Add support for making Text-To-Speech send regular text messages instead of notices

When speech-to-text/flow-type = `only_transcribe`, the bot will now send
text messages by default, not notices.

While notice messages may be less desirable with other bots in the room,
it's probably a better default for most people who enable "transcribe-only" mode.

This is an improvement related to https://github.com/etkecc/baibot/issues/14

src/entity/room_config_context.rs

@@ -4,7 +4,8 @@ use super::globalconfig::GlobalConfig;
 use super::roomconfig::RoomConfig;
 
 use crate::entity::roomconfig::{
-    defaults as roomconfig_defaults, SpeechToTextFlowType, TextGenerationAutoUsage,
+    defaults as roomconfig_defaults, SpeechToTextFlowType,
+    SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages, TextGenerationAutoUsage,
     TextGenerationPrefixRequirementType, TextToSpeechBotMessagesFlowType,
     TextToSpeechUserMessagesFlowType,
 };
@@ -37,6 +38,24 @@ impl RoomConfigContext {
             .unwrap_or(roomconfig_defaults::SPEECH_TO_TEXT_FLOW_TYPE)
     }
 
+    pub fn speech_to_text_msg_type_for_non_threaded_only_transcribed_messages(
+        &self,
+    ) -> SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages {
+        self.room_config
+            .settings
+            .speech_to_text
+            .msg_type_for_non_threaded_only_transcribed_messages
+            .or({
+                self.global_config
+                    .fallback_room_settings
+                    .speech_to_text
+                    .msg_type_for_non_threaded_only_transcribed_messages
+            })
+            .unwrap_or(
+                roomconfig_defaults::SPEECH_TO_TEXT_ONLY_TRANSCRIBE_NON_THREADED_MESSAGE_TYPE,
+            )
+    }
+
     pub fn speech_to_text_language(&self) -> Option<String> {
         self.room_config
             .settings

src/entity/roomconfig/defaults.rs

@@ -1,4 +1,4 @@
-use super::SpeechToTextFlowType;
+use super::{SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages};
 use super::{TextGenerationAutoUsage, TextGenerationPrefixRequirementType};
 use super::{TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType};
 
@@ -15,3 +15,9 @@ pub const TEXT_TO_SPEECH_USER_MESSAGES_FLOW_TYPE: TextToSpeechUserMessagesFlowTy
 
 pub const SPEECH_TO_TEXT_FLOW_TYPE: SpeechToTextFlowType =
     SpeechToTextFlowType::TranscribeAndGenerateText;
+
+// While notice messages may be less desirable with other bots in the room,
+// it's probably a better default for most people who enable "transcribe-only" mode.
+pub const SPEECH_TO_TEXT_ONLY_TRANSCRIBE_NON_THREADED_MESSAGE_TYPE:
+    SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages =
+    SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Text;

src/entity/roomconfig/entity/mod.rs

@@ -13,7 +13,9 @@ mod text_generation;
 mod text_to_speech;
 
 pub use handler::RoomSettingsHandler;
-pub use speech_to_text::SpeechToTextFlowType;
+pub use speech_to_text::{
+    SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+};
 pub use text_generation::{TextGenerationAutoUsage, TextGenerationPrefixRequirementType};
 pub use text_to_speech::{TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType};
 

src/entity/roomconfig/entity/speech_to_text.rs

@@ -4,6 +4,20 @@ use serde::{Deserialize, Serialize};
 pub struct RoomSettingsSpeechToText {
     pub flow_type: Option<SpeechToTextFlowType>,
 
+    /// Controls how the transcribed message is posted when dealing with:
+    /// - messages that only get transcribed (and do not trigger text-generation).
+    ///   See `flow_type` and `SpeechToTextFlowType::OnlyTranscribe` for more details.
+    /// - incoming voice messages that are not part of a thread.
+    ///   For messages that are part of a thread, we need to reply within the thread in a way (with a notice message)
+    ///   that won't confuse the bot later, so we have no choice but to use a notice message.
+    ///
+    /// Text-generation may happen either as a direct result of the incoming voice message or as part of a threaded conversation.
+    /// Transcribed messages should not be attributed to the bot for the purposes of text-generation,
+    /// so any time there's a chance of text-generation happening, we should use `SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Notice`
+    /// and optionally prefix the message with `> 🦻`.
+    pub msg_type_for_non_threaded_only_transcribed_messages:
+        Option<SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages>,
+
     /// The language of the input audio.
     /// Supplying the input language in [ISO-639-1](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) format will improve accuracy and latency.
     pub language: Option<String>,
@@ -57,3 +71,39 @@ impl std::fmt::Display for SpeechToTextFlowType {
         }
     }
 }
+
+#[derive(Clone, Copy, Debug, Deserialize, Serialize, PartialEq)]
+pub enum SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages {
+    /// Send the transcribed message text as a regular message
+    #[serde(rename = "text")]
+    Text,
+
+    /// Send the transcribed message text as a notice message
+    #[serde(rename = "notice")]
+    Notice,
+}
+
+impl SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages {
+    pub fn choices() -> Vec<Self> {
+        vec![Self::Text, Self::Notice]
+    }
+
+    pub fn from_str(s: &str) -> Option<Self> {
+        match s {
+            "text" => Some(Self::Text),
+            "notice" => Some(Self::Notice),
+            _ => None,
+        }
+    }
+}
+
+impl std::fmt::Display for SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match self {
+            SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Text => write!(f, "text"),
+            SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Notice => {
+                write!(f, "notice")
+            }
+        }
+    }
+}

src/entity/roomconfig/entity/text_generation.rs

@@ -70,6 +70,8 @@ pub enum TextGenerationAutoUsage {
     /// Text Generation is to be performed when the original message was sent as audio (voice).
     /// The voice message would be transcribed to text (subject to other configuration)
     /// and text generation would be triggered.
+    ///
+    /// Also see `SpeechToTextFlowType`.
     #[serde(rename = "only_for_voice")]
     OnlyForVoice,
 

src/entity/roomconfig/mod.rs

@@ -5,8 +5,9 @@ use mxlink::helpers::account_data_config::RoomConfigManager as AccountDataRoomCo
 
 pub use entity::{RoomConfig, RoomConfigCarrierContent, RoomSettings, RoomSettingsHandler};
 pub use entity::{
-    SpeechToTextFlowType, TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
-    TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
+    SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+    TextGenerationAutoUsage, TextGenerationPrefixRequirementType, TextToSpeechBotMessagesFlowType,
+    TextToSpeechUserMessagesFlowType,
 };
 
 pub type RoomConfigurationManager =