* fix disposing the DbContext created by dbContextDefaultFactory …

…instead of the `Func<Owned<>>` factory itself

@ `DoWork()`
+ primary ctor param `replyContentImageSaver` to re-insert `ImageInReply` entites with images url filename from newly re-extracted reply contents
@ SimplifyImagesInAllReplyContentsWorker.cs
@ crawler

* using async overload for `DbContext.SaveChangesAsync()`
* add params for captured variables `processEntityCount` & `process` to fix ReSharper inspection `AccessToModifiedClosure`
* renamed from `SaveAndLog()`
@ `SaveThenLog()`
+ param `writingEntitiesAction` for `SaveThenLog()` & move param `writingEntityEntryAction` before it
@ `Transform()`
@ TransformEntityWorker.cs
@ shared
@ c#

c#/crawler/src/Worker/SimplifyImagesInAllReplyContentsWorker.cs

@@ -3,29 +3,39 @@ namespace tbm.Crawler.Worker;
 public class SimplifyImagesInAllReplyContentsWorker(
     ILogger<SimplifyImagesInAllReplyContentsWorker> logger,
     Func<Owned<CrawlerDbContext.NewDefault>> dbContextDefaultFactory,
-    Func<Owned<CrawlerDbContext.New>> dbContextFactory)
+    Func<Owned<CrawlerDbContext.New>> dbContextFactory,
+    ReplyContentImageSaver replyContentImageSaver)
     : TransformEntityWorker<CrawlerDbContext, ReplyContent, ReplyContent, Pid>(logger)
 {
     protected override async Task DoWork(CancellationToken stoppingToken)
     {
-        await using var db = dbContextDefaultFactory().Value();
+        await using var dbDefaultFactory = dbContextDefaultFactory();
+        var db = dbDefaultFactory.Value();
         foreach (var fid in from e in db.Forums select e.Fid)
         {
+            await using var dbFactory = dbContextFactory();
             await Transform(
-                () => dbContextFactory().Value(fid),
+                () => dbFactory.Value(fid),
                 saveByNthEntityCount: 10000,
-                writingEntityEntry =>
-                {
-                    var p = writingEntityEntry.Property(e => e.ProtoBufBytes);
-                    p.IsModified = !ByteArrayEqualityComparer.Instance.Equals(p.OriginalValue, p.CurrentValue);
-                },
                 readingEntity => readingEntity.Pid,
                 readingEntity =>
                 {
                     var protoBuf = Reply.Parser.ParseFrom(readingEntity.ProtoBufBytes);
                     ReplyParser.SimplifyImagesInReplyContent(logger, ref protoBuf);
                     return new() {Pid = readingEntity.Pid, ProtoBufBytes = protoBuf.ToByteArray()};
                 },
+                writingEntityEntry =>
+                {
+                    var p = writingEntityEntry.Property(e => e.ProtoBufBytes);
+                    p.IsModified = !ByteArrayEqualityComparer.Instance.Equals(p.OriginalValue, p.CurrentValue);
+                },
+                (writingDb, writingEntities) => replyContentImageSaver
+                    .Save(writingDb, writingEntities.Select(e => new ReplyPost
+                    {
+                        Pid = e.Pid,
+                        Content = null!,
+                        ContentsProtoBuf = Reply.Parser.ParseFrom(e.ProtoBufBytes).Content
+                    })),
                 stoppingToken);
         }
     }

c#/shared/src/TransformEntityWorker.cs

@@ -28,9 +28,10 @@ public abstract class TransformEntityWorker<TDbContext, TReadingEntity, TWriting
     protected async Task Transform(
         Func<TDbContext> dbContextFactory,
         int saveByNthEntityCount,
-        Action<EntityEntry<TWritingEntity>> writingEntityEntryAction,
         Func<TReadingEntity, TExceptionId> readingEntityExceptionIdSelector,
         Func<TReadingEntity, TWritingEntity> entityTransformer,
+        Action<EntityEntry<TWritingEntity>> writingEntityEntryAction,
+        Action<TDbContext, IEnumerable<TWritingEntity>> writingEntitiesAction,
         CancellationToken stoppingToken = default)
     {
         var processedEntityCount = 0;
@@ -45,26 +46,27 @@ protected async Task Transform(
             from e in readingDb.Set<TReadingEntity>().AsNoTracking() select e;
         var writingEntities = new List<TWritingEntity>();
 
-        void SaveAndLog()
+        async Task SaveThenLog(int processedCount, Process currentProcess)
         {
-            writingDb.Set<TWritingEntity>().UpdateRange(writingEntities);
+            writingDb.Set<TWritingEntity>().AttachRange(writingEntities);
             writingDb.ChangeTracker.Entries<TWritingEntity>().ForEach(writingEntityEntryAction);
-            var updatedEntityCount = writingDb.SaveChanges();
+            writingEntitiesAction(writingDb, writingEntities);
+            var updatedEntityCount = await writingDb.SaveChangesAsync(stoppingToken);
             writingEntities.Clear();
             writingDb.ChangeTracker.Clear();
 
             logger.LogTrace("processedEntityCount:{} updatedEntityCount:{} elapsed:{}ms processMemory:{}MiB exceptions:{}",
-                processedEntityCount, updatedEntityCount,
+                processedCount, updatedEntityCount,
                 stopwatch.ElapsedMilliseconds,
-                process.PrivateMemorySize64 / 1024 / 1024,
+                currentProcess.PrivateMemorySize64 / 1024 / 1024,
                 JsonSerializer.Serialize(exceptions, JsonSerializerOptions));
             stopwatch.Restart();
         }
 
         foreach (var readingEntity in readingEntities)
         {
             processedEntityCount++;
-            if (processedEntityCount % saveByNthEntityCount == 0) SaveAndLog();
+            if (processedEntityCount % saveByNthEntityCount == 0) await SaveThenLog(processedEntityCount, process);
             if (stoppingToken.IsCancellationRequested) break;
             try
             {
@@ -85,6 +87,6 @@ void SaveAndLog()
             }
         }
 
-        SaveAndLog();
+        await SaveThenLog(processedEntityCount, process);
     }
 }