* fix non-split revision entities will always be inserted with the de…

…fault value of field `DuplicateIndex` * change the constraint of type param `TRevision` from `BaseRevisionWithSplitting` to `class, TBaseRevision` @ `SaveEntitiesWithRevision()` * opt-out EFCore change tracker for existing revision entities * renamed from `AddSplitRevisions()` @ `AddRevisionsWithDuplicateIndex()` @ SaverWithRevision.cs * change the constraint of type param `TRevision` from `BaseRevisionWithSplitting` to `TBaseRevision` @ `PostSaver.Save()` * rename method `SplitToTable()` to `HasSplitTable()` @ `RevisionWithSplitting.ModelBuilder` @ c#/crawler
n0099 · Jun 1, 2024 · 522f9b3 · 522f9b3
1 parent 44a35a8
commit 522f9b3
Show file tree

Hide file tree

Showing 8 changed files with 24 additions and 24 deletions.
diff --git a/c#/crawler/src/Db/CrawlerDbContext.cs b/c#/crawler/src/Db/CrawlerDbContext.cs
@@ -80,27 +80,27 @@ protected override void OnModelCreating(ModelBuilder b)
         _ = new RevisionWithSplitting<BaseThreadRevision>
                 .ModelBuilder(b, "tbmcr_thread", e => new {e.Tid, e.TakenAt, e.DuplicateIndex})
             .HasBaseTable<ThreadRevision>()
-            .SplitToTable<SplitViewCount>("viewCount");
+            .HasSplitTable<SplitViewCount>("viewCount");
 
         _ = new RevisionWithSplitting<BaseReplyRevision>
                 .ModelBuilder(b, "tbmcr_reply", e => new {e.Pid, e.TakenAt, e.DuplicateIndex})
             .HasBaseTable<ReplyRevision>()
-            .SplitToTable<ReplyRevision.SplitAgreeCount>("agreeCount")
-            .SplitToTable<SplitSubReplyCount>("subReplyCount")
-            .SplitToTable<SplitFloor>("floor");
+            .HasSplitTable<ReplyRevision.SplitAgreeCount>("agreeCount")
+            .HasSplitTable<SplitSubReplyCount>("subReplyCount")
+            .HasSplitTable<SplitFloor>("floor");
 
         _ = new RevisionWithSplitting<BaseSubReplyRevision>
                 .ModelBuilder(b, "tbmcr_subReply", e => new {e.Spid, e.TakenAt, e.DuplicateIndex})
             .HasBaseTable<SubReplyRevision>()
-            .SplitToTable<SubReplyRevision.SplitAgreeCount>("agreeCount")
-            .SplitToTable<SplitDisagreeCount>("disagreeCount");
+            .HasSplitTable<SubReplyRevision.SplitAgreeCount>("agreeCount")
+            .HasSplitTable<SplitDisagreeCount>("disagreeCount");
 
         _ = new RevisionWithSplitting<BaseUserRevision>
                 .ModelBuilder(b, "tbmcr_user", e => new {e.Uid, e.TakenAt, e.DuplicateIndex})
             .HasBaseTable<UserRevision>()
-            .SplitToTable<SplitIpGeolocation>("ipGeolocation")
-            .SplitToTable<SplitPortraitUpdatedAt>("portraitUpdatedAt")
-            .SplitToTable<SplitDisplayName>("displayName");
+            .HasSplitTable<SplitIpGeolocation>("ipGeolocation")
+            .HasSplitTable<SplitPortraitUpdatedAt>("portraitUpdatedAt")
+            .HasSplitTable<SplitDisplayName>("displayName");
 
         b.Entity<SplitDisplayName>().Property(e => e.DisplayName).HasConversion<byte[]>();
         b.Entity<User>().Property(e => e.DisplayName).HasConversion<byte[]>();

diff --git a/c#/crawler/src/Db/Revision/Splitting/RevisionWithSplitting.cs b/c#/crawler/src/Db/Revision/Splitting/RevisionWithSplitting.cs
@@ -36,7 +36,7 @@ public ModelBuilder HasBaseTable<TRevision>() where TRevision : TBaseRevision
             return this;
         }
 
-        public ModelBuilder SplitToTable<TRevisionWithSplitting>(string tableNameSuffix)
+        public ModelBuilder HasSplitTable<TRevisionWithSplitting>(string tableNameSuffix)
             where TRevisionWithSplitting : RevisionWithSplitting<TBaseRevision>
         {
             var visitor = new ReplaceParameterTypeVisitor<TBaseRevision, TRevisionWithSplitting>();

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/Post/PostSaver.cs b/c#/crawler/src/Tieba/Crawl/Saver/Post/PostSaver.cs
@@ -29,7 +29,7 @@ protected SaverChangeSet<TPost> Save<TRevision>(
         Func<TPost, PostId> postIdSelector,
         Func<TPost, TRevision> revisionFactory,
         ExpressionStarter<TPost> existingPostPredicate)
-        where TRevision : BaseRevisionWithSplitting
+        where TRevision : TBaseRevision
     {
         var existingPostsKeyById = db.Set<TPost>().AsTracking()
             .Where(existingPostPredicate).ToDictionary(postIdSelector);

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/Post/ReplySaver.cs b/c#/crawler/src/Tieba/Crawl/Saver/Post/ReplySaver.cs
@@ -17,9 +17,9 @@ protected override Lazy<Dictionary<Type, AddSplitRevisionsDelegate>>
         AddSplitRevisionsDelegatesKeyByEntityType =>
         _addSplitRevisionsDelegatesKeyByEntityType ??= new(() => new()
         {
-            {typeof(ReplyRevision.SplitFloor), AddSplitRevisions<ReplyRevision.SplitFloor>},
-            {typeof(ReplyRevision.SplitSubReplyCount), AddSplitRevisions<ReplyRevision.SplitSubReplyCount>},
-            {typeof(ReplyRevision.SplitAgreeCount), AddSplitRevisions<ReplyRevision.SplitAgreeCount>}
+            {typeof(ReplyRevision.SplitFloor), AddRevisionsWithDuplicateIndex<ReplyRevision.SplitFloor>},
+            {typeof(ReplyRevision.SplitSubReplyCount), AddRevisionsWithDuplicateIndex<ReplyRevision.SplitSubReplyCount>},
+            {typeof(ReplyRevision.SplitAgreeCount), AddRevisionsWithDuplicateIndex<ReplyRevision.SplitAgreeCount>}
         });
 
     public override bool UserFieldUpdateIgnorance(string propName, object? oldValue, object? newValue) => propName switch

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/Post/SubReplySaver.cs b/c#/crawler/src/Tieba/Crawl/Saver/Post/SubReplySaver.cs
@@ -15,8 +15,8 @@ protected override Lazy<Dictionary<Type, AddSplitRevisionsDelegate>>
         AddSplitRevisionsDelegatesKeyByEntityType =>
         _addSplitRevisionsDelegatesKeyByEntityType ??= new(() => new()
         {
-            {typeof(SubReplyRevision.SplitAgreeCount), AddSplitRevisions<SubReplyRevision.SplitAgreeCount>},
-            {typeof(SubReplyRevision.SplitDisagreeCount), AddSplitRevisions<SubReplyRevision.SplitDisagreeCount>},
+            {typeof(SubReplyRevision.SplitAgreeCount), AddRevisionsWithDuplicateIndex<SubReplyRevision.SplitAgreeCount>},
+            {typeof(SubReplyRevision.SplitDisagreeCount), AddRevisionsWithDuplicateIndex<SubReplyRevision.SplitDisagreeCount>},
         });
 
     public override bool UserFieldUpdateIgnorance

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/Post/ThreadSaver.cs b/c#/crawler/src/Tieba/Crawl/Saver/Post/ThreadSaver.cs
@@ -17,7 +17,7 @@ protected override Lazy<Dictionary<Type, AddSplitRevisionsDelegate>>
         AddSplitRevisionsDelegatesKeyByEntityType =>
         _addSplitRevisionsDelegatesKeyByEntityType ??= new(() => new()
         {
-            {typeof(ThreadRevision.SplitViewCount), AddSplitRevisions<ThreadRevision.SplitViewCount>}
+            {typeof(ThreadRevision.SplitViewCount), AddRevisionsWithDuplicateIndex<ThreadRevision.SplitViewCount>}
         });
 
     public override bool UserFieldUpdateIgnorance

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/SaverWithRevision.cs b/c#/crawler/src/Tieba/Crawl/Saver/SaverWithRevision.cs
@@ -11,7 +11,7 @@ public abstract partial class SaverWithRevision<TBaseRevision, TEntityId>(
     protected abstract Lazy<Dictionary<Type, AddSplitRevisionsDelegate>>
         AddSplitRevisionsDelegatesKeyByEntityType { get; }
 
-    protected void AddSplitRevisions<TRevision>(CrawlerDbContext db, IEnumerable<TBaseRevision> revisions)
+    protected void AddRevisionsWithDuplicateIndex<TRevision>(CrawlerDbContext db, IEnumerable<TBaseRevision> revisions)
         where TRevision : TBaseRevision
     {
         var newRevisions = revisions.OfType<TRevision>().ToList();
@@ -20,7 +20,7 @@ protected void AddSplitRevisions<TRevision>(CrawlerDbContext db, IEnumerable<TBa
 
         // https://github.com/npgsql/npgsql/issues/4437
         // https://github.com/dotnet/efcore/issues/32092
-        var existingRevisions = dbSet
+        var existingRevisions = dbSet.AsNoTracking()
             .Where(newRevisions.Aggregate(
                 LinqKit.PredicateBuilder.New<TRevision>(),
                 (predicate, newRevision) => predicate.Or(LinqKit.PredicateBuilder
@@ -62,7 +62,7 @@ protected void SaveEntitiesWithRevision<TEntity, TRevision>(
         UserSaver.FieldChangeIgnorance? userFieldUpdateIgnorance = null,
         UserSaver.FieldChangeIgnorance? userFieldRevisionIgnorance = null)
         where TEntity : RowVersionedEntity
-        where TRevision : BaseRevisionWithSplitting
+        where TRevision : class, TBaseRevision
     {
         db.Set<TEntity>().AddRange(isExistingEntityLookup[false]); // newly added
         var newRevisions = isExistingEntityLookup[true].Select(newEntity =>
@@ -147,7 +147,7 @@ bool IsTimestampingFieldName(string name) => name is nameof(BasePost.LastSeenAt)
         }).OfType<TRevision>().ToList();
         if (newRevisions.Count == 0) return; // quick exit to prevent execute sql with WHERE FALSE clause
 
-        db.Set<TRevision>().AddRange(
+        AddRevisionsWithDuplicateIndex<TRevision>(db,
             newRevisions.Where(rev => !rev.IsAllFieldsIsNullExceptSplit()));
         newRevisions.OfType<RevisionWithSplitting<TBaseRevision>>()
             .SelectMany(rev => rev.SplitEntities)

diff --git a/c#/crawler/src/Tieba/Crawl/Saver/UserSaver.cs b/c#/crawler/src/Tieba/Crawl/Saver/UserSaver.cs
@@ -9,9 +9,9 @@ protected override Lazy<Dictionary<Type, AddSplitRevisionsDelegate>>
         AddSplitRevisionsDelegatesKeyByEntityType =>
         _addSplitRevisionsDelegatesKeyByEntityType ??= new(() => new()
         {
-            {typeof(UserRevision.SplitDisplayName), AddSplitRevisions<UserRevision.SplitDisplayName>},
-            {typeof(UserRevision.SplitPortraitUpdatedAt), AddSplitRevisions<UserRevision.SplitPortraitUpdatedAt>},
-            {typeof(UserRevision.SplitIpGeolocation), AddSplitRevisions<UserRevision.SplitIpGeolocation>}
+            {typeof(UserRevision.SplitDisplayName), AddRevisionsWithDuplicateIndex<UserRevision.SplitDisplayName>},
+            {typeof(UserRevision.SplitPortraitUpdatedAt), AddRevisionsWithDuplicateIndex<UserRevision.SplitPortraitUpdatedAt>},
+            {typeof(UserRevision.SplitIpGeolocation), AddRevisionsWithDuplicateIndex<UserRevision.SplitIpGeolocation>}
         });
 
     protected override Uid RevisionEntityIdSelector(BaseUserRevision entity) => entity.Uid;