Add forward spam detection via quote text and external reply photo OCR

Spam forwarded as "reply to external message" was bypassing ML because
the bot only analyzed message.Text/Caption and message.Photo. Now we
extract text from message.Quote and run OCR on message.ExternalReply.Photo,
prepending the result before the original message text so the ML pipeline
sees the full content.

Feature-flagged via FORWARD_SPAM_DETECTION_ENABLED (default: true).

Made-with: Cursor

Author: Szer

src/VahterBanBot.Tests/ContainerTestBase.fs

@@ -128,12 +128,15 @@ type VahterTestContainers(mlEnabled: bool) =
                 .WithEnvironment("REACTION_SPAM_ENABLED", "true")
                 .WithEnvironment("REACTION_SPAM_MIN_MESSAGES", "3")
                 .WithEnvironment("REACTION_SPAM_MAX_REACTIONS", "5")
+                // Forward spam detection
+                .WithEnvironment("FORWARD_SPAM_DETECTION_ENABLED", "true")
                 .Build()
         else
             builder
                 .WithEnvironment("ML_ENABLED", "false")
                 .WithEnvironment("OCR_ENABLED", "false")
                 .WithEnvironment("REACTION_SPAM_ENABLED", "false")
+                .WithEnvironment("FORWARD_SPAM_DETECTION_ENABLED", "false")
                 .Build()
 
     let startContainers() = task {

src/VahterBanBot.Tests/MLBanTests.fs

@@ -494,4 +494,70 @@ type MLBanTests(fixture: MlEnabledVahterTestContainers, _unused: MlAwaitFixture)
         Assert.True(userBanned, "User should be auto-banned after reaching karma threshold via soft spam")
     }
 
+    [<Fact>]
+    let ``Message with spam in quote text triggers auto-delete`` () = task {
+        let msgUpdate = Tg.quickMsg(
+            chat = fixture.ChatsToMonitor[0],
+            text = "hello",
+            quote = Tg.textQuote("2222222")
+        )
+        let! _ = fixture.SendMessage msgUpdate
+
+        let! msgBanned = fixture.MessageIsAutoDeleted msgUpdate.Message
+        Assert.True msgBanned
+    }
+
+    [<Fact>]
+    let ``Message with non-spam quote text does NOT trigger auto-delete`` () = task {
+        let msgUpdate = Tg.quickMsg(
+            chat = fixture.ChatsToMonitor[0],
+            text = "hello",
+            quote = Tg.textQuote("b")
+        )
+        let! _ = fixture.SendMessage msgUpdate
+
+        let! msgBanned = fixture.MessageIsAutoDeleted msgUpdate.Message
+        Assert.False msgBanned
+    }
+
+    [<Fact>]
+    let ``Quote text is prepended to message text`` () = task {
+        let msgUpdate = Tg.quickMsg(
+            chat = fixture.ChatsToMonitor[0],
+            text = "hello",
+            quote = Tg.textQuote("b")
+        )
+        let! _ = fixture.SendMessage msgUpdate
+
+        let! dbMsg = fixture.TryGetDbMessage msgUpdate.Message
+        Assert.True dbMsg.IsSome
+        Assert.Equal("b\nhello", dbMsg.Value.text)
+    }
+
+    [<Fact>]
+    let ``Spam in external reply photo triggers auto-delete via OCR`` () = task {
+        let msgUpdate = Tg.quickMsg(
+            chat = fixture.ChatsToMonitor[0],
+            text = "hello",
+            externalReply = Tg.externalReply(photos = [| Tg.spamPhoto |])
+        )
+        let! _ = fixture.SendMessage msgUpdate
+
+        let! msgBanned = fixture.MessageIsAutoDeleted msgUpdate.Message
+        Assert.True msgBanned
+    }
+
+    [<Fact>]
+    let ``Ham in external reply photo does NOT trigger auto-delete`` () = task {
+        let msgUpdate = Tg.quickMsg(
+            chat = fixture.ChatsToMonitor[0],
+            text = "hello",
+            externalReply = Tg.externalReply(photos = [| Tg.hamPhoto |])
+        )
+        let! _ = fixture.SendMessage msgUpdate
+
+        let! msgBanned = fixture.MessageIsAutoDeleted msgUpdate.Message
+        Assert.False msgBanned
+    }
+
     interface IClassFixture<MlAwaitFixture>

src/VahterBanBot.Tests/TgMessageUtils.fs

@@ -60,7 +60,16 @@ type Tg() =
     static member emoji(?offset: int) = MessageEntity(Type = MessageEntityType.CustomEmoji, Offset = defaultArg offset 0 , Length = 1)
     static member emojies(n: int) = Array.init n (fun i -> Tg.emoji i)
 
-    static member quickMsg (?text: string, ?chat: Chat, ?from: User, ?date: DateTime, ?callback: CallbackQuery, ?caption: string, ?editedText: string, ?entities: MessageEntity[], ?photos: PhotoSize[], ?isAutomaticForward: bool, ?senderChat: Chat) =
+    static member textQuote(text: string) =
+        TextQuote(Text = text, Position = 0)
+
+    static member externalReply(?photos: PhotoSize[], ?chat: Chat) =
+        ExternalReplyInfo(
+            Photo = (photos |> Option.defaultValue null),
+            Chat = (chat |> Option.defaultValue null)
+        )
+
+    static member quickMsg (?text: string, ?chat: Chat, ?from: User, ?date: DateTime, ?callback: CallbackQuery, ?caption: string, ?editedText: string, ?entities: MessageEntity[], ?photos: PhotoSize[], ?isAutomaticForward: bool, ?senderChat: Chat, ?quote: TextQuote, ?externalReply: ExternalReplyInfo) =
         let updateId = next()
         let msgId = next()
         Update(
@@ -77,7 +86,9 @@ type Tg() =
                     Entities = (entities |> Option.defaultValue null),
                     Photo = (photos |> Option.defaultValue null),
                     IsAutomaticForward = (isAutomaticForward |> Option.defaultValue false),
-                    SenderChat = (senderChat |> Option.defaultValue null)
+                    SenderChat = (senderChat |> Option.defaultValue null),
+                    Quote = (quote |> Option.defaultValue null),
+                    ExternalReply = (externalReply |> Option.defaultValue null)
                 ),
             EditedMessage =
                 if editedText |> Option.isSome then

src/VahterBanBot/Bot.fs

@@ -889,6 +889,92 @@ let onMessage
         do! justMessage botUser botClient botConfig logger ml message
 }
 
+let private ocrPhotos
+    (botClient: ITelegramBotClient)
+    (botConfig: BotConfiguration)
+    (computerVision: IComputerVision)
+    (logger: ILogger)
+    (photos: PhotoSize array)
+    (messageId: int) = task {
+    let candidatePhotos =
+        photos
+        |> Array.filter (fun p ->
+            let size = int64 p.FileSize
+            size = 0L || size <= botConfig.OcrMaxFileSizeBytes)
+
+    if candidatePhotos.Length = 0 then
+        logger.LogWarning(
+            "No photos under OCR limit of {LimitBytes} bytes for message {MessageId}",
+            botConfig.OcrMaxFileSizeBytes,
+            messageId)
+        return None
+    else
+        let largestPhoto =
+            candidatePhotos
+            |> Seq.filter (fun p -> p.FileSize.HasValue)
+            |> Seq.maxBy (fun p -> p.FileSize.Value)
+
+        let! file = botClient.GetFile(largestPhoto.FileId)
+
+        if String.IsNullOrWhiteSpace file.FilePath then
+            logger.LogWarning("Failed to resolve file path for photo {PhotoId}", largestPhoto.FileId)
+            return None
+        else
+            let fileUrl = $"https://api.telegram.org/file/bot{botConfig.BotToken}/{file.FilePath}"
+            let! ocrText = computerVision.TextFromImageUrl fileUrl
+            if String.IsNullOrWhiteSpace ocrText then
+                return None
+            else
+                return Some ocrText
+}
+
+let tryEnrichMessageWithForwardedContent
+    (botClient: ITelegramBotClient)
+    (botConfig: BotConfiguration)
+    (computerVision: IComputerVision)
+    (logger: ILogger)
+    (update: Update) = task {
+    if botConfig.ForwardSpamDetectionEnabled then
+        let message = update.EditedOrMessage
+        if not (isNull message) then
+            use activity = botActivity.StartActivity("forwardedContentEnrichment")
+            try
+                let mutable forwardedText: string = null
+
+                if not (isNull message.Quote)
+                   && not (String.IsNullOrWhiteSpace message.Quote.Text) then
+                    forwardedText <- message.Quote.Text
+                    %activity.SetTag("quoteTextLength", message.Quote.Text.Length)
+
+                if botConfig.OcrEnabled
+                   && not (isNull message.ExternalReply)
+                   && not (isNull message.ExternalReply.Photo)
+                   && message.ExternalReply.Photo.Length > 0 then
+                    let! ocrText = ocrPhotos botClient botConfig computerVision logger message.ExternalReply.Photo message.MessageId
+                    match ocrText with
+                    | Some text ->
+                        forwardedText <-
+                            if isNull forwardedText then text
+                            else $"{forwardedText}\n{text}"
+                        %activity.SetTag("externalReplyOcrLength", text.Length)
+                    | None -> ()
+
+                if not (String.IsNullOrWhiteSpace forwardedText) then
+                    let baseText = message.TextOrCaption
+                    let enrichedText =
+                        if String.IsNullOrWhiteSpace baseText then forwardedText
+                        else $"{forwardedText}\n{baseText}"
+                    logger.LogDebug(
+                        "Enriched message {MessageId} with forwarded content of length {ForwardedLength}",
+                        message.MessageId,
+                        forwardedText.Length
+                    )
+                    message.Text <- enrichedText
+                    %activity.SetTag("enrichedTextLength", enrichedText.Length)
+            with ex ->
+                logger.LogError(ex, "Failed to process forwarded content for message {MessageId}", update.EditedOrMessage.MessageId)
+}
+
 let tryEnrichMessageWithOcr
     (botClient: ITelegramBotClient)
     (botConfig: BotConfiguration)
@@ -1222,6 +1308,7 @@ let onUpdate
     elif update.MessageReaction <> null then
         do! onMessageReaction botClient botConfig logger update.MessageReaction
     elif update.EditedOrMessage <> null then
+        do! tryEnrichMessageWithForwardedContent botClient botConfig computerVision logger update
         do! tryEnrichMessageWithOcr botClient botConfig computerVision logger update
         do! onMessage botUser botClient botConfig logger ml update.EditedOrMessage
     elif update.ChatMember <> null || update.MyChatMember <> null then

src/VahterBanBot/Program.fs

@@ -101,7 +101,9 @@ let botConf =
       // Reaction spam detection
       ReactionSpamEnabled = getEnvOr "REACTION_SPAM_ENABLED" "false" |> bool.Parse
       ReactionSpamMinMessages = getEnvOr "REACTION_SPAM_MIN_MESSAGES" "10" |> int
-      ReactionSpamMaxReactions = getEnvOr "REACTION_SPAM_MAX_REACTIONS" "5" |> int }
+      ReactionSpamMaxReactions = getEnvOr "REACTION_SPAM_MAX_REACTIONS" "5" |> int
+      // Forward spam detection
+      ForwardSpamDetectionEnabled = getEnvOr "FORWARD_SPAM_DETECTION_ENABLED" "true" |> bool.Parse }
 
 let validateApiKey (ctx : HttpContext) =
     match ctx.TryGetRequestHeader "X-Telegram-Bot-Api-Secret-Token" with

src/VahterBanBot/Types.fs

@@ -60,7 +60,9 @@ type BotConfiguration =
       // Reaction spam detection
       ReactionSpamEnabled: bool
       ReactionSpamMinMessages: int
-      ReactionSpamMaxReactions: int }
+      ReactionSpamMaxReactions: int
+      // Forward spam detection
+      ForwardSpamDetectionEnabled: bool }
 
 [<CLIMutable>]
 type DbUser =