Merge pull request #3371 from fearmear/release

Decrease TTS generation delay by splitting a message on a new line

public/scripts/extensions/tts/index.js

@@ -120,7 +120,7 @@ async function onNarrateOneMessage() {
     }
 
     resetTtsPlayback();
-    ttsJobQueue.push(message);
+    processAndQueueTtsMessage(message);
     moduleWorker();
 }
 
@@ -147,7 +147,7 @@ async function onNarrateText(args, text) {
     }
 
     resetTtsPlayback();
-    ttsJobQueue.push({ mes: text, name: name });
+    processAndQueueTtsMessage({ mes: text, name: name });
     await moduleWorker();
 
     // Return back to the chat voices
@@ -220,6 +220,36 @@ function isTtsProcessing() {
     return processing;
 }
 
+/**
+ * Splits a message into lines and adds each non-empty line to the TTS job queue.
+ * @param {Object} message - The message object to be processed.
+ * @param {string} message.mes - The text of the message to be split into lines.
+ * @param {string} message.name - The name associated with the message.
+ * @returns {void}
+ */
+function processAndQueueTtsMessage(message) {
+    if (!extension_settings.tts.narrate_by_paragraphs) {
+        ttsJobQueue.push(message);
+        return;
+    }
+
+    const lines = message.mes.split('\n');
+
+    for (let i = 0; i < lines.length; i++) {
+        const line = lines[i];
+
+        if (line.length === 0) {
+            continue;
+        }
+
+        ttsJobQueue.push(
+            Object.assign({}, message, {
+                mes: line,
+            }),
+        );
+    }
+}
+
 function debugTtsPlayback() {
     console.log(JSON.stringify(
         {
@@ -350,7 +380,7 @@ function onAudioControlClicked() {
         talkingAnimation(false);
     } else {
         // Default play behavior if not processing or playing is to play the last message.
-        ttsJobQueue.push(context.chat[context.chat.length - 1]);
+        processAndQueueTtsMessage(context.chat[context.chat.length - 1]);
     }
     updateUiAudioPlayState();
 }
@@ -569,6 +599,7 @@ function loadSettings() {
     $('#tts_narrate_quoted').prop('checked', extension_settings.tts.narrate_quoted_only);
     $('#tts_auto_generation').prop('checked', extension_settings.tts.auto_generation);
     $('#tts_periodic_auto_generation').prop('checked', extension_settings.tts.periodic_auto_generation);
+    $('#tts_narrate_by_paragraphs').prop('checked', extension_settings.tts.narrate_by_paragraphs);
     $('#tts_narrate_translated_only').prop('checked', extension_settings.tts.narrate_translated_only);
     $('#tts_narrate_user').prop('checked', extension_settings.tts.narrate_user);
     $('#tts_pass_asterisks').prop('checked', extension_settings.tts.pass_asterisks);
@@ -638,6 +669,11 @@ function onPeriodicAutoGenerationClick() {
     saveSettingsDebounced();
 }
 
+function onNarrateByParagraphsClick() {
+    extension_settings.tts.narrate_by_paragraphs = !!$('#tts_narrate_by_paragraphs').prop('checked');
+    saveSettingsDebounced();
+}
+
 
 function onNarrateDialoguesClick() {
     extension_settings.tts.narrate_dialogues_only = !!$('#tts_narrate_dialogues').prop('checked');
@@ -816,7 +852,12 @@ async function onMessageEvent(messageId, lastCharIndex) {
     lastChatId = context.chatId;
 
     console.debug(`Adding message from ${message.name} for TTS processing: "${message.mes}"`);
-    ttsJobQueue.push(message);
+
+    if (extension_settings.tts.periodic_auto_generation) {
+        ttsJobQueue.push(message);
+    } else {
+        processAndQueueTtsMessage(message);
+    }
 }
 
 async function onMessageDeleted() {
@@ -1156,6 +1197,7 @@ jQuery(async function () {
         $('#tts_pass_asterisks').on('click', onPassAsterisksClick);
         $('#tts_auto_generation').on('click', onAutoGenerationClick);
         $('#tts_periodic_auto_generation').on('click', onPeriodicAutoGenerationClick);
+        $('#tts_narrate_by_paragraphs').on('click', onNarrateByParagraphsClick);
         $('#tts_narrate_user').on('click', onNarrateUserClick);
 
         $('#playback_rate').on('input', function () {

public/scripts/extensions/tts/settings.html

@@ -30,6 +30,10 @@
                     <input type="checkbox" id="tts_periodic_auto_generation">
                     <small data-i18n="Narrate by paragraphs (when streaming)">Narrate by paragraphs (when streaming)</small>
                 </label>
+                <label class="checkbox_label" for="tts_narrate_by_paragraphs">
+                    <input type="checkbox" id="tts_narrate_by_paragraphs">
+                    <small data-i18n="Narrate by paragraphs (when not streaming)">Narrate by paragraphs (when not streaming)</small>
+                </label>
                 <label class="checkbox_label" for="tts_narrate_quoted">
                     <input type="checkbox" id="tts_narrate_quoted">
                     <small data-i18n="Only narrate quotes">Only narrate "quotes"</small>