Apply write normalization wrapper to tests, remove manual NFD workarounds

Now that we have proper normalization wrappers, the test base applies both:
- MiniLcmWriteApiNormalizationWrapperFactory (normalizes writes to NFD)
- MiniLcmApiStringNormalizationWrapperFactory (normalizes search queries to NFD)

This removes the manual NFD normalization workarounds in QueryEntryTestsBase
that were marked with "remove next line in #2065" comments.

https://claude.ai/code/session_01MYAMRqK5NJmFkQ7a7tVkTd

Author: claude

backend/FwLite/MiniLcm.Tests/MiniLcmTestBase.cs

@@ -17,7 +17,13 @@ public virtual async Task InitializeAsync()
     {
         BaseApi = await NewApi();
         BaseApi.Should().NotBeNull();
-        Api = BaseApi.WrapWith([new MiniLcmApiStringNormalizationWrapperFactory()], null!);
+        // Apply both normalization wrappers:
+        // - Write wrapper normalizes user input to NFD on create/update
+        // - String wrapper normalizes search queries and filter options to NFD
+        Api = BaseApi.WrapWith([
+            new MiniLcmWriteApiNormalizationWrapperFactory(),
+            new MiniLcmApiStringNormalizationWrapperFactory()
+        ], null!);
         Api.Should().NotBeNull();
     }
 

backend/FwLite/MiniLcm.Tests/QueryEntryTestsBase.cs

@@ -353,12 +353,11 @@ public async Task SuccessfulMatches(string searchTerm, string word, bool identic
     {
         // identical is to make the test cases more readable when they only differ in their normalization
         (searchTerm == word).Should().Be(identical);
-        // remove next line in https://github.com/sillsdev/languageforge-lexbox/issues/2065
-        word = word.Normalize(NormalizationForm.FormD);
+        // Word is stored as-is; the write normalization wrapper converts it to NFD
         await Api.CreateEntry(new Entry { LexemeForm = { ["en"] = word } });
         var words = await Api.SearchEntries(searchTerm).Select(e => e.LexemeForm["en"]).ToArrayAsync();
         words.Should().NotBeEmpty();
-        // Like LicLCM the MiniLcm API should normalize to NFD
+        // The API normalizes to NFD on write, so results should be NFD regardless of input
         words.Should().Contain(word.Normalize(NormalizationForm.FormD));
     }
 
@@ -370,12 +369,10 @@ public async Task SuccessfulMatches(string searchTerm, string word, bool identic
     [InlineData("É", "È")] // Different accents
     public async Task NegativeMatches(string searchTerm, string word)
     {
-        word = word.Normalize(NormalizationForm.FormD);
-        //should we be normalizing the search term internally?
-        searchTerm = searchTerm.Normalize(NormalizationForm.FormD);
+        // Both writes and searches are normalized to NFD by the API wrappers
         await Api.CreateEntry(new Entry { LexemeForm = { ["en"] = word } });
         var words = await Api.SearchEntries(searchTerm).Select(e => e.LexemeForm["en"]).ToArrayAsync();
-        words.Should().NotContain(word);
+        words.Should().NotContain(word.Normalize(NormalizationForm.FormD));
     }
 
     [Theory]