Skip to content
Commits on Feb 4, 2016
  1. @nno

    Merge pull request #422 from yarikoptic/bf-openfmri-ds

    nno committed
    BF+ENH: set TR from scan_key for openfmri datasets' with TR=1 in nifti files
Commits on Feb 1, 2016
  1. @yarikoptic

    BF+ENH: set TR from scan_key for openfmri datasets' with TR=1 in nift…

    yarikoptic committed
    …i files
    
    also added datalad external, if crawled openfmri ds 105 is present - smoke testing above
    fix on it
Commits on Jan 27, 2016
  1. @hanke

    Merge pull request #421 from yarikoptic/nf-savez

    hanke committed
    NF: to/from_npz for Datasets
  2. @yarikoptic

    NF: to/from_npz for Datasets

    yarikoptic committed
Commits on Jan 26, 2016
  1. @yarikoptic
  2. @yarikoptic

    Merge pull request #407 from hanke/isnot

    yarikoptic committed
    RF: not X is None -> X is not None
Commits on Jan 25, 2016
  1. @yarikoptic

    Merge pull request #419 from hanke/statsmodels

    yarikoptic committed
    BF: Sample attribute name convenience can fail for non-default settings
  2. @hanke
  3. @hanke

    Merge pull request #418 from yarikoptic/master

    hanke committed
    ENH: make partitioner (aka generator) optional for CrossValidation and ad-hoc searchlights
Commits on Jan 22, 2016
  1. @yarikoptic

    BF: simplify testing of cv_no_generator tests since we need really a …

    yarikoptic committed
    …classifier there
    
    and _predict could be called on training or testing data, which messed up testing a bit if all ca
    are enabled
  2. @yarikoptic
  3. @yarikoptic

    BF: adhoc (e.g. gnb) searchlights now can tollerate awkward partition…

    yarikoptic committed
    …ings
    
    E.g. the ones which do not mark some samples as training or testing (so partitions==0 for
    them), so custom splitter in gnbsearchlight also got its values hardcoded to (1, 2) not
    just a sorted range of all possible which would start from 0
  4. @yarikoptic

    ENH: make partitioner optional for CrossValidation

    yarikoptic committed
    primarily to facilitate cross- classifications, so people could just specify
    a Splitter on the values of the necessary attribute and be done
  5. @yarikoptic

    Merge pull request #417 from hanke/fixtypo

    yarikoptic committed
    BF: Typo + PEP8
  6. @hanke

    BF: Typo + PEP8

    hanke committed
Commits on Jan 18, 2016
  1. @hanke

    Merge pull request #416 from yarikoptic/master

    hanke committed
    BF+ENH: explicitly use float for "sum of" estimation in anova so we get proper float division
  2. @yarikoptic

    Merge remote-tracking branch 'gh-hanke/numpy1.10_fix'

    yarikoptic committed
    * gh-hanke/numpy1.10_fix:
      PEP8 compliance
      BF: Make ANOVA operation compatible with non-float data
  3. @yarikoptic

    BF+ENH: explicitly use float for "sum of" estimation in anova so we g…

    yarikoptic committed
    …et proper float division
    
    also OPT to use some inplace operations without temporary arrays being created
Commits on Jan 17, 2016
  1. @hanke

    Merge pull request #414 from hanke/fix_coverage

    hanke committed
    BF: Random attempt to fix the coverage setup
  2. @hanke

    Merge pull request #412 from hanke/enh_coverage

    hanke committed
    ENH: Modernize SMLR test (run more of its code)
  3. @hanke
  4. @hanke

    BF: Random attempt to fix the coverage setup

    hanke committed
    Aims to fix gh-411
Commits on Jan 16, 2016
  1. @yarikoptic

    Merge pull request #408 from hanke/remove_hamster

    yarikoptic committed
    RM: Removal of Hamster()
  2. @hanke

    Merge pull request #410 from hanke/reality_update

    hanke committed
    BF: Adjust code to match reality
  3. @hanke
  4. @hanke

    BF: Adjust code to match reality

    hanke committed
  5. @nno

    Merge pull request #406 from hanke/nomutabledefaults

    nno committed
    BF: Replace all occurrences of mutable default values with a safe default
  6. @hanke

    RM: Removal of Hamster()

    hanke committed
    I don't think this class serves any purpose anymore. I haven seen it
    being used in five years. I propose its removal.
  7. @hanke

    Merge pull request #402 from yarikoptic/enh-permutator-limit

    hanke committed
    ENH: allow limit for AttributePermutator to be a list
  8. @hanke

    RF: not X is None -> X is not None

    hanke committed
  9. @hanke
  10. @nno

    Merge pull request #405 from hanke/variousfixes

    nno committed
    Lots of tiny bits for general beautification
  11. @hanke
  12. @hanke
  13. @hanke
Something went wrong with that request. Please try again.