Closed alfiopuglisi closed 3 years ago
commit cd797822a0899f07ce97941305b776de8f5da8b6 forces numpy > 1.17 that fixes these checks. On my computer with numpy=1.18 fft units are properly managed, with 1.16 they are not.
Il giorno mar 28 apr 2020 alle ore 21:37 Alfio Puglisi < notifications@github.com> ha scritto:
in discrete_fourier_transform: fft direct and inverse do not work with astropy units and numpy < 1.18, because units are not conserved. We need a way of checking the numpy version, or write code that works in both cases (sometimes not easy).
In various points in the library I've put unit checks that maybe were only needed on my computer, which was running numpy == 1.16 up to five minutes ago.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ArcetriAdaptiveOptics/arte/issues/16, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB544W5GNHXQHY4LNH7VPBDRO4V7BANCNFSM4MTDSCSA .
I will close this, as it looks like numpy version is now properly managed
in discrete_fourier_transform: fft direct and inverse do not work with astropy units and numpy < 1.18, because units are not conserved. We need a way of checking the numpy version, or write code that works in both cases (sometimes not easy).
In various points in the library I've put unit checks that maybe were only needed on my computer, which was running numpy == 1.16 up to five minutes ago.