CIA.vc
swaroopgj
Real-time open source activity stats
Stats » Authors » swaroopgj
informationsyndicateUTC clock
07:57 on May 28, 2012
event counters
The last message was received 2.04 months ago at 03:55 on Mar 28, 2012
0 messages so far today, 0 messages yesterday
0 messages so far this week, 0 messages last week
0 messages so far this month, 0 messages last month
145 messages since the first one, 2.3 years ago, for an average of 0.83 weeks between messages
recent messages
dateReversed sort columnprojectcontentlink
00:03 on Mar 28pymvpa
Commit by swaroopgj on dist2/debian/proper/sid :: r538f550 pymvpa/ (77 files in 28 dirs):
Merge remote branch 'origin/master'
#
00:03 on Mar 28pymvpa
Commit by swaroopgj on dist2/debian/proper/sid :: r6b605ba pymvpa/mvpa2/ (3 files in 3 dirs):
ENH: Added dgesvd as an optional svd implementation in procrusteanmapper
#
17:02 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: rd3ff78b pymvpa/doc/examples/cachedkernel.py:
D: Adding a simple example on how to use cached kernel
#
17:01 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r48f1801 pymvpa/ (6 files in 5 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: RF+BIG: Heavy refactoring of HDF5 code. NF: Support for recursive references in HDF5 IO BF: Simplify object detection in h5load() DOC: adding gnbsearchlight to modref RF: not running daily tests for mh/master upon mh request ;-) ENH: provide "filler" argument to FeatureSliceMapper which could be used instead of 0 upon reverse call. Close gh-28
#
17:01 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r4772dbb pymvpa/ (39 files in 20 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (27 commits) DOC: Update sensanas example output figure. BF: HDF5 can handle single 0d array(object) BF: test_svmkernels ValueError should be raised only in __debug__ mode BF: order voxels for xyz order now in eventrelated example RF: 2nd wave of deprecating demo_blocfmri RF: deprecating demo_blockfmri -- we use tutorial_data -- 1st step --renaming NF: oops -- forgotten to commit asciiart -- what if comes useful BF: provide exit=False for nose whenever ran from cmdline RF: deprecating get_unique_length_n_combinations in favor of (x)unique_combinations BF: removed duplicate get_unique_length_n_combinations DOC: extended Emanuele's comment on returned datasets in feature selection BF: removed duplicate *getUniqueLengthNCombinations -- appeared due to cherry pick and later merge BF: provide output of git describe not within action so it could be seen actually DOC: documenting unexpected sharing of .labels and .chunks. ENH: option to enforce recaching of CachedKernel in compute() BF(RF): use init_origids in SamplesLookup, and assure that all those are unique ENH: atlaslabeler -- enhanced statistics + reporting coordinates via qform OPT: using np.ix_ in selecting a subkernel in CachedKernel ENH: debug messages for CachedKernel + unittest which hoped to ring the bell for Swaroop's issue (it didn't) ENH: debug messages for SamplesLookup ...
#
16:59 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r22ccacf pymvpa/ (107 files in 24 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (150 commits) DOC: More references. BF: in-place (before it would operate on copy) zscoring of ndarrays. . Close gh-27 BF: moving scipy-dependent statement in gnbsearchlight BF+RF: apparently epsilon is tube_epsilon for sg.LibSVR + sg.svm.seed DOC: Restructure paper listing. DOC: Note new users with Nature Neuroscience paper ;-) BF: workaround for shogun -- epsilon in constructor for regressions has no effect NF: mvpa.support.ipython + ipy_pymvpa_completer BF: import debug for kernels.sg ENH: complement CorrErrorFx with just CorrFx ENH: fslatlas.get_maps -- accepts now 'key_attr' defining on what to use for keys ENH: more debugging for sg kernels and svm BF+DOC: Wrong link to tutorial. RF: atlas.levels now corresponds to previous atlas.levels_dict RF/ENH: atlaslabeler -- ability to provide a stream of voxels indices RF: atlases __getitem__ for queries by voxel indices, __call__ -- by coordinates BF: actually fail (error exist status) whenever unittest-debug fails (uff) BF: access selected_ids from .ca BF: unittest-debug -- do not rely on MVPA_SEED or OK being in the tail -- just go to the end ;) BF: nr_class can be None, so just use %s in debug printout ...
#
16:54 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r69d5918 pymvpa/ (81 files in 20 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (61 commits) DOC: Note NiBabel as alternative MRI IO backend. Close gh-18 BF: Remove print statements; allow hdf5 tests also for nibabel. BF: fixed up or RF a bit magical parameters/docstrings handling RF+BIG: Allow loading MRI data with nibabel. RF: searchlights -- consistent prefix roi_ (instead of roi in roisizes) and roi_ids instead of center_ids (which is still used in sphere_searchlight) RF: minimal RF of searchlight.py so we could have subclasses with specific implementations RF: taking larger dataset for test_hdf5_clf so error is more stable RF: Allow for nibabel as alternative storage backend for MRI. RF+BIG: Change fMRI samples from z,y,x to x,y,z PL: minor in searchlight DOCstrings: Default: in params descriptions gets repr, no unneeded indent for ca listing DOC (+BF): GNB -- just minor touches removing test_dataset.py -- if want to look at old code -- checkout maint/0.4 ;-) RF: explicitly call suppress warnings for scipy if scipy is known to be present RF: Prep for an alternative MRI IO backend. NF: additional testing for storing/restoring/training of classifier minor: for convenience bind sim_modulated to sin_modulated_train as well NF+BF: obj2hdf to puke when trying to store lambdas, hdf2obj handle types, HDF5ConversionError new refactor_0.5.sh to reflect the fact that those were deprecated RF: second_axis_*_of_abs transformers deprecated into favor of *_of_abs ...
#
16:52 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r4cc32b6 pymvpa/ (27 files in 12 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: DOC: Typos DOC: Add comments about tutorial results and 'alt' hints Put tutorial_lib.py into tools/ BF: fixed support text link + removed Freenode for IRC DOC+RF: Tutorial data now in separate location. DOC+BK: Complete draft of event-related tutorial; mapper bug pending BF: Unittest for 1D multispace searchlights. RF: FeatureSelection returns a single dataset if possible. RF+NF: chunk -> block within searchlight (confusing with chunks) + SLC_ debug target BF: make Sphere to deal with scalars as coordinates (i.e. degenerate 1D ;-)) DOC: Bits for the event-related tutorial. BK: IndexQueryEngine is broken. BF: Unbelievable -- erdataset tests never ran. RF: plot_err_line() returns line objects for later modification. DOC: Updated event-related figure. RF: Reuse as much as we can. BF: include mvpa.support.nipy into distribution
#
16:52 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r2e418c4 pymvpa/doc/source/tutorial_eventrelated.rst:
DOC: Typos
#
16:51 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r90dc99e pymvpa/ (241 files in 35 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (131 commits) DOC: Complete draft of the sensitivity tutorial. DOC: tiny adjustment to TreeClassifier docstring RF: tune up ipython completion only when importing mvpa.suite NF: by default in ipython do not complete protected/private DOC: First half of the sensitivity tutorial. print meaningful msg for knn plot example DOC: minor RF: raiseException -> raise_ (following PEP) DOC: Note in release notes about N -> np RF (HUGE): 3nd wave -- import numpy/pylab as np/pl correspondingly to be inline with most of other python projects RF: minor -- commented out imports * in easy-to-debug code within plot_samples_distance to please Python RF: 2nd wave -- perchunk -> chunks_attr + some places added targets_attr argument (e.g. permute_labels) RF: 1st wave -- chunks='chunks' -> chunks_attr='chunks' NF: Splitter now perform basic slicing by default if possible. DOC: minor in tutorial_significance -- so that I could proceed with refactorings BF: temporary to assign versions upon request (if no external checking was done before, like if their presence was loaded from a config file) DOC: adjustments to significance testing blurb DOC: put licenses/copyrights for 3rd party code into COPYING "enabling" nipy"s emp_null within PyMVPA (adjusting unittest/ adding tosuite) Borrowing emp_null NiPy to be carried within PyMVPA to avoid evil 'reinvent the wheel' ...
#
16:50 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r175ff8c pymvpa/ (4 files in 3 dirs):
DOC: Typos
#
16:45 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: rd3ff78b pymvpa/doc/examples/cachedkernel.py:
D: Adding a simple example on how to use cached kernel
#
16:45 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r48f1801 pymvpa/ (6 files in 5 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: RF+BIG: Heavy refactoring of HDF5 code. NF: Support for recursive references in HDF5 IO BF: Simplify object detection in h5load() DOC: adding gnbsearchlight to modref RF: not running daily tests for mh/master upon mh request ;-) ENH: provide "filler" argument to FeatureSliceMapper which could be used instead of 0 upon reverse call. Close gh-28
#
16:45 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r4772dbb pymvpa/ (39 files in 20 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (27 commits) DOC: Update sensanas example output figure. BF: HDF5 can handle single 0d array(object) BF: test_svmkernels ValueError should be raised only in __debug__ mode BF: order voxels for xyz order now in eventrelated example RF: 2nd wave of deprecating demo_blocfmri RF: deprecating demo_blockfmri -- we use tutorial_data -- 1st step --renaming NF: oops -- forgotten to commit asciiart -- what if comes useful BF: provide exit=False for nose whenever ran from cmdline RF: deprecating get_unique_length_n_combinations in favor of (x)unique_combinations BF: removed duplicate get_unique_length_n_combinations DOC: extended Emanuele's comment on returned datasets in feature selection BF: removed duplicate *getUniqueLengthNCombinations -- appeared due to cherry pick and later merge BF: provide output of git describe not within action so it could be seen actually DOC: documenting unexpected sharing of .labels and .chunks. ENH: option to enforce recaching of CachedKernel in compute() BF(RF): use init_origids in SamplesLookup, and assure that all those are unique ENH: atlaslabeler -- enhanced statistics + reporting coordinates via qform OPT: using np.ix_ in selecting a subkernel in CachedKernel ENH: debug messages for CachedKernel + unittest which hoped to ring the bell for Swaroop's issue (it didn't) ENH: debug messages for SamplesLookup ...
#
16:40 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r22ccacf pymvpa/ (107 files in 24 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (150 commits) DOC: More references. BF: in-place (before it would operate on copy) zscoring of ndarrays. . Close gh-27 BF: moving scipy-dependent statement in gnbsearchlight BF+RF: apparently epsilon is tube_epsilon for sg.LibSVR + sg.svm.seed DOC: Restructure paper listing. DOC: Note new users with Nature Neuroscience paper ;-) BF: workaround for shogun -- epsilon in constructor for regressions has no effect NF: mvpa.support.ipython + ipy_pymvpa_completer BF: import debug for kernels.sg ENH: complement CorrErrorFx with just CorrFx ENH: fslatlas.get_maps -- accepts now 'key_attr' defining on what to use for keys ENH: more debugging for sg kernels and svm BF+DOC: Wrong link to tutorial. RF: atlas.levels now corresponds to previous atlas.levels_dict RF/ENH: atlaslabeler -- ability to provide a stream of voxels indices RF: atlases __getitem__ for queries by voxel indices, __call__ -- by coordinates BF: actually fail (error exist status) whenever unittest-debug fails (uff) BF: access selected_ids from .ca BF: unittest-debug -- do not rely on MVPA_SEED or OK being in the tail -- just go to the end ;) BF: nr_class can be None, so just use %s in debug printout ...
#
16:38 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: rd3ff78b pymvpa/doc/examples/cachedkernel.py:
D: Adding a simple example on how to use cached kernel
#
16:38 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r48f1801 pymvpa/ (6 files in 5 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: RF+BIG: Heavy refactoring of HDF5 code. NF: Support for recursive references in HDF5 IO BF: Simplify object detection in h5load() DOC: adding gnbsearchlight to modref RF: not running daily tests for mh/master upon mh request ;-) ENH: provide "filler" argument to FeatureSliceMapper which could be used instead of 0 upon reverse call. Close gh-28
#
16:37 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: r4772dbb pymvpa/ (39 files in 20 dirs):

Merge branch 'mh/master' of git://github.com/hanke/PyMVPA

  • 'mh/master' of git://github.com/hanke/PyMVPA: (27 commits) DOC: Update sensanas example output figure. BF: HDF5 can handle single 0d array(object) BF: test_svmkernels ValueError should be raised only in __debug__ mode BF: order voxels for xyz order now in eventrelated example RF: 2nd wave of deprecating demo_blocfmri RF: deprecating demo_blockfmri -- we use tutorial_data -- 1st step --renaming NF: oops -- forgotten to commit asciiart -- what if comes useful BF: provide exit=False for nose whenever ran from cmdline RF: deprecating get_unique_length_n_combinations in favor of (x)unique_combinations BF: removed duplicate get_unique_length_n_combinations DOC: extended Emanuele's comment on returned datasets in feature selection BF: removed duplicate *getUniqueLengthNCombinations -- appeared due to cherry pick and later merge BF: provide output of git describe not within action so it could be seen actually DOC: documenting unexpected sharing of .labels and .chunks. ENH: option to enforce recaching of CachedKernel in compute() BF(RF): use init_origids in SamplesLookup, and assure that all those are unique ENH: atlaslabeler -- enhanced statistics + reporting coordinates via qform OPT: using np.ix_ in selecting a subkernel in CachedKernel ENH: debug messages for CachedKernel + unittest which hoped to ring the bell for Swaroop's issue (it didn't) ENH: debug messages for SamplesLookup ...
#
16:32 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: rd3ff78b pymvpa/doc/examples/cachedkernel.py:
D: Adding a simple example on how to use cached kernel
#
16:32 on Jun 10, 2010pymvpa
Commit by swaroopgj on per/master :: rd3ff78b pymvpa/doc/examples/cachedkernel.py:
D: Adding a simple example on how to use cached kernel
#