Merge pull request JuliaStats#286 from lwabeke/master

dbscan(): drop num points > num dimensions check

it is not necessary

Author: alyst

src/dbscan.jl

@@ -101,7 +101,6 @@ function dbscan(points::AbstractMatrix, radius::Real;
     if metric !== nothing
         # points are point coordinates
         dim, num_points = size(points)
-        num_points <= dim && throw(ArgumentError("points has $dim rows and $num_points columns. Must be a D x N matric with D < N"))
         kdtree = KDTree(points, metric; nntree_kwargs...)
         data = (kdtree, points)
     else

test/affprop.jl

@@ -6,7 +6,6 @@ using Clustering
 using LinearAlgebra
 using Random, StableRNGs
 using Statistics
-include("test_helpers.jl")
 
 @testset "affinityprop() (affinity propagation)" begin
 

test/dbscan.jl

@@ -1,7 +1,6 @@
 using Test
 using Clustering
 using Distances
-include("test_helpers.jl")
 
 @testset "dbscan() (DBSCAN clustering)" begin
 
@@ -14,6 +13,51 @@ include("test_helpers.jl")
     @test @inferred(dbscan(randn(2, 2), 0.5, metric=nothing, min_neighbors=1)) isa DbscanResult
 end
 
+@testset "Simple 2D tests" begin
+    X = [10.0  0.0   10.5
+          0.0  10.0  0.1]
+
+    @testset "n=3 samples" begin
+        X3 = X
+
+        R = dbscan(X3, 20)
+        @test nclusters(R) == 1
+
+        R = dbscan(X3, 1.0)
+        @test nclusters(R) == 2
+
+        R = dbscan(X3, 0.1)
+        @test nclusters(R) == 3
+    end
+
+    @testset "n=2 samples" begin
+        X2 = X[:, 1:2]
+
+        R = dbscan(X2, 20)
+        @test nclusters(R) == 1
+
+        R = dbscan(X2, 1.0)
+        @test nclusters(R) == 2
+    end
+
+    @testset "n=1 samples" begin
+        X1 = X[:, 1:1]
+
+        R = dbscan(X1, 20)
+        @test nclusters(R) == 1
+
+        R = dbscan(X1, 1.0)
+        @test nclusters(R) == 1
+    end
+
+    @testset "n=0 samples" begin
+        X0 = X[:, 1:0]
+
+        R = dbscan(X0, 20)
+        @test nclusters(R) == 0
+    end
+end
+
 @testset "clustering synthetic data with 3 clusters" begin
     Random.seed!(34568)
 

test/kmedoids.jl

@@ -1,7 +1,6 @@
 using Test
 using Distances
 using Clustering
-include("test_helpers.jl")
 
 @testset "kmedoids() (k-medoids)" begin
 

test/runtests.jl

@@ -6,6 +6,8 @@ using SparseArrays
 using StableRNGs
 using Statistics
 
+include("test_helpers.jl")
+
 tests = ["seeding",
          "kmeans",
          "kmedoids",