From 5a6d9180fed81a30cb91ef3fed30176cd4402116 Mon Sep 17 00:00:00 2001
From: JPVenson <github@jpb.email>
Date: Thu, 25 Sep 2025 00:20:30 +0300
Subject: Add People Dedup and multiple progress fixes (#14848)

---
 .../ScheduledTasks/Tasks/PeopleValidationTask.cs   | 66 +++++++++++++++++++++-
 1 file changed, 63 insertions(+), 3 deletions(-)

(limited to 'Emby.Server.Implementations/ScheduledTasks')

diff --git a/Emby.Server.Implementations/ScheduledTasks/Tasks/PeopleValidationTask.cs b/Emby.Server.Implementations/ScheduledTasks/Tasks/PeopleValidationTask.cs
index 18162ad2fc..6e4e5c7808 100644
--- a/Emby.Server.Implementations/ScheduledTasks/Tasks/PeopleValidationTask.cs
+++ b/Emby.Server.Implementations/ScheduledTasks/Tasks/PeopleValidationTask.cs
@@ -1,10 +1,14 @@
 using System;
+using System.Buffers;
 using System.Collections.Generic;
+using System.Linq;
 using System.Threading;
 using System.Threading.Tasks;
+using Jellyfin.Database.Implementations;
 using MediaBrowser.Controller.Library;
 using MediaBrowser.Model.Globalization;
 using MediaBrowser.Model.Tasks;
+using Microsoft.EntityFrameworkCore;
 
 namespace Emby.Server.Implementations.ScheduledTasks.Tasks;
 
@@ -15,16 +19,19 @@ public class PeopleValidationTask : IScheduledTask, IConfigurableScheduledTask
 {
     private readonly ILibraryManager _libraryManager;
     private readonly ILocalizationManager _localization;
+    private readonly IDbContextFactory<JellyfinDbContext> _dbContextFactory;
 
     /// <summary>
     /// Initializes a new instance of the <see cref="PeopleValidationTask" /> class.
     /// </summary>
     /// <param name="libraryManager">Instance of the <see cref="ILibraryManager"/> interface.</param>
     /// <param name="localization">Instance of the <see cref="ILocalizationManager"/> interface.</param>
-    public PeopleValidationTask(ILibraryManager libraryManager, ILocalizationManager localization)
+    /// <param name="dbContextFactory">Instance of the <see cref="IDbContextFactory{TContext}"/> interface.</param>
+    public PeopleValidationTask(ILibraryManager libraryManager, ILocalizationManager localization, IDbContextFactory<JellyfinDbContext> dbContextFactory)
     {
         _libraryManager = libraryManager;
         _localization = localization;
+        _dbContextFactory = dbContextFactory;
     }
 
     /// <inheritdoc />
@@ -62,8 +69,61 @@ public class PeopleValidationTask : IScheduledTask, IConfigurableScheduledTask
     }
 
     /// <inheritdoc />
-    public Task ExecuteAsync(IProgress<double> progress, CancellationToken cancellationToken)
+    public async Task ExecuteAsync(IProgress<double> progress, CancellationToken cancellationToken)
     {
-        return _libraryManager.ValidatePeopleAsync(progress, cancellationToken);
+        IProgress<double> subProgress = new Progress<double>((val) => progress.Report(val / 2));
+        await _libraryManager.ValidatePeopleAsync(subProgress, cancellationToken).ConfigureAwait(false);
+
+        subProgress = new Progress<double>((val) => progress.Report((val / 2) + 50));
+        var context = await _dbContextFactory.CreateDbContextAsync(cancellationToken).ConfigureAwait(false);
+        await using (context.ConfigureAwait(false))
+        {
+            var dupQuery = context.Peoples
+                    .GroupBy(e => new { e.Name, e.PersonType })
+                    .Where(e => e.Count() > 1)
+                    .Select(e => e.Select(f => f.Id).ToArray());
+
+            var total = dupQuery.Count();
+
+            const int PartitionSize = 100;
+            var iterator = 0;
+            int itemCounter;
+            var buffer = ArrayPool<Guid[]>.Shared.Rent(PartitionSize)!;
+            try
+            {
+                do
+                {
+                    itemCounter = 0;
+                    await foreach (var item in dupQuery
+                        .Take(PartitionSize)
+                        .AsAsyncEnumerable()
+                        .WithCancellation(cancellationToken)
+                        .ConfigureAwait(false))
+                    {
+                        buffer[itemCounter++] = item;
+                    }
+
+                    for (int i = 0; i < itemCounter; i++)
+                    {
+                        var item = buffer[i];
+                        var reference = item[0];
+                        var dups = item[1..];
+                        await context.PeopleBaseItemMap.WhereOneOrMany(dups, e => e.PeopleId)
+                            .ExecuteUpdateAsync(e => e.SetProperty(f => f.PeopleId, reference), cancellationToken)
+                            .ConfigureAwait(false);
+                        await context.Peoples.Where(e => dups.Contains(e.Id)).ExecuteDeleteAsync(cancellationToken).ConfigureAwait(false);
+                        subProgress.Report(100f / total * ((iterator * PartitionSize) + i));
+                    }
+
+                    iterator++;
+                } while (itemCounter == PartitionSize && !cancellationToken.IsCancellationRequested);
+            }
+            finally
+            {
+                ArrayPool<Guid[]>.Shared.Return(buffer);
+            }
+
+            subProgress.Report(100);
+        }
     }
 }
-- 
cgit v1.2.3