From 1100e1c368a36c61f975b8e1d7532d925bc9e6e5 Mon Sep 17 00:00:00 2001
From: Tyrrrz <1935960+Tyrrrz@users.noreply.github.com>
Date: Wed, 29 Nov 2023 23:09:04 +0200
Subject: [PATCH] Expect YouTube to provide 3-letter language codes by itself
 sometimes

---
 YoutubeExplode.Converter/Converter.cs         |  2 +-
 .../Utils/Extensions/LanguageExtensions.cs    | 21 ++++++++++++-------
 .../Utils/Extensions/StringExtensions.cs      | 14 ++++++++++++-
 .../Utils/Extensions/StringExtensions.cs      |  1 -
 .../Videos/ClosedCaptions/Language.cs         |  3 ++-
 5 files changed, 30 insertions(+), 11 deletions(-)

diff --git a/YoutubeExplode.Converter/Converter.cs b/YoutubeExplode.Converter/Converter.cs
index b53268f4..2c16f531 100644
--- a/YoutubeExplode.Converter/Converter.cs
+++ b/YoutubeExplode.Converter/Converter.cs
@@ -137,7 +137,7 @@ private async ValueTask ProcessAsync(
             // three-letter codes, so we'll try to convert to that first.
             var languageCode =
                 subtitleInput.Info.Language.TryGetThreeLetterCode()
-                ?? subtitleInput.Info.Language.GetTwoLetterCode();
+                ?? subtitleInput.Info.Language.Code;
 
             arguments
                 .Add($"-metadata:s:s:{i}")
diff --git a/YoutubeExplode.Converter/Utils/Extensions/LanguageExtensions.cs b/YoutubeExplode.Converter/Utils/Extensions/LanguageExtensions.cs
index 7449e7b0..8b999eb4 100644
--- a/YoutubeExplode.Converter/Utils/Extensions/LanguageExtensions.cs
+++ b/YoutubeExplode.Converter/Utils/Extensions/LanguageExtensions.cs
@@ -1,17 +1,23 @@
-using YoutubeExplode.Videos.ClosedCaptions;
+using System;
+using YoutubeExplode.Videos.ClosedCaptions;
 
 namespace YoutubeExplode.Converter.Utils.Extensions;
 
 internal static class LanguageExtensions
 {
-    public static string GetTwoLetterCode(this Language language)
+    public static string? TryGetThreeLetterCode(this Language language)
     {
-        var dashIndex = language.Code.IndexOf('-');
-        return dashIndex >= 0 ? language.Code[..dashIndex] : language.Code;
-    }
+        // YouTube provides either a two-letter or a three-letter language code,
+        // which may or may not also contain a region identifier.
+        var regionNeutralLanguageCode = language
+            .Code
+            .SubstringUntil("-", StringComparison.OrdinalIgnoreCase);
+
+        // Already a three-letter code
+        if (regionNeutralLanguageCode.Length == 3)
+            return regionNeutralLanguageCode;
 
-    public static string? TryGetThreeLetterCode(this Language language) =>
-        language.GetTwoLetterCode().ToLowerInvariant() switch
+        return regionNeutralLanguageCode.ToLowerInvariant() switch
         {
             "aa" => "aar",
             "ab" => "abk",
@@ -205,4 +211,5 @@ public static string GetTwoLetterCode(this Language language)
             "zu" => "zul",
             _ => null
         };
+    }
 }
diff --git a/YoutubeExplode.Converter/Utils/Extensions/StringExtensions.cs b/YoutubeExplode.Converter/Utils/Extensions/StringExtensions.cs
index 3be3a493..4a1c6252 100644
--- a/YoutubeExplode.Converter/Utils/Extensions/StringExtensions.cs
+++ b/YoutubeExplode.Converter/Utils/Extensions/StringExtensions.cs
@@ -1,7 +1,19 @@
-namespace YoutubeExplode.Converter.Utils.Extensions;
+using System;
+
+namespace YoutubeExplode.Converter.Utils.Extensions;
 
 internal static class StringExtensions
 {
     public static string? NullIfWhiteSpace(this string s) =>
         !string.IsNullOrWhiteSpace(s) ? s : null;
+
+    public static string SubstringUntil(
+        this string str,
+        string sub,
+        StringComparison comparison = StringComparison.Ordinal
+    )
+    {
+        var index = str.IndexOf(sub, comparison);
+        return index < 0 ? str : str[..index];
+    }
 }
diff --git a/YoutubeExplode/Utils/Extensions/StringExtensions.cs b/YoutubeExplode/Utils/Extensions/StringExtensions.cs
index f23e6041..b067efc1 100644
--- a/YoutubeExplode/Utils/Extensions/StringExtensions.cs
+++ b/YoutubeExplode/Utils/Extensions/StringExtensions.cs
@@ -18,7 +18,6 @@ public static string SubstringUntil(
     )
     {
         var index = str.IndexOf(sub, comparison);
-
         return index < 0 ? str : str[..index];
     }
 
diff --git a/YoutubeExplode/Videos/ClosedCaptions/Language.cs b/YoutubeExplode/Videos/ClosedCaptions/Language.cs
index 0907295b..7f05b4cc 100644
--- a/YoutubeExplode/Videos/ClosedCaptions/Language.cs
+++ b/YoutubeExplode/Videos/ClosedCaptions/Language.cs
@@ -9,7 +9,8 @@ namespace YoutubeExplode.Videos.ClosedCaptions;
 public readonly partial struct Language
 {
     /// <summary>
-    /// Two-letter (ISO 639-1) language code, possibly with a regional identifier (e.g. 'en' or 'en-US').
+    /// Two-letter or three-letter language code, possibly with a regional identifier
+    /// (e.g. 'en' or 'en-US' or 'eng').
     /// </summary>
     public string Code { get; }