feat: expand the dataset name regex to allow periods

Author: mortenpi

src/DataSets.jl

@@ -106,8 +106,9 @@ end
 const DATASET_NAME_REGEX_STRING = raw"""
 [[:alnum:]]
 (?:
-    [-[:alnum:]_]     |
-    / (?=[[:alnum:]])
+    [-[:alnum:]_]      |
+    \.(?=[[:alnum:]])  |
+    \/ (?=[[:alnum:]])
 )*
 """
 const DATASET_NAME_REGEX = Regex("^\n$(DATASET_NAME_REGEX_STRING)\n\$", "x")

test/dataset-names-invalid.txt

@@ -1,9 +1,23 @@
 a b
-a.b
 a/b/
 a//b
 /a/b
 a/-
 a/ _/b
 a/-a
 a/-1
+.a
+..a
+a.
+a..
+.a.
+a..b
+.abc
+abc.
+abc/.def
+abc/def.
+a./b
+a.-
+_._
+a._b
+a.-b

test/dataset-names-valid.txt

@@ -14,3 +14,11 @@ a/1
 1-2-3
 x_-__
 a---
+a.b
+a.b
+abc.def
+abc/def.ghi
+abc-def.ghi_jkl
+a.b.c
+a_.c
+foo__-.csv

test/runtests.jl

@@ -105,19 +105,47 @@ end
 @testset "Data set name parsing" begin
     @testset "Valid names" begin
         valid_names = load_list("dataset-names-valid.txt")
-        @test length(valid_names) == 16
+        @test !isempty(valid_names)
         @testset "Valid name: $name" for name in valid_names
             @test DataSets.check_dataset_name(name) === nothing
             @test DataSets._split_dataspec(name) == (name, nothing, nothing)
+            # Also test that the name is still valid when it appears as part of
+            # a path elements.
+            let path_name = "foo/$(name)"
+                @test DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (path_name, nothing, nothing)
+            end
+            let path_name = "$(name)/foo"
+                @test DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (path_name, nothing, nothing)
+            end
+            let path_name = "foo/$(name)/bar"
+                @test DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (path_name, nothing, nothing)
+            end
         end
     end
 
     @testset "Invalid names" begin
         invalid_names = load_list("dataset-names-invalid.txt")
-        @test length(invalid_names) == 9
+        @test !isempty(invalid_names)
         @testset "Invalid name: $name" for name in invalid_names
             @test_throws ErrorException DataSets.check_dataset_name(name)
             @test DataSets._split_dataspec(name) == (nothing, nothing, nothing)
+            # Also test that the name is still invalid when it appears as part of
+            # a path elements.
+            let path_name = "foo/$(name)"
+                @test_throws ErrorException DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (nothing, nothing, nothing)
+            end
+            let path_name = "$(name)/foo"
+                @test_throws ErrorException DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (nothing, nothing, nothing)
+            end
+            let path_name = "foo/$(name)/bar"
+                @test_throws ErrorException DataSets.check_dataset_name(path_name) === nothing
+                @test DataSets._split_dataspec(path_name) == (nothing, nothing, nothing)
+            end
         end
     end
 end