Closed dnjohnstone closed 4 years ago
Yep, will get a local copy of the branch, review & then add changes + make PR
Which notebook, the Basic SED Analysis - GaAs Nanowire? I haven't run through the full notebook, but calling s.change_dtype(float)
changes data type to Signal2D
. I use s.data = s.data.astype(float)
myself. Might that be the cause? Using this instead and calling decompositon()
on it now maintains the ElectronDiffraction2D
class in my case.
Which notebook, the Basic SED Analysis - GaAs Nanowire? I haven't run through the full notebook, but calling
s.change_dtype(float)
changes data type toSignal2D
. I uses.data = s.data.astype(float)
myself. Might that be the cause? Using this instead and callingdecompositon()
on it now maintains theElectronDiffraction2D
class in my case.
Ah - yes, that's almost certainly the cause. Thank you!
I guess that's a 'bug' with .change_dtype in pyxem then - which is probably to be resolved in a similar way to the previous decomposition bug?
@pc494 - I couldn't work out quickly how to update the strain demo for your new changes. Please could you do that and make a PR to this branch?
Also it seems to me the update to the decomposition methods to try and get class assignment correct from @hwagit didn't work in the case used in the demo, which I don't quite understand....