x_max to h_max conversion in AtmosphereDensityProfile

[kosack]

Please describe the use case that requires this feature.

Currently we have atmprof.line_of_sight_integral(h, zenith) that converts h (distance in m) to x (slant depth in g cm-2). For the ImPACT analysis, we need the inverse transform as well.

Describe the solution you'd like

Implement derivative() functions for all AtmosphereDensityProfiles, similar to the existing integral() and provide a top-level function to do the line_of_sight_derivative()

needed for #2305

[maxnoe]

Is it really the derivative we need?

The inverse function of an integral is not the derivative...

[kosack]

true, not the best name

[kosack]

Could just rename both of them distance_to_slant_depth() and slant_depth_to_distance()

[kosack]

Also, we should consider renaming h_max to d_max or shower_max_distance or something, to be clear that it is not the height A.S.L., but really the distance to the shower max along the shower direction. That way it's clear that this and x_max are the same measure but expressed in different units.

[maxnoe]

This was fixed: #2422