New: Use languages from Torznab/Newznab attributes if given

Closes #5654

src/NzbDrone.Core.Test/Download/Aggregation/Aggregators/AggregateLanguagesFixture.cs

@@ -32,6 +32,7 @@ namespace NzbDrone.Core.Test.Download.Aggregation.Aggregators
                                                  .With(l => l.ParsedEpisodeInfo = null)
                                                  .With(l => l.Episodes = episodes)
                                                  .With(l => l.Series = _series)
+                                                 .With(l => l.Release = new ReleaseInfo())
                                                  .Build();
         }
 

src/NzbDrone.Core/Download/Aggregation/Aggregators/AggregateLanguages.cs

@@ -20,6 +20,7 @@ namespace NzbDrone.Core.Download.Aggregation.Aggregators
         public RemoteEpisode Aggregate(RemoteEpisode remoteEpisode)
         {
             var parsedEpisodeInfo = remoteEpisode.ParsedEpisodeInfo;
+            var releaseInfo = remoteEpisode.Release;
             var languages = parsedEpisodeInfo.Languages;
             var series = remoteEpisode.Series;
             var releaseTokens = parsedEpisodeInfo.ReleaseTokens ?? parsedEpisodeInfo.ReleaseTitle;
@@ -31,34 +32,46 @@ namespace NzbDrone.Core.Download.Aggregation.Aggregators
                 _logger.Debug("Unable to aggregate languages, using parsed values: {0}", string.Join(", ", languages.ToList()));
 
                 remoteEpisode.Languages = languages;
+                remoteEpisode.Languages = releaseInfo != null && releaseInfo.Languages.Any() ? releaseInfo.Languages : languages;
 
                 return remoteEpisode;
             }
 
-            // Exclude any languages that are part of the episode title, if the episode title is in the release tokens (falls back to release title)
+            if (releaseInfo != null && releaseInfo.Languages.Any())
-            foreach (var episode in remoteEpisode.Episodes)
             {
-                var episodeTitleLanguage = LanguageParser.ParseLanguages(episode.Title);
+                _logger.Debug("Languages provided by indexer, using release values: {0}", string.Join(", ", releaseInfo.Languages));
 
-                if (!episodeTitleLanguage.Contains(Language.Unknown))
+                // Use languages from release (given by indexer or user) if available
+                languages = releaseInfo.Languages;
+            }
+            else
+            {
+                // Exclude any languages that are part of the episode title, if the episode title is in the release tokens (falls back to release title)
+                foreach (var episode in remoteEpisode.Episodes)
                 {
-                    var normalizedEpisodeTitle = Parser.Parser.NormalizeEpisodeTitle(episode.Title);
+                    var episodeTitleLanguage = LanguageParser.ParseLanguages(episode.Title);
-                    var episodeTitleIndex = normalizedReleaseTokens.IndexOf(normalizedEpisodeTitle, StringComparison.CurrentCultureIgnoreCase);
 
-                    if (episodeTitleIndex >= 0)
+                    if (!episodeTitleLanguage.Contains(Language.Unknown))
                     {
-                        releaseTokens = releaseTokens.Remove(episodeTitleIndex, normalizedEpisodeTitle.Length);
+                        var normalizedEpisodeTitle = Parser.Parser.NormalizeEpisodeTitle(episode.Title);
-                        languagesToRemove.AddRange(episodeTitleLanguage);
+                        var episodeTitleIndex = normalizedReleaseTokens.IndexOf(normalizedEpisodeTitle,
+                            StringComparison.CurrentCultureIgnoreCase);
+
+                        if (episodeTitleIndex >= 0)
+                        {
+                            releaseTokens = releaseTokens.Remove(episodeTitleIndex, normalizedEpisodeTitle.Length);
+                            languagesToRemove.AddRange(episodeTitleLanguage);
+                        }
                     }
                 }
+
+                // Remove any languages still in the title that would normally be removed
+                languagesToRemove = languagesToRemove.Except(LanguageParser.ParseLanguages(releaseTokens)).ToList();
+
+                // Remove all languages that aren't part of the updated releaseTokens
+                languages = languages.Except(languagesToRemove).ToList();
             }
 
-            // Remove any languages still in the title that would normally be removed
-            languagesToRemove = languagesToRemove.Except(LanguageParser.ParseLanguages(releaseTokens)).ToList();
-
-            // Remove all languages that aren't part of the updated releaseTokens
-            languages = languages.Except(languagesToRemove).ToList();
-
             // Use series language as fallback if we couldn't parse a language
             if (languages.Count == 0 || (languages.Count == 1 && languages.First() == Language.Unknown))
             {

src/NzbDrone.Core/Indexers/IndexerBase.cs

@@ -1,6 +1,7 @@
 using System;
 using System.Collections.Generic;
 using System.Linq;
+using System.Text.RegularExpressions;
 using FluentValidation.Results;
 using NLog;
 using NzbDrone.Common.Http;
@@ -15,6 +16,8 @@ namespace NzbDrone.Core.Indexers
     public abstract class IndexerBase<TSettings> : IIndexer
         where TSettings : IIndexerSettings, new()
     {
+        private static readonly Regex MultiRegex = new (@"\b(?<multi>multi)\b", RegexOptions.Compiled | RegexOptions.IgnoreCase);
+
         protected readonly IIndexerStatusService _indexerStatusService;
         protected readonly IConfigService _configService;
         protected readonly IParsingService _parsingService;
@@ -79,6 +82,7 @@ namespace NzbDrone.Core.Indexers
         protected virtual IList<ReleaseInfo> CleanupReleases(IEnumerable<ReleaseInfo> releases)
         {
             var result = releases.DistinctBy(v => v.Guid).ToList();
+            var settings = Definition.Settings as IIndexerSettings;
 
             result.ForEach(c =>
             {

src/NzbDrone.Core/Indexers/Newznab/NewznabRssParser.cs

@@ -4,6 +4,8 @@ using System.Linq;
 using System.Xml.Linq;
 using NzbDrone.Common.Extensions;
 using NzbDrone.Core.Indexers.Exceptions;
+using NzbDrone.Core.Languages;
+using NzbDrone.Core.Parser;
 using NzbDrone.Core.Parser.Model;
 
 namespace NzbDrone.Core.Indexers.Newznab
@@ -101,6 +103,30 @@ namespace NzbDrone.Core.Indexers.Newznab
             return ParseUrl(item.TryGetValue("comments"));
         }
 
+        protected override List<Language> GetLanguages(XElement item)
+        {
+            var languges = TryGetMultipleNewznabAttributes(item, "language");
+            var results = new List<Language>();
+
+            // Try to find <language> elements for some indexers that suck at following the rules.
+            if (languges.Count == 0)
+            {
+                languges = item.Elements("language").Select(e => e.Value).ToList();
+            }
+
+            foreach (var language in languges)
+            {
+                var mappedLanguage = IsoLanguages.FindByName(language)?.Language ?? null;
+
+                if (mappedLanguage != null)
+                {
+                    results.Add(mappedLanguage);
+                }
+            }
+
+            return results;
+        }
+
         protected override long GetSize(XElement item)
         {
             long size;
@@ -167,5 +193,22 @@ namespace NzbDrone.Core.Indexers.Newznab
 
             return defaultValue;
         }
+
+        protected List<string> TryGetMultipleNewznabAttributes(XElement item, string key)
+        {
+            var attrElements = item.Elements(ns + "attr").Where(e => e.Attribute("name").Value.Equals(key, StringComparison.OrdinalIgnoreCase));
+            var results = new List<string>();
+
+            foreach (var element in attrElements)
+            {
+                var attrValue = element.Attribute("value");
+                if (attrValue != null)
+                {
+                    results.Add(attrValue.Value);
+                }
+            }
+
+            return results;
+        }
     }
 }

src/NzbDrone.Core/Indexers/RssParser.cs

@@ -12,6 +12,7 @@ using NzbDrone.Common.Extensions;
 using NzbDrone.Common.Http;
 using NzbDrone.Common.Instrumentation;
 using NzbDrone.Core.Indexers.Exceptions;
+using NzbDrone.Core.Languages;
 using NzbDrone.Core.Parser.Model;
 
 namespace NzbDrone.Core.Indexers
@@ -158,6 +159,7 @@ namespace NzbDrone.Core.Indexers
             releaseInfo.DownloadUrl = GetDownloadUrl(item);
             releaseInfo.InfoUrl = GetInfoUrl(item);
             releaseInfo.CommentUrl = GetCommentUrl(item);
+            releaseInfo.Languages = GetLanguages(item);
 
             try
             {
@@ -224,6 +226,11 @@ namespace NzbDrone.Core.Indexers
             return ParseUrl((string)item.Element("comments"));
         }
 
+        protected virtual List<Language> GetLanguages(XElement item)
+        {
+            return new List<Language>();
+        }
+
         protected virtual long GetSize(XElement item)
         {
             if (UseEnclosureLength)

src/NzbDrone.Core/Indexers/Torznab/TorznabRssParser.cs

@@ -4,6 +4,8 @@ using System.Linq;
 using System.Xml.Linq;
 using NzbDrone.Common.Extensions;
 using NzbDrone.Core.Indexers.Exceptions;
+using NzbDrone.Core.Languages;
+using NzbDrone.Core.Parser;
 using NzbDrone.Core.Parser.Model;
 
 namespace NzbDrone.Core.Indexers.Torznab
@@ -92,6 +94,30 @@ namespace NzbDrone.Core.Indexers.Torznab
             return ParseUrl(item.TryGetValue("comments"));
         }
 
+        protected override List<Language> GetLanguages(XElement item)
+        {
+            var languges = TryGetMultipleTorznabAttributes(item, "language");
+            var results = new List<Language>();
+
+            // Try to find <language> elements for some indexers that suck at following the rules.
+            if (languges.Count == 0)
+            {
+                languges = item.Elements("language").Select(e => e.Value).ToList();
+            }
+
+            foreach (var language in languges)
+            {
+                var mappedLanguage = IsoLanguages.FindByName(language)?.Language ?? null;
+
+                if (mappedLanguage != null)
+                {
+                    results.Add(mappedLanguage);
+                }
+            }
+
+            return results;
+        }
+
         protected override long GetSize(XElement item)
         {
             long size;
@@ -206,5 +232,22 @@ namespace NzbDrone.Core.Indexers.Torznab
 
             return defaultValue;
         }
+
+        protected List<string> TryGetMultipleTorznabAttributes(XElement item, string key)
+        {
+            var attrElements = item.Elements(ns + "attr").Where(e => e.Attribute("name").Value.Equals(key, StringComparison.OrdinalIgnoreCase));
+            var results = new List<string>();
+
+            foreach (var element in attrElements)
+            {
+                var attrValue = element.Attribute("value");
+                if (attrValue != null)
+                {
+                    results.Add(attrValue.Value);
+                }
+            }
+
+            return results;
+        }
     }
 }

src/NzbDrone.Core/Parser/Model/ReleaseInfo.cs

@@ -1,11 +1,18 @@
 using System;
+using System.Collections.Generic;
 using System.Text;
 using NzbDrone.Core.Indexers;
+using NzbDrone.Core.Languages;
 
 namespace NzbDrone.Core.Parser.Model
 {
     public class ReleaseInfo
     {
+        public ReleaseInfo()
+        {
+            Languages = new List<Language>();
+        }
+
         public string Guid { get; set; }
         public string Title { get; set; }
         public long Size { get; set; }
@@ -28,6 +35,8 @@ namespace NzbDrone.Core.Parser.Model
         public string Codec { get; set; }
         public string Resolution { get; set; }
 
+        public List<Language> Languages { get; set; }
+
         public int Age
         {
             get