[Utilities] improve performance of canonical

[odow]

Initial analysis of jump-dev/JuMP.jl#2817

No changes to the memory allocations. But canonical gets called an awful lot, so this is a critical path. The current implementation was very generic with dynamic dispatch. But the functions aren't used outside MOI, or anywhere else in MOI:
https://juliahub.com/ui/Search?q=sort_and_compress%21&type=code
https://juliahub.com/ui/Search?q=is_strictly_sorted&type=code

Script:

using MathOptInterface
const MOI = MathOptInterface

function create_model(I, T = 10000)
    model = MOI.Utilities.Model{Float64}()
    v = MOI.VariableIndex[]
    for _ in 1:I
        x = MOI.add_variables(model, T)
        MOI.add_constraint.(model, x, MOI.GreaterThan(0.0))
        MOI.add_constraint.(model, x, MOI.LessThan(100.0))
        for t in 2:T
            f = MOI.ScalarAffineFunction(
                MOI.ScalarAffineTerm.([1.0, -1.0], [x[t], x[t-1]]),
                0.0,
            )
            MOI.add_constraint(model, f, MOI.GreaterThan(-10.0))
            MOI.add_constraint(model, f, MOI.LessThan(10.0))
        end
        append!(v, x)
    end
    g = MOI.ScalarAffineFunction(MOI.ScalarAffineTerm.(1.0, v), 0.0)
    MOI.set(model, MOI.ObjectiveSense(), MOI.MIN_SENSE)
    MOI.set(model, MOI.ObjectiveFunction{typeof(g)}(), g)
    return model
end

Before

julia> @benchmark create_model(100, 1000)
BenchmarkTools.Trial: 35 samples with 1 evaluation.
 Range (min … max):  113.697 ms … 176.008 ms  ┊ GC (min … max):  0.00% … 31.68%
 Time  (median):     146.131 ms               ┊ GC (median):    21.21%
 Time  (mean ± σ):   145.134 ms ±  15.243 ms  ┊ GC (mean ± σ):  20.95% ±  9.27%

  ▄                         █▄  ▁▁ ▁  ▁▄        ▁             ▁  
  █▆▁▁▁▆▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▆▁██▆▁██▆█▆▆██▁▁▁▆▁▁▆▁█▁▆▁▁▁▁▁▁▁▁▁▁▁█ ▁
  114 ms           Histogram: frequency by time          176 ms <

 Memory estimate: 77.28 MiB, allocs estimate: 799632.

After

julia> @benchmark create_model(100, 1000)
BenchmarkTools.Trial: 74 samples with 1 evaluation.
 Range (min … max):  48.633 ms … 113.591 ms  ┊ GC (min … max):  0.00% … 32.06%
 Time  (median):     67.296 ms               ┊ GC (median):    25.64%
 Time  (mean ± σ):   68.186 ms ±  10.339 ms  ┊ GC (mean ± σ):  24.84% ± 10.27%

        ▁             ▆ ▆▃█▁▃▆▁ ▃    ▁                          
  ▇▄▄▁▇▁█▁▄▁▁▁▁▁▁▄▁▇▄▇█▇███████▁█▄▄▇▄█▁▁▄▄▁▁▁▁▇▄▁▁▁▄▁▄▁▁▁▁▁▁▁▄ ▁
  48.6 ms         Histogram: frequency by time         94.1 ms <

 Memory estimate: 77.28 MiB, allocs estimate: 799632.