Sanitize contributor names (#1518)

Source/AudibleUtilities/ApiExtended.cs

@@ -1,16 +1,17 @@
-using System;
-using System.Collections.Generic;
-using System.Linq;
-using System.Threading.Channels;
-using System.Threading.Tasks;
-using System.Diagnostics;
-using AudibleApi;
+using AudibleApi;
 using AudibleApi.Common;
 using Dinah.Core;
+using LibationFileManager;
+using Newtonsoft.Json.Linq;
 using Polly;
 using Polly.Retry;
+using System;
+using System.Collections.Generic;
+using System.Diagnostics;
+using System.Linq;
 using System.Threading;
-using LibationFileManager;
+using System.Threading.Channels;
+using System.Threading.Tasks;
 
 #nullable enable
 namespace AudibleUtilities
@@ -82,6 +83,23 @@ namespace AudibleUtilities
 			return policy.ExecuteAsync(() => getItemsAsync(libraryOptions));
 		}
 
+		/// <summary>
+		/// A debugging method used to simulate a library scan from a LibraryScans.zip json file.
+		/// Simply replace the Api call to GetLibraryItemsPagesAsync() with a call to this method.
+		/// </summary>
+		private static async IAsyncEnumerable<Item[]> GetItemsFromJsonFile()
+		{
+			var libraryScanJsonPath = @"Path/to/libraryscan.json";
+			using var jsonFile = System.IO.File.OpenText(libraryScanJsonPath);
+
+			var json = await JToken.ReadFromAsync(new Newtonsoft.Json.JsonTextReader(jsonFile));
+			if (json?["Items"] is not JArray items)
+				yield break;
+
+			foreach (var batch in items.Select(i => Item.FromJson(i as JObject)).Chunk(BatchSize))
+				yield return batch;
+		}
+
 		private async Task<List<Item>> getItemsAsync(LibraryOptions libraryOptions)
 		{
 			Serilog.Log.Logger.Debug("Beginning library scan.");
@@ -162,6 +180,7 @@ namespace AudibleUtilities
 			Serilog.Log.Logger.Information("Completed indexing series episodes after {elappsed_ms} ms.", sw.ElapsedMilliseconds);
 			Serilog.Log.Logger.Information($"Completed library scan in {totalTime.TotalMilliseconds:F0} ms.");
 
+			Array.ForEach(ISanitizer.GetAllSanitizers(), s => s.Sanitize(items));
 			var allExceptions = IValidator.GetAllValidators().SelectMany(v => v.Validate(items)).ToList();
 			if (allExceptions?.Count > 0)
 				throw new ImportValidationException(items, allExceptions);

Source/AudibleUtilities/AudibleApiSanitizers.cs

@@ -0,0 +1,33 @@
+using AudibleApi.Common;
+using System.Collections.Generic;
+using System.Linq;
+
+#nullable enable
+namespace AudibleUtilities;
+
+public interface ISanitizer
+{
+	void Sanitize(IEnumerable<Item> items);
+
+	public static ISanitizer[] GetAllSanitizers() => [
+		new ContributorSanitizer()
+	];
+}
+
+public class ContributorSanitizer : ISanitizer
+{
+	public void Sanitize(IEnumerable<Item> items)
+	{
+		foreach (var item in items)
+		{
+			item.Authors = SanitizePersonArray(item.Authors);
+			item.Narrators = SanitizePersonArray(item.Narrators);
+		}
+	}
+
+	private static Person[]? SanitizePersonArray(Person?[]? contributors)
+		=> contributors
+		?.OfType<Person>()
+		.Where(c => !string.IsNullOrWhiteSpace(c.Asin) && !string.IsNullOrWhiteSpace(c.Name))
+		.ToArray();
+}

Source/AudibleUtilities/AudibleApiValidators.cs

@@ -9,17 +9,21 @@ namespace AudibleUtilities
 	{
 		IEnumerable<Exception> Validate(IEnumerable<Item> items);
 
-		public static IValidator[] GetAllValidators()
-			=> new IValidator[]
-			{
-				new LibraryValidator(),
-				new BookValidator(),
-				new CategoryValidator(),
-				new ContributorValidator(),
-				new SeriesValidator(),
-			};
+		public static IValidator[] GetAllValidators() => [
+			new LibraryValidator(),
+			new BookValidator(),
+			new CategoryValidator(),
+			new SeriesValidator(),
+		];
 	}
 
+	/// <summary>
+	/// To be used when no validation is desired
+	/// </summary>
+	public class ClearValidator : IValidator
+	{
+		public IEnumerable<Exception> Validate(IEnumerable<Item> items) => [];
+	}
 	public class LibraryValidator : IValidator
 	{
 		public IEnumerable<Exception> Validate(IEnumerable<Item> items)
@@ -68,20 +72,6 @@ namespace AudibleUtilities
 			return exceptions;
 		}
 	}
-	public class ContributorValidator : IValidator
-	{
-		public IEnumerable<Exception> Validate(IEnumerable<Item> items)
-		{
-			var exceptions = new List<Exception>();
-
-			if (items.GetAuthorsDistinct().Any(a => string.IsNullOrWhiteSpace(a.Name)))
-				exceptions.Add(new ArgumentException($"Collection contains {nameof(Item.Authors)} with null {nameof(Person.Name)}", nameof(items)));
-			if (items.GetNarratorsDistinct().Any(a => string.IsNullOrWhiteSpace(a.Name)))
-				exceptions.Add(new ArgumentException($"Collection contains {nameof(Item.Narrators)} with null {nameof(Person.Name)}", nameof(items)));
-
-			return exceptions;
-		}
-	}
 	public class SeriesValidator : IValidator
 	{
 		public IEnumerable<Exception> Validate(IEnumerable<Item> items)

Source/DtoImporterService/ContributorImporter.cs

@@ -10,7 +10,7 @@ namespace DtoImporterService
 {
 	public class ContributorImporter : ItemsImporterBase
 	{
-		protected override IValidator Validator => new ContributorValidator();
+		protected override IValidator Validator => new ClearValidator();
 
 		public Dictionary<string, Contributor> Cache { get; private set; } = new();