[DOC] Correct bounds for v in RectSphereBivariateSpline #6612
Conversation
| @@ -1529,7 +1529,7 @@ class RectSphereBivariateSpline(SphereBivariateSpline): | |||
| (0, pi). | |||
| v : array_like | |||
| 1-D array of longitude coordinates in strictly ascending order. | |||
| Coordinates must be given in radians, and must lie within (0, 2pi). | |||
| Coordinates must be given in radians, and must lie within (-pi, pi). | |||
There was a problem hiding this comment.
This isn't quite right--the FITPACK documentation requires that
-pi <= v[0] < piv[i] < v[i+1](the strictly ascending bit)v[-1] <= v[0] + 2*pi.
(Changed the indexing from the documentation for python conventions.) So for example in #6556 the test case
>>> u = np.linspace(1, 3)
>>> v = np.linspace(np.pi - 1e-6, 6)
>>> r = np.zeros([u.size, v.size])
>>> RectSphereBivariateSpline(u, v, r)works fine even though 6 is not in the interval (-pi, pi).
rgommers
added
Documentation
|
@sriki18 I wonder if you're planning to continue to working on this? No pressure, just don't want it to slip into the mists of time. |
|
Hi, Sorry I didn't catch your earlier comment. I am going to have a busy week and will have a go at it if I find the time. I should definitely find time by Friday. If you would prefer to work on it in the mean time, please let me know. Thanks! |
|
Sounds great. No rush, just wanted to make sure you were still interested. |
|
Working on this at the moment. |
|
superceded by #6665 |
This is a fix for #6556 . I changed the SciPy documentation to reflect the actual functionality of fftpack.