Backport pull request #17858 from jellyfin/release-12.z

Improve SeriesNameRegex

Original-merge: f98daa19e863468a065400fbf2e13809161d529f

Merged-by: crobibero <cody@robibe.ro>

Backported-by: Cody Robibero <cody@robibe.ro>

2 files changed, 13 insertions, 5 deletions

diff --git a/Emby.Naming/TV/SeriesResolver.cs b/Emby.Naming/TV/SeriesResolver.cs
index 733e2418c2..ce42cb69fb 100644
--- a/Emby.Naming/TV/SeriesResolver.cs
+++ b/Emby.Naming/TV/SeriesResolver.cs
@@ -10,11 +10,14 @@ namespace Emby.Naming.TV
     public static partial class SeriesResolver
     {
         /// <summary>
-        /// Regex that matches strings of at least 2 characters separated by a dot or underscore.
-        /// Used for removing separators between words, i.e turns "The_show" into "The show" while
-        /// preserving names like "S.H.O.W".
+        /// Regex that matches a run of dots or underscores that separates two words, where a word is
+        /// at least 2 characters long. Used for removing separators between words, i.e turns
+        /// "The_show" into "The show" while preserving acronyms like "S.H.O.W", whose single letters
+        /// are a word on neither side. Whitespace bounds a word too, so the dot in
+        /// "Marvel's Agents of S.H.I.E.L.D." is read against the "S" beside it rather than against
+        /// the whole run of words before it.
         /// </summary>
-        [GeneratedRegex(@"((?<a>[^\._]{2,})[\._]*)|([\._](?<b>[^\._]{2,}))")]
+        [GeneratedRegex(@"(?<=[^\s\._]{2})[\._]+|[\._]+(?=[^\s\._]{2})")]
         private static partial Regex SeriesNameRegex();
 
         /// <summary>
@@ -60,7 +63,7 @@ namespace Emby.Naming.TV
 
             if (!string.IsNullOrEmpty(seriesName))
             {
-                seriesName = SeriesNameRegex().Replace(seriesName, "${a} ${b}").Trim();
+                seriesName = SeriesNameRegex().Replace(seriesName, " ").Trim();
             }
 
             return new SeriesInfo(path)
diff --git a/tests/Jellyfin.Naming.Tests/TV/SeriesResolverTests.cs b/tests/Jellyfin.Naming.Tests/TV/SeriesResolverTests.cs
index b81b7934cd..023c6cb2fa 100644
--- a/tests/Jellyfin.Naming.Tests/TV/SeriesResolverTests.cs
+++ b/tests/Jellyfin.Naming.Tests/TV/SeriesResolverTests.cs
@@ -20,6 +20,11 @@ namespace Jellyfin.Naming.Tests.TV
         [InlineData("/some/path/The Show s02e10 720p hdtv", "The Show")]
         [InlineData("/some/path/The Show s02e10 the episode 720p hdtv", "The Show")]
         [InlineData("/some/path/1923 (2022)", "1923")]
+        // A dotted acronym keeps its dots when it follows words, whether they are space or dot separated
+        [InlineData("/some/path/Marvel's Agents of S.H.I.E.L.D.", "Marvel's Agents of S.H.I.E.L.D.")]
+        [InlineData("Marvel's.Agents.of.S.H.I.E.L.D.", "Marvel's Agents of S.H.I.E.L.D.")]
+        [InlineData("The.Show.S.H.O.W", "The Show S.H.O.W")]
+        [InlineData("/some/path/Dawson's Creek", "Dawson's Creek")]
         public void SeriesResolverResolveTest(string path, string name)
         {
             var res = SeriesResolver.Resolve(_namingOptions, path);