pmedian returns a PMedianResult object rather than a Dict{String, Any}

CHANGELOG.md

@@ -1,5 +1,6 @@
 ### 0.1.8 (Upcoming release) 
 
+- pmedian now returns a PMedianResult object rather than a Dict{String, Any}.
 
 ### 0.1.7
 

src/OperationsResearchModels.jl

@@ -40,7 +40,7 @@ import .MaximumFlow: MaximumFlowResult
 import .Assignment: AssignmentProblem, AssignmentResult
 import .Game: game, GameResult
 import .MinimumSpanningTree: hasloop, mst, MstResult
-import .PMedian: pmedian, pmedian_with_distances
+import .PMedian: pmedian, pmedian_with_distances, PMedianResult
 import .CPM: cpm, CpmActivity, earliestfinishtime, longestactivity, CpmResult
 import .CPM: pert, PertActivity, PertResult
 import .Latex: latex
@@ -51,13 +51,14 @@ export ShortestPathProblem, MaximumFlowProblem
 export AssignmentProblem, AssignmentResult
 export game, GameResult
 export hasloop, mst, MstResult
-export pmedian, pmedian_with_distances
+export pmedian, pmedian_with_distances, PMedianResult
 export cpm, CpmActivity, earliestfinishtime, longestactivity, CpmResult
 export pert, PertActivity, PertResult
 export Simplex
 export Utility
 export latex
 
+
 """
     solve(t)
 

src/pmedian.jl

@@ -1,29 +1,57 @@
 module PMedian
 
 export pmedian
+export pmedian_with_distances
 
 using JuMP, HiGHS
 
+"""
+    PMedianResult
+
+# Fields
+
+- `centers::Vector`: Indices of selected centers.
+- `model::JuMP.Model`: JuMP model.
+- `objective::Float64`: Objective value.
+- `z::Matrix`: Binary matrix of assignments. If z[i, j] = 1 then ith point is connected to the jth point.
+- `y::Vector`: Binary vector. If y[i] is 1 then ith point is a depot.
+"""
+struct PMedianResult
+    centers::Vector
+    model::JuMP.Model
+    objective::Float64
+    z::Matrix
+    y::Vector
+end
+
+
+
+
 function euclidean(u, v)::Float64
     d = (u .- v)
     return (d .* d) |> sum |> sqrt
 end
 
 
+
+
 """
 
     pmedian(data, ncenters)
 
 # Arguments 
+
 - `data::Matrix`: Coordinates of locations 
 - `ncenters::Int`: Number of centers 
 
 # Descriptions 
+
 The function calculates Euclidean distances between all possible rows of the matrix data. 
 `ncenters` locations are then selected that minimizes the total distances to the nearest rows. 
 
 # Output 
-- `Dict{String, Any}`: The dictionary object that holds the results. 
+
+- `PMedianResult`: PMedianResult object. 
 
 # Example 
 
@@ -36,17 +64,20 @@ julia> data3 = rand(10, 2) .+ 100;
 
 julia> data = vcat(data1, data2, data3);
 
-julia> result = pmedian(data, 3)
-Dict{String, Any} with 5 entries:
-  "centers"   => [6, 13, 25]
-  "model"     => A JuMP Model…
-  "objective" => 10.5921
-  "z"         => [-0.0 -0.0 … 0.0 0.0; -0.0 -0.0 … 0.0 0.0; … ; 0.0 0.0 … -0.0 0.0; 0.0 0.0 … 0.0 -0.0]
-  "y"         => [0.0, -0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, -0.0, 0.0  …  0.0, 0.0, 0.0, -0.0, 1.0, 0.0, -0.0,…
+julia> result = pmedian(data, 3);
+
+julia> result.centers
+3-element Vector{Int64}:
+  1
+ 16
+ 21
+
+ julia> result.objective
+ 11.531012240599605
 ``` 
 
 """
-function pmedian(data::Matrix, ncenters::Int)
+function pmedian(data::Matrix, ncenters::Int)::PMedianResult
 
     n, p = size(data)
 
@@ -77,9 +108,9 @@ end
 `ncenters` locations are selected that minimizes the total distances to the nearest rows. 
 
 # Output 
-- `Dict{String, Any}`: The dictionary object that holds the results. 
+- `PMedianResult`: PMedianResult object. 
 """
-function pmedian_with_distances(distancematrix::Matrix, ncenters::Int)
+function pmedian_with_distances(distancematrix::Matrix, ncenters::Int)::PMedianResult
 
     n, _ = size(distancematrix)
 
@@ -108,12 +139,12 @@ function pmedian_with_distances(distancematrix::Matrix, ncenters::Int)
 
     optimize!(model)
 
-    return Dict(
-        "z" => value.(z),
-        "y" => value.(y),
-        "centers" => findall(x -> x == 1, value.(y)),
-        "objective" => JuMP.objective_value(model),
-        "model" => model,
+    return PMedianResult(
+        findall(x -> x == 1, value.(y)),    
+        model,
+        JuMP.objective_value(model),
+        value.(z),
+        value.(y),
     )
 end
 

test/testpmedian.jl

@@ -11,9 +11,9 @@
         ]
 
         result = pmedian(coords, 2)
-        y = result["y"]
+        y = result.y
 
-        @test sort(result["centers"]) == [2, 6]
+        @test sort(result.centers) == [2, 6]
 
         @test iszero(y[1])
         @test isone(y[2])
@@ -30,8 +30,8 @@
 
         result = pmedian_with_distances(distance_matrix, number_of_depots)
 
-        @test sort(result["centers"]) == [1, 2]
-        @test result["objective"] == 5.0
-        @test result["z"] ≈ [1 0 0 0; 0 1 0 0; 1 0 0 0; 1 0 0 0]
+        @test sort(result.centers) == [1, 2]
+        @test result.objective == 5.0
+        @test result.z ≈ [1 0 0 0; 0 1 0 0; 1 0 0 0; 1 0 0 0]
     end 
 end