allow single config in solver (#92)

src/interfaces.jl

@@ -489,16 +489,33 @@ size_all_negative(::SpinGlass) = false
 size_all_positive(::SpinGlass) = false
 
 # NOTE: `findmin` and `findmax` are required by `ProblemReductions.jl`
+"""
+    GTNSolver(; optimizer=TreeSA(), single=false, usecuda=false, T=Float64)
+
+A generic tensor network based backend for the `findbest`, `findmin` and `findmax` interfaces in `ProblemReductions.jl`.
+
+Keyword arguments
+-------------------------------------
+* `optimizer` is the optimizer for the tensor network contraction.
+* `single` is a switch to return single solution instead of all solutions.
+* `usecuda` is a switch to use CUDA (when applicable), user need to call statement `using CUDA` before turning on this switch.
+* `T` is the "base" element type, sometimes can be used to reduce the memory cost.
+"""
 Base.@kwdef struct GTNSolver
     optimizer::OMEinsum.CodeOptimizer = TreeSA()
+    single::Bool = false
     usecuda::Bool = false
     T::Type = Float64
 end
-function Base.findmin(problem::AbstractProblem, solver::GTNSolver)
-    res = collect(solve(GenericTensorNetwork(problem; optimizer=solver.optimizer), ConfigsMin(; tree_storage=true); usecuda=solver.usecuda, T=solver.T)[].c)
-    return map(x -> ProblemReductions.id_to_config(problem, Int.(x) .+ 1), res)
-end
-function Base.findmax(problem::AbstractProblem, solver::GTNSolver)
-    res = collect(solve(GenericTensorNetwork(problem; optimizer=solver.optimizer), ConfigsMax(; tree_storage=true); usecuda=solver.usecuda, T=solver.T)[].c)
-    return map(x -> ProblemReductions.id_to_config(problem, Int.(x) .+ 1), res)
-end
+for (PROP, SPROP, SOLVER) in [
+        (:ConfigsMin, :SingleConfigMin, :findmin), (:ConfigsMax, :SingleConfigMax, :findmax)
+    ]
+    @eval function Base.$(SOLVER)(problem::AbstractProblem, solver::GTNSolver)
+        if solver.single
+            res = [solve(GenericTensorNetwork(problem; optimizer=solver.optimizer), $(SPROP)(); usecuda=solver.usecuda, T=solver.T)[].c.data]
+        else
+            res = collect(solve(GenericTensorNetwork(problem; optimizer=solver.optimizer), $(PROP)(; tree_storage=true); usecuda=solver.usecuda, T=solver.T)[].c)
+        end
+        return map(x -> ProblemReductions.id_to_config(problem, Int.(x) .+ 1), res)
+    end
+end

test/interfaces.jl

@@ -254,4 +254,8 @@ end
     solver2 = BruteForce()
     @test Set(findmin(sg, solver1)) == Set(findmin(sg, solver2))
     @test Set(findmax(sg, solver1)) == Set(findmax(sg, solver2))
+
+    solver3 = GTNSolver(; optimizer=TreeSA(ntrials=1), single=true)
+    @test findmin(sg, solver3)[] ∈ findmin(sg, solver2)
+    @test findmax(sg, solver3)[] ∈ findmax(sg, solver2)
 end