open tickets with 1.0.3 as the milestone. http://projects.scipy.org/scipy/numpy/query?status=newstatus=assignedstatus=reopenedmilestone=1.0.3+Release Regards, Albert ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] howto make from flat array (1-dim) 2-dimensional?
On Sun, May 13, 2007 at 07:46:47AM -0400, Darren Dale wrote: On Sunday 13 May 2007 7:36:39 am dmitrey wrote: i.e. for example from flat array [1, 2, 3] obtain array([[ 1.], [ 2.], [ 3.]]) a=array([1,2,3]) a.shape=(len(a),1) Or just a.shape = (-1,1) Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] very large matrices.
On 5/13/07, Dave P. Novakovic [EMAIL PROTECTED] wrote: They are very large numbers indeed. Thanks for giving me a wake up call. Currently my data is represented as vectors in a vectorset, a typical sparse representation. I reduced the problem significantly by removing lots of noise. I'm basically recording traces of a terms occurrence throughout a corpus and doing an analysis of the eigenvectors. I reduced my matrix to 4863 x 4863 by filtering the original corpus. Now when I attempt svd, I'm finding a memory error in the svd routine. Is there a hard upper limit of the size of a matrix for these calculations? I get the same error here with linalg.svd(eye(5000)), and the memory is indeed gone. Hmm, I think it should work, although it is sure pushing the limits of what I've got: linalg.svd(eye(1000)) works fine. I think 4GB should be enough if your memory limits are set high enough. Are you trying some sort of principal components analysis? snip Chuck ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] NumPy 1.0.3 release next week
Is it possible somehow to speedup numpy 1.0.3 appearing in Linux update channels? (as for me I'm interested in Ubuntu/Kubuntu, currently there is v 1.0.1) I tried to compile numpy 1.0.2, but, as well as in Octave compiling, it failed because c compiler can't create executable. gcc reinstallation didn't help, other c compilers are absent in update channel (I had seen only tcc, but I'm sure it will not help, and it (I mean trying to install other C compilers) requires too much efforts). D. ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] NumPy 1.0.3 release next week
On Sun, May 13, 2007 at 06:19:30PM +0300, dmitrey wrote: Is it possible somehow to speedup numpy 1.0.3 appearing in Linux update channels? (as for me I'm interested in Ubuntu/Kubuntu, currently there is v 1.0.1) I tried to compile numpy 1.0.2, but, as well as in Octave compiling, it failed because c compiler can't create executable. gcc reinstallation didn't help, other c compilers are absent in update channel (I had seen only tcc, but I'm sure it will not help, and it (I mean trying to install other C compilers) requires too much efforts). Many people here are compiling numpy fine under Ubuntu. Do you have write permissions to the output directory? What is the compiler error given? Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
[Numpy-discussion] copy object with multiple subfields, including ndarrays
hi all, does anyone know howto copy an instance of class, that contains multiple subfields, for example myObj.field1.subfield2 = 'asdf' myObj.field4.subfield8 = numpy.mat('1 2 3; 4 5 6') I tried from copy import copy myObjCopy = copy(myObj) but it seems that it doesn't work correctly Thx, D. ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
[Numpy-discussion] dtype hashes are not equal
Hi all, In the numpy.sctypes dictionary, there are two entries for uint32: In [2]: N.sctypes['uint'] Out[2]: [type 'numpy.uint8', type 'numpy.uint16', type 'numpy.uint32', type 'numpy.uint32', type 'numpy.uint64'] Comparing the dtypes of the two types gives the correct answer: In [3]: sc = N.sctypes['uint'] In [4]: N.dtype(sc[2]) == N.dtype(sc[3]) Out[4]: True But the hash values for the dtypes (and the types) differ: In [42]: for T in N.sctypes['uint']: dt = N.dtype(T) print T, dt print '=', hash(T), hash(dt) type 'numpy.uint8' uint8 = -1217082432 -1217078592 type 'numpy.uint16' uint16 = -1217082240 -1217078464 type 'numpy.uint32' uint32 = -1217081856 -1217078336 type 'numpy.uint32' uint32 = -1217082048 -1217078400 type 'numpy.uint64' uint64 = -1217081664 -1217078208 Is this expected/correct behaviour? Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] dtype hashes are not equal
Stefan van der Walt wrote: Hi all, In the numpy.sctypes dictionary, there are two entries for uint32: In [2]: N.sctypes['uint'] Out[2]: [type 'numpy.uint8', type 'numpy.uint16', type 'numpy.uint32', type 'numpy.uint32', type 'numpy.uint64'] Comparing the dtypes of the two types gives the correct answer: In [3]: sc = N.sctypes['uint'] In [4]: N.dtype(sc[2]) == N.dtype(sc[3]) Out[4]: True But the hash values for the dtypes (and the types) differ: In [42]: for T in N.sctypes['uint']: dt = N.dtype(T) print T, dt print '=', hash(T), hash(dt) type 'numpy.uint8' uint8 = -1217082432 -1217078592 type 'numpy.uint16' uint16 = -1217082240 -1217078464 type 'numpy.uint32' uint32 = -1217081856 -1217078336 type 'numpy.uint32' uint32 = -1217082048 -1217078400 type 'numpy.uint64' uint64 = -1217081664 -1217078208 Is this expected/correct behaviour? It's expected, but not desired. We haven't implemented the hash function for dtype objects, so we have the default, which is based on object identity rather than value. It is something that should be implemented, given time. -- Robert Kern I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth. -- Umberto Eco ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] NumPy 1.0.3 release next week
Stefan van der Walt wrote: On Sun, May 13, 2007 at 06:19:30PM +0300, dmitrey wrote: Is it possible somehow to speedup numpy 1.0.3 appearing in Linux update channels? (as for me I'm interested in Ubuntu/Kubuntu, currently there is v 1.0.1) I tried to compile numpy 1.0.2, but, as well as in Octave compiling, it failed because c compiler can't create executable. gcc reinstallation didn't help, other c compilers are absent in update channel (I had seen only tcc, but I'm sure it will not help, and it (I mean trying to install other C compilers) requires too much efforts). Many people here are compiling numpy fine under Ubuntu. Do you have write permissions to the output directory? What is the compiler error given? Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion Sorry, I meant compiling Python2.5 and Octave, not numpy Octave Python2.5 is already present (in Ubuntu 7.04), but I tried to compile and install it from sources because numpy compilation failed with (I have gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4), compiling as root) ... ... C compiler: gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O2 -Wall -Wstrict-prototypes -fPIC compile options: '-I/usr/include/python2.5 -Inumpy/core/src -Inumpy/core/include -I/usr/include/python2.5 -c' gcc: _configtest.c In file included from /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/syslimits.h:7, from /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/limits.h:11, from /usr/include/python2.5/Python.h:18, from _configtest.c:2: /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/limits.h:122:61: error: limits.h: No such file or directory In file included from _configtest.c:2: /usr/include/python2.5/Python.h:32:19: error: stdio.h: No such file or directory /usr/include/python2.5/Python.h:34:5: error: #error Python.h requires that stdio.h define NULL. /usr/include/python2.5/Python.h:37:20: error: string.h: No such file or directory /usr/include/python2.5/Python.h:39:19: error: errno.h: No such file or directory /usr/include/python2.5/Python.h:41:20: error: stdlib.h: No such file or directory /usr/include/python2.5/Python.h:43:20: error: unistd.h: No such file or directory /usr/include/python2.5/Python.h:55:20: error: assert.h: No such file or directory In file included from /usr/include/python2.5/Python.h:57, from _configtest.c:2: /usr/include/python2.5/pyport.h:7:20: error: stdint.h: No such file or directory In file included from /usr/include/python2.5/Python.h:57, from _configtest.c:2: /usr/include/python2.5/pyport.h:73: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘Py_uintptr_t’ /usr/include/python2.5/pyport.h:74: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘Py_intptr_t’ /usr/include/python2.5/pyport.h:97: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘Py_ssize_t’ /usr/include/python2.5/pyport.h:204:76: error: math.h: No such file or directory /usr/include/python2.5/pyport.h:211:22: error: sys/time.h: No such file or directory /usr/include/python2.5/pyport.h:212:18: error: time.h: No such file or directory /usr/include/python2.5/pyport.h:230:24: error: sys/select.h: No such file or directory /usr/include/python2.5/pyport.h:269:22: error: sys/stat.h: No such file or directory In file included from /usr/include/python2.5/Python.h:76, from _configtest.c:2: /usr/include/python2.5/pymem.h:50: warning: parameter names (without types) in function declaration /usr/include/python2.5/pymem.h:51: error: expected declaration specifiers or ‘...’ before ‘size_t’ In file included from /usr/include/python2.5/Python.h:78, from _configtest.c:2: /usr/include/python2.5/object.h:104: error: expected specifier-qualifier-list before ‘Py_ssize_t’ /usr/include/python2.5/object.h:108: error: expected specifier-qualifier-list before ‘Py_ssize_t’ /usr/include/python2.5/object.h:131: error: expected declaration specifiers or ‘...’ before ‘*’ token /usr/include/python2.5/object.h:131: warning: type defaults to ‘int’ in declaration of ‘Py_ssize_t’ /usr/include/python2.5/object.h:131: error: ‘Py_ssize_t’ declared as function returning a function /usr/include/python2.5/object.h:131: warning: function declaration isn’t a prototype . (etc,etc) and here is a part of python2.5 compilation log: gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4) configure:2093: $? = 0 configure:2095: gcc -V /dev/null 5 gcc: '-V' option must have argument configure:2098: $? = 1 configure:2121: checking for C compiler default output file name configure:2124: gcc conftest.c 5 /usr/bin/ld: crt1.o: No such file: No such file or directory collect2: ld returned 1 exit status configure:2127: $? = 1 configure: failed program was: | /* confdefs.h. */ | | #define _GNU_SOURCE 1 | #define _NETBSD_SOURCE 1 | #define __BSD_VISIBLE 1 | #define _BSD_TYPES 1 | #define _XOPEN_SOURCE 600 | #define
Re: [Numpy-discussion] NumPy 1.0.3 release next week
Hi, you have a problem with your Ubuntu installation, not with numpy. Matthieu 2007/5/13, dmitrey [EMAIL PROTECTED]: Stefan van der Walt wrote: On Sun, May 13, 2007 at 06:19:30PM +0300, dmitrey wrote: Is it possible somehow to speedup numpy 1.0.3 appearing in Linux update channels? (as for me I'm interested in Ubuntu/Kubuntu, currently there is v 1.0.1) I tried to compile numpy 1.0.2, but, as well as in Octave compiling, it failed because c compiler can't create executable. gcc reinstallation didn't help, other c compilers are absent in update channel (I had seen only tcc, but I'm sure it will not help, and it (I mean trying to install other C compilers) requires too much efforts). Many people here are compiling numpy fine under Ubuntu. Do you have write permissions to the output directory? What is the compiler error given? Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion Sorry, I meant compiling Python2.5 and Octave, not numpy Octave Python2.5 is already present (in Ubuntu 7.04), but I tried to compile and install it from sources because numpy compilation failed with (I have gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4), compiling as root) ... ... C compiler: gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O2 -Wall -Wstrict-prototypes -fPIC compile options: '-I/usr/include/python2.5 -Inumpy/core/src -Inumpy/core/include -I/usr/include/python2.5 -c' gcc: _configtest.c In file included from /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/syslimits.h:7, from /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/limits.h:11, from /usr/include/python2.5/Python.h:18, from _configtest.c:2: /usr/lib/gcc/x86_64-linux-gnu/4.1.2/include/limits.h:122:61: error: limits.h: No such file or directory In file included from _configtest.c:2: /usr/include/python2.5/Python.h:32:19: error: stdio.h: No such file or directory /usr/include/python2.5/Python.h:34:5: error: #error Python.h requires that stdio.h define NULL. /usr/include/python2.5/Python.h:37:20: error: string.h: No such file or directory /usr/include/python2.5/Python.h:39:19: error: errno.h: No such file or directory /usr/include/python2.5/Python.h:41:20: error: stdlib.h: No such file or directory /usr/include/python2.5/Python.h:43:20: error: unistd.h: No such file or directory /usr/include/python2.5/Python.h:55:20: error: assert.h: No such file or directory In file included from /usr/include/python2.5/Python.h:57, from _configtest.c:2: /usr/include/python2.5/pyport.h:7:20: error: stdint.h: No such file or directory In file included from /usr/include/python2.5/Python.h:57, from _configtest.c:2: /usr/include/python2.5/pyport.h:73: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'Py_uintptr_t' /usr/include/python2.5/pyport.h:74: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'Py_intptr_t' /usr/include/python2.5/pyport.h:97: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'Py_ssize_t' /usr/include/python2.5/pyport.h:204:76: error: math.h: No such file or directory /usr/include/python2.5/pyport.h:211:22: error: sys/time.h: No such file or directory /usr/include/python2.5/pyport.h:212:18: error: time.h: No such file or directory /usr/include/python2.5/pyport.h:230:24: error: sys/select.h: No such file or directory /usr/include/python2.5/pyport.h:269:22: error: sys/stat.h: No such file or directory In file included from /usr/include/python2.5/Python.h:76, from _configtest.c:2: /usr/include/python2.5/pymem.h:50: warning: parameter names (without types) in function declaration /usr/include/python2.5/pymem.h:51: error: expected declaration specifiers or '...' before 'size_t' In file included from /usr/include/python2.5/Python.h:78, from _configtest.c:2: /usr/include/python2.5/object.h:104: error: expected specifier-qualifier-list before 'Py_ssize_t' /usr/include/python2.5/object.h:108: error: expected specifier-qualifier-list before 'Py_ssize_t' /usr/include/python2.5/object.h:131: error: expected declaration specifiers or '...' before '*' token /usr/include/python2.5/object.h:131: warning: type defaults to 'int' in declaration of 'Py_ssize_t' /usr/include/python2.5/object.h:131: error: 'Py_ssize_t' declared as function returning a function /usr/include/python2.5/object.h:131: warning: function declaration isn't a prototype . (etc,etc) and here is a part of python2.5 compilation log: gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4) configure:2093: $? = 0 configure:2095: gcc -V /dev/null 5 gcc: '-V' option must have argument configure:2098: $? = 1 configure:2121: checking for C compiler default output file name configure:2124: gcc conftest.c 5 /usr/bin/ld: crt1.o: No such file: No such file or directory collect2: ld returned 1 exit status configure:2127: $? = 1 configure: failed program was: | /* confdefs.h. */ | | #define _GNU_SOURCE 1 | #define _NETBSD_SOURCE 1 | #define
Re: [Numpy-discussion] NumPy 1.0.3 release next week
Hi Dmitrey On Sun, May 13, 2007 at 08:21:15PM +0300, dmitrey wrote: Many people here are compiling numpy fine under Ubuntu. Do you have write permissions to the output directory? What is the compiler error given? Sorry, I meant compiling Python2.5 and Octave, not numpy Octave Python2.5 is already present (in Ubuntu 7.04), but I tried to compile and install it from sources because numpy compilation failed with (I have gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4), compiling as root) This isn't really the place to discuss compiling Python or Octave, but a good first move would be to install the 'build-essential' package. This will hopefully provide the header files and the compiler you need. Cheers Stéfan ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] copy object with multiple subfields, including ndarrays
It's alwys helpful if you can include a self contained example so it's easy to figure out exactly what you are getting at. I say that because I'm not entirely sure of the context here -- it appears that this is not numpy related issue at all, but rather a general python question. If so, I think what you are looking for is copy.deepcopy. As it name implies it does a deep copy of an object as opposed to a shallow copy, which is what copy.copydoes. If that doesn't do what you want or I misunderstood your question, please supply some more detail. On 5/13/07, dmitrey [EMAIL PROTECTED] wrote: hi all, does anyone know howto copy an instance of class, that contains multiple subfields, for example myObj.field1.subfield2 = 'asdf' myObj.field4.subfield8 = numpy.mat('1 2 3; 4 5 6') I tried from copy import copy myObjCopy = copy(myObj) but it seems that it doesn't work correctly Thx, D. ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion -- //=][=\\ [EMAIL PROTECTED] ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] very large matrices.
Are you trying some sort of principal components analysis? PCA is indeed one part of the research I'm doing. Dave ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] very large matrices.
There are definitely elements of spectral graph theory in my research too. I'll summarise We are interested in seeing the each eigenvector from svd can represent in a semantic space In addition to this we'll be testing it against some algorithms like concept indexing (uses a bipartitional k-meansish method for dim reduction) also testing against Orthogonal Locality Preserving indexing, which uses the laplacian of a similarity matrix to calculate projections of a document (or term) into a manifold. These methods have been implemented and tested for document classification, I'm interested in seeing their applicability to modelling semantics with a system known as Hyperspace to analog language. I was hoping to do svd to my HAL built out of reuters, but that was way too big. instead i'm trying with the traces idea i mentioned before (ie contextually grepping a keyword out of the docs to build a space around it.) Cheers Dave On 5/14/07, Charles R Harris [EMAIL PROTECTED] wrote: On 5/13/07, Dave P. Novakovic [EMAIL PROTECTED] wrote: Are you trying some sort of principal components analysis? PCA is indeed one part of the research I'm doing. I had the impression you were trying to build a linear space in which to embed a model, like atmospheric folk do when they try to invert spectra to obtain thermal profiles. Model based compression would be another aspect of this. I wonder if there aren't some algorithms out there for this sort of thing. Chuck ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion