refactor: deduplicate IRC parsing

CHANGELOG.md

@@ -113,6 +113,7 @@
 - Dev: Refactored static `MessageBuilder` helpers to standalone functions. (#5652)
 - Dev: Decoupled reply parsing from `MessageBuilder`. (#5660, #5668)
 - Dev: Refactored IRC message building. (#5663)
+- Dev: Unified parsing of historic and live IRC messages. (#5678)
 
 ## 2.5.1
 

src/CMakeLists.txt

@@ -39,6 +39,7 @@ set(SOURCE_FILES
         common/WindowDescriptors.cpp
         common/WindowDescriptors.hpp
 
+        common/enums/MessageContext.hpp
         common/enums/MessageOverflow.hpp
 
         common/network/NetworkCommon.cpp
@@ -281,6 +282,9 @@ set(SOURCE_FILES
         messages/MessageElement.cpp
         messages/MessageElement.hpp
         messages/MessageFlag.hpp
+        messages/MessageSimilarity.cpp
+        messages/MessageSimilarity.hpp
+        messages/MessageSink.hpp
         messages/MessageThread.cpp
         messages/MessageThread.hpp
 
@@ -526,6 +530,8 @@ set(SOURCE_FILES
         util/Twitch.hpp
         util/TypeName.hpp
         util/Variant.hpp
+        util/VectorMessageSink.cpp
+        util/VectorMessageSink.hpp
         util/WidgetHelpers.cpp
         util/WidgetHelpers.hpp
         util/WindowsHelper.cpp

src/common/Channel.cpp

@@ -3,7 +3,9 @@
 #include "Application.hpp"
 #include "messages/Message.hpp"
 #include "messages/MessageBuilder.hpp"
+#include "messages/MessageSimilarity.hpp"
 #include "providers/twitch/IrcMessageHandler.hpp"
+#include "providers/twitch/TwitchIrcServer.hpp"
 #include "singletons/Emotes.hpp"
 #include "singletons/Logging.hpp"
 #include "singletons/Settings.hpp"
@@ -121,10 +123,10 @@ void Channel::addSystemMessage(const QString &contents)
     this->addMessage(msg, MessageContext::Original);
 }
 
-void Channel::addOrReplaceTimeout(MessagePtr message)
+void Channel::addOrReplaceTimeout(MessagePtr message, QTime now)
 {
     addOrReplaceChannelTimeout(
-        this->getMessageSnapshot(), std::move(message), QTime::currentTime(),
+        this->getMessageSnapshot(), std::move(message), now,
         [this](auto /*idx*/, auto msg, auto replacement) {
             this->replaceMessage(msg, replacement);
         },
@@ -287,10 +289,15 @@ void Channel::clearMessages()
 }
 
 MessagePtr Channel::findMessage(QString messageID)
+{
+    return this->findMessageByID(messageID);
+}
+
+MessagePtr Channel::findMessageByID(QStringView messageID)
 {
     MessagePtr res;
 
-    if (auto msg = this->messages_.rfind([&messageID](const MessagePtr &msg) {
+    if (auto msg = this->messages_.rfind([messageID](const MessagePtr &msg) {
             return msg->id == messageID;
         });
         msg)
@@ -301,6 +308,19 @@ MessagePtr Channel::findMessage(QString messageID)
     return res;
 }
 
+void Channel::applySimilarityFilters(const MessagePtr &message) const
+{
+    setSimilarityFlags(message, this->messages_.getSnapshot());
+}
+
+MessageSinkTraits Channel::sinkTraits() const
+{
+    return {
+        MessageSinkTrait::AddMentionsToGlobalChannel,
+        MessageSinkTrait::RequiresKnownChannelPointReward,
+    };
+}
+
 bool Channel::canSendMessage() const
 {
     return false;

src/common/Channel.hpp

@@ -1,8 +1,10 @@
 #pragma once
 
+#include "common/enums/MessageContext.hpp"
 #include "controllers/completion/TabCompletionModel.hpp"
 #include "messages/LimitedQueue.hpp"
 #include "messages/MessageFlag.hpp"
+#include "messages/MessageSink.hpp"
 
 #include <magic_enum/magic_enum.hpp>
 #include <pajlada/signals/signal.hpp>
@@ -26,15 +28,7 @@ enum class TimeoutStackStyle : int {
     Default = DontStackBeyondUserMessage,
 };
 
-/// Context of the message being added to a channel
-enum class MessageContext {
-    /// This message is the original
-    Original,
-    /// This message is a repost of a message that has already been added in a channel
-    Repost,
-};
-
-class Channel : public std::enable_shared_from_this<Channel>
+class Channel : public std::enable_shared_from_this<Channel>, public MessageSink
 {
 public:
     // This is for Lua. See scripts/make_luals_meta.py
@@ -55,7 +49,7 @@ class Channel : public std::enable_shared_from_this<Channel>
     };
 
     explicit Channel(const QString &name, Type type);
-    virtual ~Channel();
+    ~Channel() override;
 
     // SIGNALS
     pajlada::Signals::Signal<const QString &, const QString &, bool &>
@@ -85,17 +79,18 @@ class Channel : public std::enable_shared_from_this<Channel>
     // overridingFlags can be filled in with flags that should be used instead
     // of the message's flags. This is useful in case a flag is specific to a
     // type of split
-    void addMessage(MessagePtr message, MessageContext context,
-                    std::optional<MessageFlags> overridingFlags = std::nullopt);
+    void addMessage(
+        MessagePtr message, MessageContext context,
+        std::optional<MessageFlags> overridingFlags = std::nullopt) final;
     void addMessagesAtStart(const std::vector<MessagePtr> &messages_);
 
     void addSystemMessage(const QString &contents);
 
     /// Inserts the given messages in order by Message::serverReceivedTime.
     void fillInMissingMessages(const std::vector<MessagePtr> &messages);
 
-    void addOrReplaceTimeout(MessagePtr message);
-    void disableAllMessages();
+    void addOrReplaceTimeout(MessagePtr message, QTime now) final;
+    void disableAllMessages() final;
     void replaceMessage(MessagePtr message, MessagePtr replacement);
     void replaceMessage(size_t index, MessagePtr replacement);
     void deleteMessage(QString messageID);
@@ -104,9 +99,14 @@ class Channel : public std::enable_shared_from_this<Channel>
     void clearMessages();
 
     MessagePtr findMessage(QString messageID);
+    MessagePtr findMessageByID(QStringView messageID) final;
 
     bool hasMessages() const;
 
+    void applySimilarityFilters(const MessagePtr &message) const final;
+
+    MessageSinkTraits sinkTraits() const final;
+
     // CHANNEL INFO
     virtual bool canSendMessage() const;
     virtual bool isWritable() const;  // whether split input will be usable

src/common/enums/MessageContext.hpp

@@ -0,0 +1,13 @@
+#pragma once
+
+namespace chatterino {
+
+/// Context of the message being added to a channel
+enum class MessageContext {
+    /// This message is the original
+    Original,
+    /// This message is a repost of a message that has already been added in a channel
+    Repost,
+};
+
+}  // namespace chatterino

src/messages/MessageSimilarity.cpp

@@ -0,0 +1,51 @@
+#include "messages/MessageSimilarity.hpp"
+
+#include <algorithm>
+#include <vector>
+
+namespace chatterino::similarity::detail {
+
+float relativeSimilarity(QStringView str1, QStringView str2)
+{
+    using SizeType = QStringView::size_type;
+
+    // Longest Common Substring Problem
+    std::vector<std::vector<int>> tree(str1.size(),
+                                       std::vector<int>(str2.size(), 0));
+    int z = 0;
+
+    for (SizeType i = 0; i < str1.size(); ++i)
+    {
+        for (SizeType j = 0; j < str2.size(); ++j)
+        {
+            if (str1[i] == str2[j])
+            {
+                if (i == 0 || j == 0)
+                {
+                    tree[i][j] = 1;
+                }
+                else
+                {
+                    tree[i][j] = tree[i - 1][j - 1] + 1;
+                }
+                z = std::max(tree[i][j], z);
+            }
+            else
+            {
+                tree[i][j] = 0;
+            }
+        }
+    }
+
+    // ensure that no div by 0
+    if (z == 0)
+    {
+        return 0.F;
+    }
+
+    auto div = std::max<>({static_cast<SizeType>(1), str1.size(), str2.size()});
+
+    return float(z) / float(div);
+}
+
+}  // namespace chatterino::similarity::detail

src/messages/MessageSimilarity.hpp

@@ -0,0 +1,73 @@
+#pragma once
+
+#include "Application.hpp"
+#include "controllers/accounts/AccountController.hpp"
+#include "messages/Message.hpp"
+#include "providers/twitch/TwitchAccount.hpp"
+#include "singletons/Settings.hpp"
+
+#include <ranges>
+
+namespace chatterino::similarity::detail {
+
+float relativeSimilarity(QStringView str1, QStringView str2);
+
+float inMessages(const MessagePtr &msg,
+                 const std::ranges::bidirectional_range auto &messages)
+{
+    float similarityPercent = 0.0F;
+
+    for (const auto &prevMsg :
+         messages | std::views::reverse |
+             std::views::take(getSettings()->hideSimilarMaxMessagesToCheck))
+    {
+        if (prevMsg->parseTime.secsTo(QTime::currentTime()) >=
+            getSettings()->hideSimilarMaxDelay)
+        {
+            break;
+        }
+        if (getSettings()->hideSimilarBySameUser &&
+            msg->loginName != prevMsg->loginName)
+        {
+            continue;
+        }
+        similarityPercent = std::max(
+            similarityPercent,
+            relativeSimilarity(msg->messageText, prevMsg->messageText));
+    }
+
+    return similarityPercent;
+}
+
+}  // namespace chatterino::similarity::detail
+
+namespace chatterino {
+
+void setSimilarityFlags(const MessagePtr &message,
+                        const std::ranges::bidirectional_range auto &messages)
+{
+    if (getSettings()->similarityEnabled)
+    {
+        bool isMyself =
+            message->loginName ==
+            getApp()->getAccounts()->twitch.getCurrent()->getUserName();
+        bool hideMyself = getSettings()->hideSimilarMyself;
+
+        if (isMyself && !hideMyself)
+        {
+            return;
+        }
+
+        if (similarity::detail::inMessages(message, messages) >
+            getSettings()->similarityPercentage)
+        {
+            message->flags.set(MessageFlag::Similar);
+            if (getSettings()->colorSimilarDisabled)
+            {
+                message->flags.set(MessageFlag::Disabled);
+            }
+        }
+    }
+}
+
+}  // namespace chatterino

src/messages/MessageSink.hpp

@@ -0,0 +1,58 @@
+#pragma once
+
+#include "common/enums/MessageContext.hpp"
+#include "messages/MessageFlag.hpp"
+
+#include <memory>
+
+class QStringView;
+class QTime;
+
+namespace chatterino {
+
+struct Message;
+using MessagePtr = std::shared_ptr<const Message>;
+
+enum class MessageSinkTrait : uint8_t {
+    None = 0,
+    AddMentionsToGlobalChannel = 1 << 0,
+    RequiresKnownChannelPointReward = 1 << 1,
+};
+using MessageSinkTraits = FlagsEnum<MessageSinkTrait>;
+
+/// A generic interface for a managed buffer of `Message`s
+class MessageSink
+{
+public:
+    virtual ~MessageSink() = default;
+
+    /// Add a message to this sink
+    ///
+    /// @param message The message to add (non-null)
+    /// @param ctx The context in which this message is being added.
+    /// @param overridingFlags
+    virtual void addMessage(
+        MessagePtr message, MessageContext ctx,
+        std::optional<MessageFlags> overridingFlags = std::nullopt) = 0;
+
+    /// Adds a timeout message or merges it into an existing one
+    virtual void addOrReplaceTimeout(MessagePtr clearchatMessage,
+                                     QTime now) = 0;
+
+    /// Flags all messages as `Disabled`
+    virtual void disableAllMessages() = 0;
+
+    /// Searches for similar messages and flags this message as similar
+    /// (based on the current settings).
+    virtual void applySimilarityFilters(const MessagePtr &message) const = 0;
+
+    /// @brief Searches for a message by an ID
+    ///
+    /// If there is no message found, an empty shared-pointer is returned.
+    virtual MessagePtr findMessageByID(QStringView id) = 0;
+
+    ///
+    virtual MessageSinkTraits sinkTraits() const = 0;
+};
+
+}  // namespace chatterino