These are some updates that resolve some tasks from #115, and implements #229

Notably, slab() now uses dice() so users will not get confused when they get a deprecation message RE: slice() while using slab(). I have had several questions from users about this (essentially: how do I use dice() with slab()?)

library(aqp)
#> This is aqp 1.43
data(sp1, package = 'aqp')
depths(sp1) <- id ~ top + bottom
a.max <- slab(sp1, fm = ~ prop)
#> Note: aqp::slice() will be deprecated in aqp version 2.0
#> --> Please consider using the more efficient aqp::dice()

This PR has attempted to cover / ensure that the following are addressed:

• issues related to the fm = 0:z ~ . results in z+1 slices (fixed insofar as they affect slab(), I don't think it is fully fixed on dice() side yet)
• handled 0 thickness horizons in slab(), whether or not strict=TRUE
• fixed mismatching lengths of slab labels v.s rows in data
• added and exported slab_function() a helper method for accessing the internal slab methods provided by the package; now used in method definitions rather than internal functions (could use more documentation on arguments/patterns for custom functions, and ensuring extra arguments all works)
• added a weighted variant of the numeric/quantile method via Hmisc::wtd.quantile(), and support for custom weighted functions (such as weighted.mean())

I am sure this has not yet caught everything, but I spent some time ensuring all the existing tests worked, and added some more that stress a few other features, as well as new tests for the now-implemented weights argument. It all appears to be functioning as expected in the test cases, though I bet there are still edge cases or un-tested use cases that still have issues.