Due to data collection strategies, it would be more flexible if we did not assume "junk" frames always are at the start of an image stack - but also allow for them to be at the back, or a combination of front and back.
Second Option (or 1+):
Ideally you would also have the option of brining in an array as a mask and only use frame/omega wedges you choose for the series. I believe a version of this was implemented somewhere for when we wanted to mask out frames with the intermittent beam drops.
First Option:
Second Option (or 1+):