Refactor `IIF.Mixture` as direct Density

[dehann]

Another question here is whether we should just update Mixture to be a regular <:SamplableBelief density rather than quasi factor-density structure it is now. i.e. change everything so that PriorPose2(Mixture((.., ...), (0.5;0.5))). Should never have gone down the road of Mixture(PriorPose2, (..,...),(0.5,0.5)) as it is now.

Originally posted by @dehann in #1407 (comment)

[dehann]

From later duplicate: #1504

Project Epic board

https://github.com/orgs/JuliaRobotics/projects/5/views/1

Discussed in JuliaRobotics/Caesar.jl#808

Continue chatting on Discussion page, this Issue is just boiler plate link to Project Epics.
( we can unlock this conversation if necessary ).