pycbc_ifar_catalog : more choices for 'sigma' shading

[tdent]

The cumulative count significance 'sigma' shading is generally somewhat misleading for events that are not the loudest as it does not give individual event significance. It does indicate whether the cumulative event count at high FAR is consistent with noise, but that is not its main objective as an 'open box' plot.

To address this / provide (mutually exclusive?) alternatives :

• Make the cumulative sigma shading optional
• Add an option for 'N^1/2 and 2*N^1/2 Errors' curves as in pycbc_page_ifar - 2 sigma is usually sufficient to check for consistency without people taking it to mean 'significant detection'
• Add an option for vertical sigma shading over the whole plot based on only the loudest event (ie 'sigma' for individual event significance only, not cumulative) as in pycbc_page_snrifar with the --not-cumulative option - nb the shadings there are horizontal but IFAR is on the y-axis, whereas here it's on the x-axis
• Add an option for vertical sigma shading over the whole plot based on fixed IFAR levels, eg 1/(5 months) corresponding to O4 CBC significant online alerts, 1/(100 yr) corresponding to hierarchical removal threshold for loud detections