Injection placement and efficiency calculation code for PyGRB Suggestions

[rpfisher]

In order to solve the problems with inaccurate efficiency calculations, I have two ideas that I think should be implemented in the O4 pipeline (these may already be in or not, I need to learn about it)

1. Injections that are used in the efficiency calculation should not be placed in data that has been gated/autogated
2. Injections that are missed due to failing the X^2 tests should not be counted in the calculation of the exclusion distance and the efficiency curves:
   a. The calculation of the exclusion distance is meant to indicate the effective range to which we could detect a GW. This should be a continuous drop off in efficiency as the distance is increased. Due to the design of the search, unless we were to request deglitched frames, whether or not there is a glitch overlaying a given injection (or onsource) is effectively a binary selection effect, not continuous. This does not match the goal of the calculation if we treat these times as if they are part of the continuum of the sensitivity of the search. This change would not mean we are handling the injections differently than the onsource: If there is a glitch in the onsource, we would cut and have no detection regardless of the range, unless the event is extremely loud, in which case it will have been detected in low latency already.

[rpfisher]

@titodalcanton @pannarale @jakeb245 I think this could be added to the board of things to include in the O4 pipeline, but I think we should get your thoughts on my reasoning.

Also, I have no permissions to label these issues or add things to the board. I'd like to get them if possible, even though Jacob's doing all the work for now. :)