The effort to redo eotest processing using DM tooling and Gen3 has raised the issue that we don't have a way to retrieve single, unflipped amplifier images via the Gen3 butler (DM-29370), and the hard part of that is the camera geometry. The way we did that in Gen2 abuses the afw cameraGeom interface, creating a lot of potential for confusion (the detector one gets from the camera and from a detector-level image are different), and I have some ideas to address that that I think might solve a few other problems.

I'll prototype those on this ticket and send them around for comment (and probably eventually RFC).