Assumptions in pixels_from_track.py

[jaafar-chakrani]

The functions in pixels_from_track.py do not use all the information from the yaml geometry files. They only use the pixel_pitch value and assume that all tiles (and pixels) are contiguous. A more general implementation should be using the tile centers as well.

This issue was observed because of a bug in the pixel pitch of the Module 2 geometry file. This geometry file mistakenly used a small pitch such that the number of pixels per tile side times the pitch is smaller than the distance between the tile centers, which leads to gaps (dead region) around each tile. When this issue was found, it was observed that this was not reflected in the output of larnd-sim. See this presentation for a sketch.

#218 addressed this by updating the pixel pitch for Module 2 to the correct value such that there are no gaps around the tile. In this specific case, the existing implementation works. However, for a more general implementation, I think that most of the functions in pixels_from_track.py need to be changed, with a particular care regarding get_neighboring_pixels since gaps between tiles would impact the list of neighboring pixels.