This caught me out while updating dodal to opyhd-async: while the numpy datatype may be useful for deciding how much space to allocate for a given stack of frames, it is decidedly not part of the shape of the frame.
For a currently not ADBaseIO detector (the TetrAMM monitor) this tripped me up as the datatype was not being passed as part of the shape and hence it was looking for a datatype with key 1000.
This caught me out while updating dodal to opyhd-async: while the numpy datatype may be useful for deciding how much space to allocate for a given stack of frames, it is decidedly not part of the shape of the frame.
For a currently not ADBaseIO detector (the TetrAMM monitor) this tripped me up as the datatype was not being passed as part of the shape and hence it was looking for a datatype with key
1000
.Closes #554