Re: [pymvpa] PyMVPA Growing Nerual Gas implementation
Hi Tony, you can actually use the MDP GNG implementation directly in PyMVPA using the MDPFlowMapper. An example of how to use MDP within PyMVPA can be found here: http://www.pymvpa.org/examples/mdp_mnist.html HTH, Tiziano On Thu 23 Jan, 15:42, Jiang, Zhiguo zhiguo.ji...@njit.edu wrote: Hi, All this might be unrelated to PyMVPA. but I am trying to find a good implementation of clustering using Grown Neural Gas algorithm. I couldn't find any information on PyMVPA. MDP offers growing neural gas class but it is not implemented for fMRI data. anyone has any idea either finding a good nerual gas python tool or modyfying the existing tool so it can be applied to fMRI time course data. Any suggestions? Thanks, Tony ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
Re: [pymvpa] PyMVPA for Python 3 -- second report
was it painful? not really, just a bit of vi gymnastics ;) ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
Re: [pymvpa] PyMVPA for Python 3 -- second report
may be some skype/shared screen sprintee next week over remaining issues etc? yes, if michael is around we could meet in berlin or magdeburg and have get you on skype. There is a bulk of changes in some guys feature branches and if we do not merge now, later it might not be as easy ;) I guess I would also need to check how changes affect minimal compatible version in 2.x series (if affect at all). FWIC (for what i care) 2.6 should be good enough for me ;) I have tried to stay compatible with 2.5 so that you can backport to squeeze, but I did not test if this is really the case. but at times going through changes I got a bit scared ;-) e.g. that is the most controversial change, I think: 1. -self._unique_values = np.unique(self.value) +# get a 1-D array +arrvalue = np.asarray(self.value).ravel() +# check if we have None somewhere +nones = np.equal(arrvalue, None) +if nones.any(): +uniques = np.unique(arrvalue[nones == False]) +# put back one None +self._unique_values = np.insert(uniques, 0, None) +else: +self._unique_values = np.unique(arrvalue) does numpy handles its own arrays differently among pythons or what is the reason? the reason is that sometimes self.values contains None's. in python 3 arithmetic comparisons like or are not allowed among heterogeneous types, so you can not sort a list or an array that looks for example like this: [ 0, 1, None, -2]. and np.unique uses sort internally. the dance with the None's is not needed in py2, so we could special case with a sys.version = '3', but you have to decide if it is worth it. the new masked array type in newcoming numpy could probably used instead... it is not an inferior behaviour when compared to py2: actually I think I spotted at least one place (the tests is still failing) where PyMVPA was comparing non-comparable types and getting a basically random result: == ERROR: test_basic (mvpa2.tests.test_datameasure.SensitivityAnalysersTests) -- Traceback (most recent call last): File /home/tiziano/git/PyMVPA/build/py3k/mvpa2/testing/sweepargs.py, line 67 , in do_sweep method(*args_, **kwargs_) File /home/tiziano/git/PyMVPA/build/py3k/mvpa2/tests/test_datameasure.py, li ne 79, in test_basic self.assertTrue(f[0] 0.1) # some reasonably large value TypeError: unorderable types: Dataset() float() 2. -sample += random.sample((targets == r).nonzero()[0], npertarget[i] ) +sample += random.sample(list((targets == r).nonzero()[0]), npertarget[i] ) why does it need explicit conversion to list, are numpy arrays not iterables any longer, or random.sample now can't tollerate them... the reason is: Python 3.2.3rc2 (default, Mar 21 2012, 05:47:04) [GCC 4.6.3] on linux2 Type help, copyright, credits or license for more information. import numpy x = numpy.zeros((10,)) import random random.sample(x, 1) Traceback (most recent call last): File stdin, line 1, in module File /usr/lib/python3.2/random.py, line 298, in sample raise TypeError(Population must be a sequence or set. For dicts, use list(d).) TypeError: Population must be a sequence or set. For dicts, use list(d). I do not know why numpy arrays are not sequences in py3, but this is it. +Convert *py files with lib2to3. +Taken from MDP and numpy. so -- copyright of numpy and MDP developers? years? (BSD-3 I guess ;) ) will do :) ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
Re: [pymvpa] PyMVPA for Python 3 -- second report
cool -- you can consider me now a contributor to this little project ;-) I have sent a pull request against your branch with few changes, please review. I merged your pull request, thank you :) Also it might be nice to markup those commits which deal with python3 compatibility as PY3 (e.g. instead of BF), not sure if it is worth rewriting the history for this cause though :-) but might be good to start at least now that's a good idea. I fear my git foo stops at git rebase -i, to do what you want we would need a git filter-branch, right? I would prefer to rewrite history, if you give me a couple of hints on how to do it properly. ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
[pymvpa] PyMVPA for Python 3 -- second report
hi all, with my today's efforts I got down to 13 errors and 15 failures out of 284 tests, so it's getting better :) on the same branch the same set of unittests passes without errors or failures with python 2.7, so I assume I did not break anything important. I attach here the error logs in case someone else feels like having a look. the remaining errors and failures are probably out of my reach, as fixing them requires a much deeper understanding of the very rich data types structure of PyMVPA ;) I am available for a short sprint in case someone is orbiting around northern germany... there are some commits that I think would require a deeper review by some core developer, as I am not 100% sure those changes are really as neutral as the lack of failures in the unittests seems to suggest. I had to disable a test that was hanging in an infinite loop: test_nfold_random_counted_selection_partitioner no idea what the problem there is, really. one point where I can directly ask for a solution is the usage of the variable nproc in BaseSearchlight (mvpa2/measures/searchlight.py): the doc-string seems to imply that setting it to None would use all available cores, but if the pprocess module is not installed this is not true, and an error is spit only in the case nproc 1 (which is false if nproc is None). so maybe the doc-string could be updated to say something like: nproc : None or int How many processes to use for computation. Requires `pprocess` external module. If None -- all available cores will be used if `pprocess` is available, a single core will be used otherwise. ciao, tiziano python3 setup.py config --noisy Converting to Python3 via 2to3... running config python3 setup.py build_ext --inplace Converting to Python3 via 2to3... running build_ext running build_src build_src building extension mvpa2.clfs.libsmlrc.smlrc sources building extension mvpa2.clfs.libsvmc._svmc sources building data_files sources build_src: building npy-pkg config files customize UnixCCompiler customize UnixCCompiler using build_ext customize UnixCCompiler customize UnixCCompiler using build_ext running scons touch build3-stamp I: Running py3-compatible unittests. None of them should ever fail. T: MVPA_SEED=1251509818 T: Testing for availability of external software packages. WARNING: libsvm verbosity control is not available. * Please note: warnings are printed only once, but underlying problem might occur many times * test_externals (mvpa2.tests.test_externals.TestExternals) ... ok test_externals_correct2nd_invocation (mvpa2.tests.test_externals.TestExternals) ... ok test_externals_no_double_invocation (mvpa2.tests.test_externals.TestExternals) ... ok test_basic (mvpa2.tests.test_dochelpers.DochelpersTests) ... ok test_borrow_doc (mvpa2.tests.test_dochelpers.DochelpersTests) ... ok test_borrow_kwargs (mvpa2.tests.test_dochelpers.DochelpersTests) ... ok test_searchlight_doc (mvpa2.tests.test_dochelpers.DochelpersTests) ... ok test_simple_som (mvpa2.tests.test_som.SOMMapperTests) ... ERROR test_blank_state (mvpa2.tests.test_state.StateTests) ... ok test_deep_copying_state_variable (mvpa2.tests.test_state.StateTests) ... ok Check if we can store/restore set of enabled ca ... ok test_parametrized (mvpa2.tests.test_state.StateTests) ... ERROR test_proper_state (mvpa2.tests.test_state.StateTests) ... ok Simple test if child gets conditional attributes from the parent as well ... ok test_read_only (mvpa2.tests.test_state.StateTests) ... ok To test new ca ... ok test_stored_temporarily (mvpa2.tests.test_state.StateTests) ... ok test_value_in_constructor (mvpa2.tests.test_state.StateTests) ... ok test_blank (mvpa2.tests.test_params.ParamsTests) ... ok test_classifier (mvpa2.tests.test_params.ParamsTests) ... ok test_incorrect_parameter_error (mvpa2.tests.test_params.ParamsTests) ... ok test_mixed (mvpa2.tests.test_params.ParamsTests) ... ok test_simple (mvpa2.tests.test_params.ParamsTests) ... ok test_config (mvpa2.tests.test_config.ConfigTests) ... ok test_asobjarray (mvpa2.tests.test_support.SupportFxTests) ... ok test_break_points (mvpa2.tests.test_support.SupportFxTests) ... ok test_correlation (mvpa2.tests.test_support.SupportFxTests) ... ok test_event (mvpa2.tests.test_support.SupportFxTests) ... ok test_id_hash (mvpa2.tests.test_support.SupportFxTests) ... FAIL test_map_overlap (mvpa2.tests.test_support.SupportFxTests) ... ok test_mof_n_combinations (mvpa2.tests.test_support.SupportFxTests) ... ok Test our ad-hoc SmartVersion ... ERROR test_transform_with_boxcar (mvpa2.tests.test_support.SupportFxTests) ... ok Test conversion of versions from strings ... ok test_xrandom_unique_combinations (mvpa2.tests.test_support.SupportFxTests) ... ok mvpa2.tests.test_support.test_value2idx ... ok mvpa2.tests.test_support.test_limit_filter ... ok Test if works fine with carriage return (cr) symbol ... ok test_debug (mvpa2.tests.test_verbosity.VerboseOutputTest) ... ok test_debug_rgexp
Re: [pymvpa] PyMVPA for Python 3
if you push your work somewhere (I don't see your clone at https://github.com/PyMVPA/PyMVPA/network) we will add it to daily testing I don't know why you don't see it, but I do have a fork on github: https://github.com/otizonaizit/PyMVPA/ but please, slow down, this is very much work in progress and it is by no means ready to be daily tested by anyone who is not me ;) most quick tests are failing anyway *and* I want to be free to rewrite history a bit before having other people pulling from it... MVPA_TESTS_QUICK=yes for the final test run since that would exclude looping over parameters in many parametric tests, and just run on the first argument only, thus not exploring as much fair enough. I'll remember to switch it off again as soon as my py3 port starts passing the quick tests. MVPA_MATPLOTLIB_BACKEND=agg would it help? yes, it is helping already. P.S. Meanwhile I will upgrade our testing box and will setup everything we have for python3 so far ;) whichis ATM, I fear, only python3-numpy and python3-mdp... I hope python3-scipy and python3-matplotlib make it in time for wheezy, but given the amount of nit-picking going on on python-debian ML, I fear we may be stuck without a proper py3 environment in debian for another couple of years... ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
Re: [pymvpa] PyMVPA for Python 3 -- first report
My guess is that it is still related to how that AttributesCollector injects those additional attributes into the freshly constructed classes. I hope above clues would be of help it helped indeed ;) ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
Re: [pymvpa] PyMVPA for Python 3
MVPA_TESTS_QUICK=yes MVPA_TESTS_LOWMEM=yes make unittest-nonlabile [...] Let us know if that helps for a start. yep, it works fine this way, thanks. it even finds more unit tests (326) ;) ciao, tiziano ___ Pkg-ExpPsy-PyMVPA mailing list Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa