backporting scipy.stats fixes
I added 3 changesets in trunk, that should be backported to 0.9 7030 is a trivial fix to avoid an exception 7028 and 7029 correct the integration bounds in the expect methods and have, I think, sufficient test cases. Thanks, Josef
On Fri, Jan 14, 2011 at 4:31 PM, <josef.pktd@gmail.com> wrote:
I added 3 changesets in trunk, that should be backported to 0.9
7030 is a trivial fix to avoid an exception
7028 and 7029 correct the integration bounds in the expect methods and have, I think, sufficient test cases.
7031, 7032, 7033 are cases that only show up on the first call to the distribution and are difficult to test. 7031, 7032 add protection if there are no valid arguments, initially reported on stackoverflow 7032, 7033 add argcheck call to moment method and fixes ticket 934 for distributions with support shifted by args, and also protects against the no valid args case They don't break any tests, but I only checked them on the commandline for a few cases. I think, it's pretty safe to backport them also.
python -c "from scipy import stats; print stats.wrapcauchy.moment(2,1.)" 1.#QNAN python -c "from scipy import stats; print stats.poisson.moment(2,-2)" 1.#QNAN python -c "from scipy import stats; print stats.truncnorm.moment(2,-3,3)" 0.973336924663
Josef
Thanks,
Josef
On Sat, Jan 15, 2011 at 10:25 AM, <josef.pktd@gmail.com> wrote:
On Fri, Jan 14, 2011 at 4:31 PM, <josef.pktd@gmail.com> wrote:
I added 3 changesets in trunk, that should be backported to 0.9
7030 is a trivial fix to avoid an exception
7028 and 7029 correct the integration bounds in the expect methods and have, I think, sufficient test cases.
7031, 7032, 7033 are cases that only show up on the first call to the distribution and are difficult to test. 7031, 7032 add protection if there are no valid arguments, initially reported on stackoverflow 7032, 7033 add argcheck call to moment method and fixes ticket 934 for distributions with support shifted by args, and also protects against the no valid args case
They don't break any tests, but I only checked them on the commandline for a few cases. I think, it's pretty safe to backport them also.
Looks pretty safe to me too, I backported them. Ralf
python -c "from scipy import stats; print stats.wrapcauchy.moment(2,1.)" 1.#QNAN python -c "from scipy import stats; print stats.poisson.moment(2,-2)" 1.#QNAN python -c "from scipy import stats; print stats.truncnorm.moment(2,-3,3)" 0.973336924663
Josef
Thanks,
Josef
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
participants (2)
-
josef.pktd＠gmail.com -
Ralf Gommers