WIP: Rework local coordinate systems

This MR aims to change the local detector coordinate systems to be more intuitive. It continues the discussion which evolved in !84 (merged).

@nurnberg and me have realized that we have a problem with local coordinates systems. Currently, the local frame is defined on the implant side of the sensor, starting with pixel (0, 0) on the lower left corner. However, with a hybrid detector this would mean, that on the readout chip (flipped along rows) the pixel in the lower left corner (which is 0,0 in the real world) turns up as (0, npixelsX) instead.

There are a couple of possibilities to fix this:

1. Add the translation (flip along rows) in the SimpleTransfer module - it is supposed to map sensor pixels to chip pixels anyway
2. Change the coordinate system to be defined by the readout chip. This could make sense when thinking a bit more careful about monolithic sensors. Currently, they are facing "away" from the beam, i.e. the backside of the sensor is facing upstream (as is correct with hybrid detectors) and the implants face downstream. This should be reversed, since the chip is glued to the carrier PCB always with the non-processed side (we have wire bond pads on the implant side!)

Any ideas or contributions?

@kwolters @dhynds