Albert Y. Shih
Albert Y. Shih
Now tested: I don't encounter any conflicts with SunPy's existing code. The changes in this PR are cleaner logic, but I'll point out that the performance improvements are very small....
Confirmed: the inheritance-diagram links do get fixed
Unfortunately, no. The Sphinx PR (sphinx-doc/sphinx#10614) hasn't yet been merged, so the fixes aren't in any release.
It looks like these methods as currently written would be inaccurate when: - The pixel scales at the reference pixel are different from the pixel scales within the image extent...
Although rectangular pixels are not common, I think they are more common in simulated data and more common in heliophysics than astrophysics. For example, I was Co-I an instrument concept...
Re-iterating [my comment above](https://github.com/astropy/astropy/pull/12102#issuecomment-1056997224): > It looks like these methods as currently written would be inaccurate when: > > * The pixel scales at the reference pixel are different from...
> > Re-iterating [my comment above](https://github.com/astropy/astropy/pull/12102#issuecomment-1056997224): > > Thanks for the reminder. I've added a note in the docstrings with [a1d98b0](https://github.com/astropy/astropy/commit/a1d98b0ce7401f4ef6aff819fa093c20d40a161c). Thanks!
This appears exactly like a bug that graphviz had a while back (see sunpy/sunpy#3545 and https://gitlab.com/graphviz/graphviz/issues/1605): the scaling factor was accidentally inverted (enlarging instead of shrinking). Perhaps the graphviz bug...
Ah, Astropy docs are being built with graphviz 2.43.0, which still has the bug
I don't know. Apparently SunPy docs are also being built with graphviz 2.43.0 on RTD, which explains sunpy/sunpy#5857