Should we deprecate lsfWho in favor of just lsf

[Affie]

From PR discussion: https://github.com/JuliaRobotics/DistributedFactorGraphs.jl/pull/705#discussion_r540034178

new name is: 'lsfTypes(fg, LinearRelative)'

@dehann There is no function like that. Only lsfTypes(dfg) -> Return Vector{Symbol} of all unique factor types in factor graph.

Did you perhaps mean just lsf as the test just above that?

Should we deprecate lsfWho in favour of just lsf

In this case lsfWho and lsf have almost identical functionality with different implementations. I would say we keep on only lsf for example lsf(fg, LinearRelative)

It can work if there is not a need for lsf(fg, :LinearRelative)