diff --git a/public/scripts/extensions/tts/index.js b/public/scripts/extensions/tts/index.js
index 035d485c3..ebb83070d 100644
--- a/public/scripts/extensions/tts/index.js
+++ b/public/scripts/extensions/tts/index.js
@@ -5,6 +5,7 @@ import { EdgeTtsProvider } from './edge.js'
import { ElevenLabsTtsProvider } from './elevenlabs.js'
import { SileroTtsProvider } from './silerotts.js'
import { SystemTtsProvider } from './system.js'
+import { NovelTtsProvider } from './novel.js'
const UPDATE_INTERVAL = 1000
@@ -62,6 +63,7 @@ let ttsProviders = {
Silero: SileroTtsProvider,
System: SystemTtsProvider,
Edge: EdgeTtsProvider,
+ Novel: NovelTtsProvider,
}
let ttsProvider
let ttsProviderName
@@ -244,7 +246,7 @@ async function playAudioData(audioBlob) {
window['tts_preview'] = function (id) {
const audio = document.getElementById(id)
- if (!$(audio).data('disabled')) {
+ if (audio && !$(audio).data('disabled')) {
audio.play()
}
else {
@@ -265,7 +267,9 @@ async function onTtsVoicesClick() {
${voice.name}
`
- popupText += ``
+ if (voice.preview_url) {
+ popupText += ``
+ }
}
} catch {
popupText = 'Could not load voices list. Check your API key.'
@@ -327,7 +331,7 @@ function completeCurrentAudioJob() {
*/
async function addAudioJob(response) {
const audioData = await response.blob()
- if (!audioData.type in ['audio/mpeg', 'audio/wav', 'audio/x-wav', 'audio/wave']) {
+ if (!audioData.type in ['audio/mpeg', 'audio/wav', 'audio/x-wav', 'audio/wave', 'audio/webm']) {
throw `TTS received HTTP response with invalid data format. Expecting audio/mpeg, got ${audioData.type}`
}
audioJobQueue.push(audioData)
diff --git a/public/scripts/extensions/tts/novel.js b/public/scripts/extensions/tts/novel.js
new file mode 100644
index 000000000..539c45eda
--- /dev/null
+++ b/public/scripts/extensions/tts/novel.js
@@ -0,0 +1,130 @@
+import { getRequestHeaders } from "../../../script.js"
+import { getPreviewString } from "./index.js"
+
+export { NovelTtsProvider }
+
+class NovelTtsProvider {
+ //########//
+ // Config //
+ //########//
+
+ settings
+ voices = []
+ separator = ' . '
+ audioElement = document.createElement('audio')
+
+ defaultSettings = {
+ voiceMap: {}
+ }
+
+ get settingsHtml() {
+ let html = `Use NovelAI's TTS engine.
+ The Voice IDs in the preview list are only examples, as it can be any string of text. Feel free to try different options!
+ Hint: Save an API key in the NovelAI API settings to use it here.`;
+ return html;
+ }
+
+ onSettingsChange() {
+ }
+
+ loadSettings(settings) {
+ // Populate Provider UI given input settings
+ if (Object.keys(settings).length == 0) {
+ console.info("Using default TTS Provider settings")
+ }
+
+ // Only accept keys defined in defaultSettings
+ this.settings = this.defaultSettings
+
+ for (const key in settings) {
+ if (key in this.settings) {
+ this.settings[key] = settings[key]
+ } else {
+ throw `Invalid setting passed to TTS Provider: ${key}`
+ }
+ }
+
+ console.info("Settings loaded")
+ }
+
+
+ async onApplyClick() {
+ return
+ }
+
+ //#################//
+ // TTS Interfaces //
+ //#################//
+
+ async getVoice(voiceName) {
+ if (!voiceName) {
+ throw `TTS Voice name not provided`
+ }
+
+ return { name: voiceName, voice_id: voiceName, lang: 'en-US', preview_url: false}
+ }
+
+ async generateTts(text, voiceId) {
+ const response = await this.fetchTtsGeneration(text, voiceId)
+ return response
+ }
+
+ //###########//
+ // API CALLS //
+ //###########//
+ async fetchTtsVoiceIds() {
+ const voices = [
+ { name: 'Ligeia', voice_id: 'Ligeia', lang: 'en-US', preview_url: false },
+ { name: 'Aini', voice_id: 'Aini', lang: 'en-US', preview_url: false },
+ { name: 'Orea', voice_id: 'Orea', lang: 'en-US', preview_url: false },
+ { name: 'Claea', voice_id: 'Claea', lang: 'en-US', preview_url: false },
+ { name: 'Lim', voice_id: 'Lim', lang: 'en-US', preview_url: false },
+ { name: 'Aurae', voice_id: 'Aurae', lang: 'en-US', preview_url: false },
+ { name: 'Naia', voice_id: 'Naia', lang: 'en-US', preview_url: false },
+ { name: 'Aulon', voice_id: 'Aulon', lang: 'en-US', preview_url: false },
+ { name: 'Elei', voice_id: 'Elei', lang: 'en-US', preview_url: false },
+ { name: 'Ogma', voice_id: 'Ogma', lang: 'en-US', preview_url: false },
+ { name: 'Raid', voice_id: 'Raid', lang: 'en-US', preview_url: false },
+ { name: 'Pega', voice_id: 'Pega', lang: 'en-US', preview_url: false },
+ { name: 'Lam', voice_id: 'Lam', lang: 'en-US', preview_url: false },
+ ];
+
+ return voices;
+ }
+
+
+ async previewTtsVoice(id) {
+ this.audioElement.pause();
+ this.audioElement.currentTime = 0;
+
+ const text = getPreviewString('en-US')
+ const response = await this.fetchTtsGeneration(text, id)
+ if (!response.ok) {
+ throw new Error(`HTTP ${response.status}`)
+ }
+
+ const audio = await response.blob();
+ const url = URL.createObjectURL(audio);
+ this.audioElement.src = url;
+ this.audioElement.play();
+ }
+
+ async fetchTtsGeneration(inputText, voiceId) {
+ console.info(`Generating new TTS for voice_id ${voiceId}`)
+ const response = await fetch(`/novel_tts`,
+ {
+ method: 'POST',
+ headers: getRequestHeaders(),
+ body: JSON.stringify({
+ "text": inputText,
+ "voice": voiceId,
+ })
+ }
+ )
+ if (!response.ok) {
+ toastr.error(response.statusText, 'TTS Generation Failed');
+ throw new Error(`HTTP ${response.status}`);
+ }
+ return response
+ }
+}
diff --git a/server.js b/server.js
index 20133d1e2..7bc8ec65a 100644
--- a/server.js
+++ b/server.js
@@ -6,7 +6,7 @@ const { hideBin } = require('yargs/helpers');
const net = require("net");
// work around a node v20 bug: https://github.com/nodejs/node/issues/47822#issuecomment-1564708870
if (net.setDefaultAutoSelectFamily) {
- net.setDefaultAutoSelectFamily(false);
+ net.setDefaultAutoSelectFamily(false);
}
const cliArguments = yargs(hideBin(process.argv))
@@ -3197,6 +3197,40 @@ app.post('/google_translate', jsonParser, async (request, response) => {
});
});
+app.post('/novel_tts', jsonParser, async (request, response) => {
+ const token = readSecret(SECRET_KEYS.NOVEL);
+
+ if (!token) {
+ return response.sendStatus(401);
+ }
+
+ const text = request.body.text;
+ const voice = request.body.voice;
+
+ if (!text || !voice) {
+ return response.sendStatus(400);
+ }
+
+ try {
+ const fetch = require('node-fetch').default;
+ const url = `${api_novelai}/ai/generate-voice?text=${encodeURIComponent(text)}&voice=-1&seed=${encodeURIComponent(voice)}&opus=false&version=v2`;
+ const result = await fetch(url, { method: 'GET', headers: { 'Authorization': `Bearer ${token}`, 'Accept': 'audio/webm' } });
+
+ if (!result.ok) {
+ return response.sendStatus(result.status);
+ }
+
+ const chunks = await readAllChunks(result.body);
+ const buffer = Buffer.concat(chunks);
+ response.setHeader('Content-Type', 'audio/webm');
+ return response.send(buffer);
+ }
+ catch (error) {
+ console.error(error);
+ return response.sendStatus(500);
+ }
+});
+
app.post('/delete_sprite', jsonParser, async (request, response) => {
const label = request.body.label;
const name = request.body.name;
@@ -3343,6 +3377,26 @@ function readSecret(key) {
return secrets[key];
}
+async function readAllChunks(readableStream) {
+ return new Promise((resolve, reject) => {
+ // Consume the readable stream
+ const chunks = [];
+ readableStream.on('data', (chunk) => {
+ chunks.push(chunk);
+ });
+
+ readableStream.on('end', () => {
+ console.log('Finished reading the stream.');
+ resolve(chunks);
+ });
+
+ readableStream.on('error', (error) => {
+ console.error('Error while reading the stream:', error);
+ reject();
+ });
+ });
+}
+
async function getImageBuffers(zipFilePath) {
return new Promise((resolve, reject) => {
// Check if the zip file exists