Cubic interpolation with SciPy 1.2.0rc1
Hi, I just noticed our test suite is failing with SciPy 1.2.0rc1 on any test involving cubic interpolation through griddata. The differences aren't huge, at most 2.5%. I'm working on reducing one of the tests to a simple example. In the mean time, is there any reason to expect changes to the results from cubic interpolation? I didn't see anything in the release notes, and a quick search on GitHub didn't bring any obvious issues/PRs. Ryan -- Ryan May
ti, 2018-12-04 kello 15:08 -0700, Ryan May kirjoitti:
In the mean time, is there any reason to expect changes to the results from cubic interpolation? I didn't see anything in the release notes, and a quick search on GitHub didn't bring any obvious issues/PRs.
There was a bug for triangles on the boundary, resulting to weird choice of free degrees of freedom: https://github.com/scipy/scipy/pull/9176
Anything I should consider here for rc2, since that's delayed slightly (again)? Presumably the bug fix is higher priority than the modest behavior change? On Tue, 4 Dec 2018 at 15:12, Pauli Virtanen <pav@iki.fi> wrote:
ti, 2018-12-04 kello 15:08 -0700, Ryan May kirjoitti:
In the mean time, is there any reason to expect changes to the results from cubic interpolation? I didn't see anything in the release notes, and a quick search on GitHub didn't bring any obvious issues/PRs.
There was a bug for triangles on the boundary, resulting to weird choice of free degrees of freedom: https://github.com/scipy/scipy/pull/9176
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@python.org https://mail.python.org/mailman/listinfo/scipy-dev
The fix seems fine. It just took a little bit to find the fix Pauli linked to--it's a bit of a leap from griddata to CloughTocher2DInterpolator. Just wanted to double check before updating my tests. Ryan On Tue, Dec 4, 2018 at 4:31 PM Tyler Reddy <tyler.je.reddy@gmail.com> wrote:
Anything I should consider here for rc2, since that's delayed slightly (again)?
Presumably the bug fix is higher priority than the modest behavior change?
On Tue, 4 Dec 2018 at 15:12, Pauli Virtanen <pav@iki.fi> wrote:
ti, 2018-12-04 kello 15:08 -0700, Ryan May kirjoitti:
In the mean time, is there any reason to expect changes to the results from cubic interpolation? I didn't see anything in the release notes, and a quick search on GitHub didn't bring any obvious issues/PRs.
There was a bug for triangles on the boundary, resulting to weird choice of free degrees of freedom: https://github.com/scipy/scipy/pull/9176
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@python.org https://mail.python.org/mailman/listinfo/scipy-dev
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@python.org https://mail.python.org/mailman/listinfo/scipy-dev
-- Ryan May
participants (3)
-
Pauli Virtanen -
Ryan May -
Tyler Reddy