Validate WritingSystemIds on creation (#1230)

* Validate WritingSystemIds on creation

Special ID "default" is allowed, all other writing system IDs must be
considered valid by the SIL.WritingSystems validation logic.

* Teach AutoFaker.Generate about valid writing systems

Now when AutoFaker.Generate creates a random writing system ID, it will
be chosen from the list of all valid 2-letter writing system IDs.

* Add some writing system ID validation tests
* move faker overrides into miniLcm tests
* make test WritingSystemCodes ValidTwoLetterCodes readonly
* move MultiStringOverride.cs into auto faker helpers

---------

Co-authored-by: Kevin Hahn <[email protected]>

backend/FwLite/FwLiteProjectSync.Tests/UpdateDiffTests.cs

@@ -11,7 +11,7 @@ public class UpdateDiffTests
 {
     private readonly AutoFaker _autoFaker = new(new AutoFakerConfig()
     {
-        Overrides = [new MultiStringOverride()]
+        Overrides = [new MultiStringOverride(), new WritingSystemIdOverride()]
     });
 
     [Fact]

backend/FwLite/LcmCrdt.Tests/DataModelSnapshotTests.cs

@@ -7,14 +7,21 @@
 using Microsoft.Extensions.DependencyInjection.Extensions;
 using Microsoft.Extensions.Logging;
 using Microsoft.Extensions.Options;
+using MiniLcm.Tests.AutoFakerHelpers;
 using SIL.Harmony.Changes;
 using SIL.Harmony.Entities;
 using Soenneker.Utils.AutoBogus;
+using Soenneker.Utils.AutoBogus.Config;
 
 namespace LcmCrdt.Tests;
 
 public class DataModelSnapshotTests : IAsyncLifetime
 {
+    private static AutoFaker _faker = new AutoFaker(new AutoFakerConfig()
+    {
+        Overrides = [new MultiStringOverride(), new WritingSystemIdOverride()]
+    });
+
     protected readonly AsyncServiceScope _services;
     private readonly LcmCrdtDbContext _crdtDbContext;
     private CrdtConfig _crdtConfig;
@@ -76,15 +83,14 @@ public async Task VerifyIObjectWithIdModels()
     [Fact]
     public void VerifyIObjectWithIdsMatchAdapterGetObjectTypeName()
     {
-        var faker = new AutoFaker();
         var jsonSerializerOptions = _crdtConfig.JsonSerializerOptions;
         var types = jsonSerializerOptions.GetTypeInfo(typeof(IObjectWithId)).PolymorphismOptions?.DerivedTypes ?? [];
         using (new AssertionScope())
         {
             foreach (var jsonDerivedType in types)
             {
                 var typeDiscriminator = jsonDerivedType.TypeDiscriminator.Should().BeOfType<string>().Subject;
-                var obj = faker.Generate(jsonDerivedType.DerivedType);
+                var obj = _faker.Generate(jsonDerivedType.DerivedType);
                 new MiniLcmCrdtAdapter((IObjectWithId)obj).GetObjectTypeName().Should().Be(typeDiscriminator);
             }
         }

backend/FwLite/LcmCrdt.Tests/EntityCopyMethodTests.cs

@@ -1,5 +1,6 @@
 using SIL.Harmony.Entities;
 using LcmCrdt.Objects;
+using MiniLcm.Tests.AutoFakerHelpers;
 using SIL.Harmony;
 using Soenneker.Utils.AutoBogus;
 using Soenneker.Utils.AutoBogus.Config;
@@ -8,7 +9,11 @@ namespace LcmCrdt.Tests;
 
 public class EntityCopyMethodTests
 {
-    private readonly AutoFaker _autoFaker = new(new AutoFakerConfig());
+    private readonly AutoFaker _autoFaker = new(new AutoFakerConfig()
+    {
+        Overrides = [new MultiStringOverride(), new WritingSystemIdOverride()]
+    });
+
     public static IEnumerable<object[]> GetEntityTypes()
     {
         var crdtConfig = new CrdtConfig();

backend/FwLite/MiniLcm.Tests/AutoFakerHelpers/MultiStringOverride.cs

@@ -16,9 +16,7 @@ public override void Generate(AutoFakerOverrideContext context)
         var wordsArray = context.Faker.Random.WordsArray(1, 4);
         foreach (var word in wordsArray)
         {
-            var writingSystemId = validWs is not null
-                ? context.Faker.Random.ArrayElement(validWs)
-                : context.Faker.Random.String(2, 'a', 'z');
+            var writingSystemId = context.Faker.Random.ArrayElement(validWs ?? WritingSystemCodes.ValidTwoLetterCodes);
             target[writingSystemId] = word;
         }
     }

backend/FwLite/MiniLcm.Tests/AutoFakerHelpers/WritingSystemIdOverride.cs

@@ -0,0 +1,16 @@
+using MiniLcm.Models;
+using Soenneker.Utils.AutoBogus.Context;
+using Soenneker.Utils.AutoBogus.Override;
+
+namespace MiniLcm.Tests.AutoFakerHelpers;
+
+public class WritingSystemIdOverride: AutoFakerOverride<WritingSystemId>
+{
+    public override bool Preinitialize => false;
+
+    public override void Generate(AutoFakerOverrideContext context)
+    {
+        var ws = context.Faker.Random.ArrayElement(WritingSystemCodes.ValidTwoLetterCodes);
+        context.Instance = new WritingSystemId(ws);
+    }
+}

backend/FwLite/MiniLcm.Tests/WritingSystemCodes.cs

@@ -0,0 +1,8 @@
+using SIL.WritingSystems;
+
+namespace MiniLcm.Tests;
+
+public static class WritingSystemCodes
+{
+    public static readonly string[] ValidTwoLetterCodes = StandardSubtags.RegisteredLanguages.Select(lang => lang.Code).ToArray();
+}

backend/FwLite/MiniLcm.Tests/WritingSystemIdTests.cs

@@ -0,0 +1,34 @@
+using MiniLcm.Models;
+
+namespace MiniLcm.Tests;
+
+public class WritingSystemIdTests
+{
+    [Theory]
+    [InlineData("en")]
+    [InlineData("th")]
+    [InlineData("en-Zxxx-x-audio")]
+    public void ValidWritingSystemId_ShouldNotThrow(string code)
+    {
+        var ws = new WritingSystemId(code);
+        ws.Should().NotBeNull();
+    }
+
+    [Theory]
+    [InlineData("gx")]
+    [InlineData("oo")]
+    [InlineData("eng")] // Three-letter codes not allowed when there's a valid two-letter code
+    [InlineData("eng-Zxxx-x-audio")]
+    [InlineData("nonsense")]
+    public void InvalidWritingSystemId_ShouldThrow(string code)
+    {
+        Assert.Throws<ArgumentException>(() => new WritingSystemId(code));
+    }
+
+    [Fact]
+    public void DefaultWritingSystemId_IsValid()
+    {
+        var ws = new WritingSystemId("default");
+        ws.Should().NotBeNull();
+    }
+}

backend/FwLite/MiniLcm/MiniLcm.csproj

@@ -9,6 +9,7 @@
     </PropertyGroup>
 
     <ItemGroup>
+      <PackageReference Include="SIL.WritingSystems" Version="14.2.0-beta*" />
       <PackageReference Include="System.Text.Json" Version="8.0.5" />
       <PackageReference Include="SystemTextJsonPatch" Version="3.2.1" />
     </ItemGroup>

backend/FwLite/MiniLcm/Models/WritingSystemId.cs

@@ -1,5 +1,6 @@
 using System.Text.Json;
 using System.Text.Json.Serialization;
+using SIL.WritingSystems;
 
 namespace MiniLcm.Models;
 
@@ -26,11 +27,26 @@ public override void WriteAsPropertyName(Utf8JsonWriter writer, WritingSystemId
 }
 
 [JsonConverter(typeof(WritingSystemIdJsonConverter))]
-public readonly record struct WritingSystemId(string Code): ISpanFormattable, ISpanParsable<WritingSystemId>
+public readonly record struct WritingSystemId: ISpanFormattable, ISpanParsable<WritingSystemId>
 {
+    public string Code { get; init; }
+
+    public WritingSystemId(string code)
+    {
+        if (code == "default" || IetfLanguageTag.IsValid(code))
+        {
+            Code = code;
+        }
+        else
+        {
+            throw new ArgumentException($"Invalid writing system ID '{code}'", nameof(code));
+        }
+    }
+
     public static implicit operator string(WritingSystemId ws) => ws.Code;
     public static implicit operator WritingSystemId(string code) => new(code);
     public static implicit operator WritingSystemId(ReadOnlySpan<char> code) => new(new string(code));
+
     public override string ToString()
     {
         return Code;