Bug#683292: unblock: nipy/0.2.0-1
Hi Adam,
Thanks you looking at the package.
On Fri, 17 Aug 2012, Adam D. Barratt wrote:
> > I am attaching a gzipped debdiff on .dsc from 0.2.0~rc2+git27-g7b9b5a5-1 --
> > there is a bulk of changes indeed BUT most of them are examples and
> > documentation fixes; all functionality changes/fixes were accompanied with
> > unittests and passed peer-review. Neither of those changes should effect
> > python-nipy -- the only package (loosly) depending on nipy.
> Even after dropping doc/*, examples/*, tests/* and whitespace-only
> changes, the debdiff is still
> 137 files changed, 16586 insertions(+), 8818 deletions(-)
> I'm not sure how we're supposed to sanely review that.
Sorry for dumping that big diff on you... even besides whitespace changes
there was a bulk of docstrings changes which could not be filtered easily indeed.
it might be worth looking at the upstream log of changes instead:
$> git log --pretty=oneline debian/0.2.0_rc2+git27-g7b9b5a5-1...debian/0.2.0-1 -- nipy
1cbe46b0038255bd6fad9438236427f397133b8f REL: bump info to indicate 0.2.0 release
bacaa6c98758211df69e6e5a6cf737ccc94399e9 MISC: make interfaces tests a package
83ff905c38cf40e7610a9ac693191307c36a7898 DOC: harmonize README and info.py
c6b4dc1b1cf13be42ea2aa2a2ebe4eb46c7e1df0 DOC: point to collated data installation page
f66cefa8daeea5e4b9a17fd729939aa27a84aafd DOC: fix section heading in fmri.utils
10cd8c168232dbd821cc96f6a0c0b5c8183cb75a Merge pull request #203 from matthew-brett/convolve-review
4e913137a2d822bd336e1119c4ad0d1991b854d6 HACK: add diag_indices helper for numpy < 1.4.0
e14dbaa0a4158cd9e8434ffd72a51c20f9893422 Merge branch 'multichannel-segmentation' into main-master
192a085cd64526e98de7d8f9e947669676b3f72a MISC: update generated .c files from .pyx files
a3ec7b96724944a4cf1736c7a9318c6f95c6f2a1 fixed type error related to numpy.exp
280746985c4d4c89c8c2385df7887340c6e899aa RF+TST: refactor convolve_functions interface
8a18eaed46eec59074121ff4b1756de6f1239b6f DOC: add note from Jonathan about bug in convolve
377d246f4d0b45bf6f3a7dbba3691d70e28329a8 RF+API: change interp behavior for fill arguments
dc1fb39d9a0b3ba227cee77f9e358850066c2979 Merge pull request #200 from matthew-brett/fmri-utils-fixes
9dca074b8be768140f2d5412488d4d2915f84721 fixed 64-bit bug in histogram cython function
d20376a81e281d4e4f1586ebc4ec7f8910f93727 add optional user-specified initialization for brain seg class
ee1d022d99ed8394a9e6e7ca94081568fd8d4066 fixed 64-bit bug + add related tests
b5b6dbae14c9eec357df1dc5ae22a375cba8b788 minor clean up
744f6e009bb951137055b8b9dccf03556b114fb1 update reference parameters for segmentation initialization
d04598f62dbe16b7bece15430fe7e722be2f1d85 fix numerical issue
1381389181f029fb2803d86dce3da01cdd2db3fd slight refactoring of the brain segmentation class and further tests
fdf8029db06217d69bc0a1acc7588f47269c633b add short comment
231ea9c71a1a959dbf30f09dd27a863ccd744337 enable ngb_size input argument in BrainSegmentation class
119b91da725d35d71d3f74c3d0850103554d4b8e adding a histogram test
4adf405b207d151fcfbee4711a7adfee174a0def add segmentation tests + some code cleaning
438b52a68399e08a636f6bba4a563558d998593b remove nibabel dependency in tissue classification example
c8a3dd1103f1d80484ae607a1eb27eb3425a84f9 generalize brain segmentation model
fa7aed950016eaa929ceabde94bcef3faf5464b7 fixed imports
b564596fa7cf057203ea78e7f7a9446e584a130e fixed tests
f53d8cf6a6c9c2449a7d377be3792704b0fe0ff1 adding a fast histogram routine in algorithms.statistics
f5019dfa041e1c12b7bcdeaeef8121d5297b40c1 fix memory leak
7ee5a0f49eab5a1159b8462e734cf910582a72df redesign brain segmentation model
965cd4f3bef45f6d339818168cec43f88bdc6f91 copy input parameter arrays in segmentation
db3a533db05a9c3a6ebbca1be24b2da1a0dfbc32 removed print statementsw
e5398287b4172d220220d23dfba08dcc687caadd small API change in Segmentation class
147bec1ee46d206300eb7a269086320c92f364fd fixed free energy computation under no zero-beta
035989a5e37862af6a83e4bc7cc3821b11f293f8 initialize VEM using proper uniform distributions
60ed6b92727dfd400aea4c20978dfdd681cbb460 fix segfault
02db1b46b15bfafc95cf1aa35a52e854585a6550 implement free energy computation
f0aed02cf7754a7577f0699fb6c1330ebd84b1f9 some cleaning
cec78e65488ad95ab572d0cab7f788d316f3b12b improve behavior of edge making function
32b41ba05abb52ea81cd1fbe858b53df09cf97a8 Fixed error message, minor change
4c899421d8f0262d014418bae0cc39449dc9c3e8 implement edge computation method and rigorous boundary checks
ac34942e1010e1d32ff8023ca52e8d6b2c9cc874 make customizable neighborhood system in segmentation
b7f00ea81d16ec21be9b28039f1ffa3267e1d3fc generalizing segmentation code to multichannel problems
5b3511525f51d4b2d434696ae37f74d88ce0b2a0 more general MRF model
4aac1573b55c94b17cf81eef423947cc35f038a3 TST: add test for step_function and block names
b1aeb67a5addfd9a018ccffcaf8b00567aa70eaf NF: adding a name to blocks
cfdad830baaf77620b4ba7650131c99b4344333b RF+TEST: refactor, fix, test isestimable function
9c888a4a0a38cca6b0555c5871bb64130cd8525d Merge pull request #197 from matthew-brett/rollimg-fix0-default
0a53a5e5a0b89b74985b579d63ee8074cc4ee335 further changes to resampling suggested by Matthew
38eca2306dfa952cef6bfafbb12e40fda7a6359c rewrote tests for the resampling function
7ecc4485a4babf7f2c81aa10c2370deb09d3732c fix resampling function
e21ebda4eba01d403a4b93b45778d230ec273ccc fixed tests
b7506ef5667f099d2d75ef45ac99e38ccc93811b rename methods
db314d1af3a184b01274d098e2ae11c32f6b80f4 improve masking strategy in image registration
d6c0f3b90abc5625582ff2d96e6f28f548599210 rename subsample method in HistogramRegistration object as reduce_from_image
30d6113ce6b4e5756d0f4e36f6ae1915753c96da remove useless input argument
0c149b7381641c0ae461f77056354d2a012027b2 RF: use fix0 to fix rollimg axis errors for TR=0
c69075c4c61594f2afb99677e28f89d69e6aa283 BF+TEST: adapt code to fix in rollimg behavior
fa07e6ce90349179c2999ba8acb1328e4a65fbe9 BF+TEST: rollimg start wrong when start > axis
5fce0f38e1e399706c723ab2117cce37fc3b2984 BF: remove use of nibabel and nipy image get_shape
b1be9a3bfb826858e08713ab4a0248efccab7742 DOC: adapt docstrings to numpy standard
as you see -- (nearly) all of them are fixes, compatibility improvements,
extended testing and docs changes. Upstream worked quite hard to polish that
~rc into a stable release. Judging from the commits post release, it is quite
stable -- there are only docs-related fixes after the release. Altogether, as
I have mentioned in the original request, with all the unittesting at build
time and growing popularity of the package [1] without any bug reports being
filed, it is in an exceptional shape worth inclusion into Debian wheezy instead
of the release-candidate it carries now. I really do not think that there
would be any cons- from such action -- only pros- of getting stable version to
be maintained through the years of wheezy existence. Due to their number,
cherry-picking of individual fixes would IMHO be not worthwhile and would
leave the -doc package in some funky state as well since there were tremendous
improvements in the documentation.
[1] http://qa.debian.org/popcon.php?package=nipy
--
Yaroslav O. Halchenko
Postdoctoral Fellow, Department of Psychological and Brain Sciences
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419
WWW: http://www.linkedin.com/in/yarik
Reply to: