Problem with matching by set distance

I'd like some help in figuring out what's going wrong in a synthetic example I created. I'm trying to apply global_continuation using AttractorsViaFeaturizing, first with a matching by set distance. The system's real behavior is as described in the left column of the figure below.  The first attractor, which starts at position 0, moves to -5 for params 3 and 3.5, but moves back to 0 at param>=4. Matching by centroid misses this behavior, and 'incorrectly' matches the attractors, as shown in the right column. I had run these results before the reformulation of `global_continuation` and everything. Now I am getting an error:  There are missing values of the basin fractions. As far as I understand, the problem is in the `rmap` that you get. This simple example shows what's happening: ```julia fs = Dict(4 => 0.4097706879361914, 2 => 0.19042871385842472, 3 => 0.18843469591226322, 1 => 0.21136590229312063) rmap = Dict(2 => 3) #error # rmap = Dict(2 => 3, 3=>2) #fine swap_dict_keys!(fs, rmap) ``` The rmap for those parameters is missing one entry (e.g. 3=>2). This makes the swap function drop one of the keys. I am not sure why this is happening, but I know that removing line 112 of `sssdistance.jl`, which is ```julia filter!(p -> p.first ≠ p.second, rmap) ``` solves it, because then rmap has 2=>3 and 3=>2. I don't know why, though. Could you please confirm this is not just on my machine? I've updated everything now. If you don't want to run the continuation, it all boils down to these guys: ```julia pcurve = [[1 => 1.0], [1 => 1.5], [1 => 2.0], [1 => 2.5], [1 => 3.0], [1 => 3.5], [1 => 4.0], [1 => 4.5], [1 => 5.0]] attractors_cont_simple = Dict{Int64, SVector{1, Float64}}[Dict(1 => [0.0]), Dict(1 => [0.0]), Dict(2 => [2.0], 1 => [0.0]), Dict(2 => [0.0], 1 => [2.0]), Dict(2 => [2.0], 3 => [4.0], 1 => [-5.0]), Dict(2 => [4.0], 3 => [-5.0], 1 => [2.0]), Dict(4 => [6.0], 2 => [0.0], 3 => [2.0], 1 => [4.0]), Dict(4 => [4.0], 2 => [0.0], 3 => [2.0], 1 => [6.0]), Dict(5 => [8.0], 4 => [6.0], 2 => [0.0], 3 => [2.0], 1 => [4.0])] attractors_cont = [Dict(k=>StateSpaceSet(Vector(v)) for (k,v) in atts) for atts in attractors_cont_simple] fractions_cont = [Dict(1 => 1.0), Dict(1 => 1.0), Dict(2 => 0.8091908091908092, 1 => 0.19080919080919082), Dict(2 => 0.1906187624750499, 1 => 0.8093812375249501), Dict(2 => 0.18862275449101795, 3 => 0.6207584830339321, 1 => 0.1906187624750499), Dict(2 => 0.6211365902293121, 3 => 0.19042871385842472, 1 => 0.18843469591226322), Dict(4 => 0.4097706879361914, 2 => 0.19042871385842472, 3 => 0.18843469591226322, 1 => 0.21136590229312063), Dict(4 => 0.21115537848605578, 2 => 0.1902390438247012, 3 => 0.18824701195219123, 1 => 0.4103585657370518), Dict(5 => 0.21613545816733068, 4 => 0.1942231075697211, 2 => 0.1902390438247012, 3 => 0.18824701195219123, 1 => 0.21115537848605578)] rmaps = match_sequentially!( attractors_cont, MatchBySSSetDistance(); pcurve, ds = Attractors.referenced_dynamical_system(mapper) ) match_sequentially!(fractions_cont, rmaps) ``` Note that e.g. `fractions_cont[7]` is now incomplete. The problem seems to be with retract_keys somehow. Thanks!