added the count analysis for the new egraph

continue fixing the sampling, not happy with the solution

fix fixpoint lamdba searcher

integrated sampler into the other logic

factored out find_min functions and iterators

refactor choice_iter into extract

sampling for the egraph instead of enumerating

fix for the tree iterator, now much worse runtime

better count impl

explicit ignore-labels flag

added the structural diff option

pretty match

minor cleanup

also fixed 0.0 to 0.0f in all the jsons

fix parsing of floats/doubles