echogarden 2.10.2 → 3.0.0

package/src/synthesis/KokoroTTS.ts

@@ -3,11 +3,11 @@ import { Logger } from '../utilities/Logger.js'
 import { readdir, readFileAsBinary } from '../utilities/FileSystem.js'
 import { joinPath } from '../utilities/PathUtilities.js'
 import { getOnnxSessionOptions, OnnxExecutionProvider } from '../utilities/OnnxUtilities.js'
-import { SynthesisVoice } from '../api/Synthesis.js'
+import { SynthesisCallbacks, SynthesisVoice } from '../api/Synthesis.js'
 import { concatAudioSegments, getRawAudioDuration, RawAudio } from '../audio/AudioUtilities.js'
 import { defaultEspeakOptions, EspeakOptions } from '../synthesis/EspeakTTS.js'
 import { Lexicon } from '../nlp/Lexicon.js'
-import { indexOfLastMatchingNumberInRange } from '../utilities/Utilities.js'
+import { indexOfLastMatchingNumberInRange, yieldToEventLoop } from '../utilities/Utilities.js'
 import { simplifyPunctuationCharacters } from '../nlp/TextNormalizer.js'
 import { getShortLanguageCode } from '../utilities/Locale.js'
 import { substituteStringUsingLookup } from '../utilities/StringUtilities.js'
@@ -21,8 +21,9 @@ export async function synthesizeSentence(
 	lexicons: Lexicon[],
 	modelPath: string,
 	voicesPath: string,
-	executionProviders: OnnxExecutionProvider[]
-) {
+	executionProviders: OnnxExecutionProvider[],
+	callbacks: SynthesisCallbacks) {
+
 	const cacheLookupKey = `${modelPath} ${voicesPath}`
 
 	let kokoroTTS: KokoroTTS | undefined = cachedInstanceLookup.get(cacheLookupKey)
@@ -34,7 +35,13 @@ export async function synthesizeSentence(
 		cachedInstanceLookup.set(cacheLookupKey, kokoroTTS)
 	}
 
-	const result = await kokoroTTS.synthesizeSentence(text, voice, speed, lexicons)
+	const result = await kokoroTTS.synthesizeSentence(
+		text,
+		voice,
+		speed,
+		lexicons,
+		callbacks,
+	)
 
 	return result
 }
@@ -49,10 +56,10 @@ export class KokoroTTS {
 	) {
 	}
 
-	async synthesizeSentence(sentenceText: string, voice: SynthesisVoice, speed: number, lexicons: Lexicon[]) {
-		await this.initializeSessionIfNeeded()
+	async synthesizeSentence(sentenceText: string, voice: SynthesisVoice, speed: number, lexicons: Lexicon[], callbacks: SynthesisCallbacks) {
+		await this.initializeSessionIfNeeded(callbacks)
 
-		const logger = new Logger()
+		const logger = new Logger(callbacks.logLevel)
 
 		const Onnx = await import('onnxruntime-node')
 
@@ -103,12 +110,16 @@ export class KokoroTTS {
 			useKlatt: false
 		}
 
-		//await logger.startAsync('Phonemize text')
-
 		const {
 			referenceSynthesizedAudio,
 			referenceTimeline,
-		} = await Espeak.preprocessAndSynthesize(sentenceText, voiceLanguage, espeakOptions, lexicons)
+		} = await Espeak.preprocessAndSynthesize(
+			sentenceText,
+			voiceLanguage,
+			espeakOptions,
+			lexicons,
+			{ logLevel: callbacks.logLevel },
+		)
 
 		logger.end()
 
@@ -271,6 +282,8 @@ export class KokoroTTS {
 		const audioParts: Float32Array[][] = []
 
 		for (let partIndex = 0; partIndex < parts.length; partIndex++) {
+			callbacks.abortSignal?.throwIfAborted()
+
 			if (parts.length === 1) {
 				await logger.startAsync(`Synthesize sentence with ONNX model`)
 			} else {
@@ -298,6 +311,8 @@ export class KokoroTTS {
 
 		logger.end()
 
+		callbacks.abortSignal?.throwIfAborted()
+
 		const concatenatedAudioParts = audioParts.length > 0 ? concatAudioSegments(audioParts) : [new Float32Array(0)]
 
 		const synthesizedAudio: RawAudio = { audioChannels: concatenatedAudioParts, sampleRate }
@@ -309,19 +324,27 @@ export class KokoroTTS {
 		const referenceWordTimeline = referenceTimeline.flatMap(phrase => phrase.timeline!)
 
 		const dtwWindowDuration = Math.max(5, Math.ceil(0.2 * getRawAudioDuration(synthesizedAudio)))
-		const mappedTimeline = await alignUsingDtw(synthesizedAudio, referenceSynthesizedAudio, referenceWordTimeline, ['high'], [dtwWindowDuration])
+		const mappedTimeline = await alignUsingDtw(
+			synthesizedAudio,
+			referenceSynthesizedAudio,
+			referenceWordTimeline,
+			['high'],
+			[dtwWindowDuration],
+			{ ...callbacks, logLevel: 'warning' },
+		)
 
 		logger.end()
 
 		return { rawAudio: synthesizedAudio, timeline: mappedTimeline }
 	}
 
-	async initializeSessionIfNeeded() {
+	async initializeSessionIfNeeded(callbacks: SynthesisCallbacks) {
 		if (this.session) {
 			return
 		}
 
-		const logger = new Logger()
+		const logger = new Logger(callbacks.logLevel)
+
 		await logger.startAsync('Initialize Kokoro ONNX synthesis model')
 
 		const filesInModelPath = await readdir(this.modelPath)

package/src/synthesis/MicrosoftEdgeTTS.ts

@@ -1,15 +1,16 @@
-import { request } from 'gaxios'
 import WebSocket from 'ws'
 
 import * as AzureCognitiveServicesTTS from './AzureCognitiveServicesTTS.js'
 import * as FFMpegTranscoder from '../codecs/FFMpegTranscoder.js'
 import { Logger } from '../utilities/Logger.js'
 import { OpenPromise } from '../utilities/OpenPromise.js'
-import { concatUint8Arrays, getRandomHexString, logToStderr } from '../utilities/Utilities.js'
+import { concatUint8Arrays, getRandomHexString, logToStderr, sha256AsHex } from '../utilities/Utilities.js'
 import { RawAudio, getEmptyRawAudio, getRawAudioDuration } from '../audio/AudioUtilities.js'
 import { Timer } from '../utilities/Timer.js'
 import { decodeUtf8, encodeUtf8 } from '../encodings/Utf8.js'
 import { escapeHtml } from '../encodings/HtmlEscape.js'
+import { requestHttp } from 'easier-http-request'
+import { VoiceListRequestCallbacks, SynthesisCallbacks } from '../api/API.js'
 
 const traceEnabled = false
 
@@ -21,26 +22,38 @@ export async function synthesize(
 	voice = 'Microsoft Server Speech Text to Speech Voice (en-US, AvaNeural)',
 	ssmlPitchString = '+0Hz',
 	ssmlRateString = '+0%',
-	ssmlVolumeString = '+0%') {
+	ssmlVolumeString = '+0%',
+	callbacks: SynthesisCallbacks) {
+
+	const logger = new Logger(callbacks.logLevel)
 
-	const logger = new Logger()
 	logger.start('Request synthesis from Microsoft Edge cloud API')
 
-	const { audioData, events } = await requestSynthesis(text, trustedClientToken, voice, ssmlPitchString, ssmlRateString, ssmlVolumeString)
-	logger.end()
+	const { audioData, events } = await requestSynthesis(
+		text,
+		trustedClientToken,
+		voice,
+		ssmlPitchString,
+		ssmlRateString,
+		ssmlVolumeString,
+		callbacks,
+	)
 
-	//logToStderr(`Audio length: ${audioData.length}`)
+	logger.end()
 
 	let rawAudio: RawAudio
 
 	try {
-		rawAudio = await FFMpegTranscoder.decodeToChannels(audioData, 24000, 1)
+		rawAudio = await FFMpegTranscoder.decodeToChannels(
+			audioData,
+			24000,
+			1,
+			{ abortSignal: callbacks.abortSignal, logLevel: 'warning' }
+		)
 	} catch (e) {
 		rawAudio = getEmptyRawAudio(1, 24000)
 	}
 
-	//logToStderr(`Raw audio length: ${rawAudio.audioChannels[0].length}`)
-
 	logger.start('Convert boundary events to timeline')
 	const timeline = AzureCognitiveServicesTTS.boundaryEventsToTimeline(events, getRawAudioDuration(rawAudio))
 	logger.end()
@@ -56,7 +69,8 @@ async function requestSynthesis(
 	voice = 'Microsoft Server Speech Text to Speech Voice (en-US, AriaNeural)',
 	ssmlPitchString = '+0Hz',
 	ssmlRateString = '+0%',
-	ssmlVolumeString = '+0%') {
+	ssmlVolumeString = '+0%',
+	callbacks: SynthesisCallbacks) {
 
 	const synthesisOpenPromise = new OpenPromise<SynthesisRequestResult>()
 
@@ -95,8 +109,8 @@ async function requestSynthesis(
 	const requestId = getRandomHexString()
 
 	try {
-		webSocket = await initializeWebsocketConnection(trustedClientToken)
-	} catch(e) {
+		webSocket = await initializeWebsocketConnection(trustedClientToken, callbacks)
+	} catch (e) {
 		connectionFailed = true
 
 		throw e
@@ -212,7 +226,7 @@ async function requestSynthesis(
 
 let existingWebSocketConnection: WebSocket | undefined = undefined
 
-export async function initializeWebsocketConnection(trustedClientToken: string) {
+export async function initializeWebsocketConnection(trustedClientToken: string, callbacks: SynthesisCallbacks) {
 	const requestOpenPromise = new OpenPromise<WebSocket>()
 
 	if (existingWebSocketConnection && existingWebSocketConnection.readyState == WebSocket.OPEN) {
@@ -230,7 +244,9 @@ export async function initializeWebsocketConnection(trustedClientToken: string)
 	const requestURL = 'wss://speech.platform.bing.com/' +
 		'consumer/speech/synthesize/readaloud/edge/v1' +
 		`?TrustedClientToken=${trustedClientToken}` +
-		'&ConnectionId=' + connectionId
+		`&ConnectionId=${connectionId}` +
+		`&Sec-MS-GEC=${await getCurrentSecMsGecToken(trustedClientToken)}` +
+		`&Sec-MS-GEC-Version=1-143.0.3650.75`
 
 	log(requestURL)
 	log('')
@@ -239,9 +255,9 @@ export async function initializeWebsocketConnection(trustedClientToken: string)
 		headers: {
 			'Pragma': 'no-cache',
 			'Cache-Control': 'no-cache',
-			'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.5060.66 Safari/537.36 Edg/103.0.1264.44',
+			'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Safari/537.36 Edg/143.0.0.0',
 			'Origin': 'chrome-extension://jdiccldimpdaibmpdkjnbmckianbfold',
-			'Accept-Encoding': 'gzip, deflate, br',
+			'Accept-Encoding': 'gzip, deflate, br, zstd',
 			'Accept-Language': 'en-US,en;q=0.9',
 		}
 	})
@@ -298,15 +314,23 @@ export async function initializeWebsocketConnection(trustedClientToken: string)
 	webSocket.on('close', onClose)
 	webSocket.on('error', onError)
 
+	callbacks?.abortSignal?.addEventListener('abort', () => {
+		requestOpenPromise.reject(new DOMException('Microsoft Edge TTS WebSocket connection aborted by user.', 'AbortError'))
+		webSocket.terminate()
+	})
+
 	return requestOpenPromise.promise
 }
 
-export async function getVoiceList(trustedClientToken: string) {
-	const response = await request<any>({
+export async function getVoiceList(trustedClientToken: string, callbacks: VoiceListRequestCallbacks) {
+	const response = await requestHttp({
 		method: 'GET',
 
-		url: 'https://speech.platform.bing.com/consumer/speech/synthesize/' +
-			`readaloud/voices/list?trustedclienttoken=${trustedClientToken}`,
+		url: 'https://speech.platform.bing.com/consumer/speech/synthesize/readaloud/voices/list',
+
+		params: {
+			trustedclienttoken: trustedClientToken
+		},
 
 		headers: {
 			'sec-ch-ua': `" Not;A Brand";v="99", "Microsoft Edge";v="103", "Chromium";v="103"`,
@@ -322,10 +346,12 @@ export async function getVoiceList(trustedClientToken: string) {
 			'accept-language': 'en-US,en;q=0.9',
 		},
 
-		responseType: 'json'
+		abortSignal: callbacks?.abortSignal,
 	})
 
-	return response.data as any[]
+	const responseObject = await response.json()
+
+	return responseObject as any[]
 }
 
 function getTimestampString() {
@@ -391,3 +417,25 @@ function indexOfByteSubsequence(sequence: Uint8Array, subsequence: Uint8Array) {
 
 	return -1
 }
+
+async function getCurrentSecMsGecToken(trustedClientToken: string) {
+	const formatted = `${getCurrentWindowsFileTimeTicks()}${trustedClientToken}`
+
+	const hash = (await sha256AsHex(formatted)).toUpperCase()
+
+	return hash
+}
+
+function getCurrentWindowsFileTimeTicks() {
+	const winEpoch100ns = 11644473600n * 10000000n // 116444736000000000n
+	const fiveMin100ns = 300n * 10000000n // 3000000000n
+
+	const currentUnixTime = BigInt(Date.now())
+
+	// Date.now() is ms; 1 ms = 10_000 × 100ns
+	let currentWindowsFileTimeTicks = (currentUnixTime * 10000n) + winEpoch100ns
+
+	currentWindowsFileTimeTicks -= currentWindowsFileTimeTicks % fiveMin100ns
+
+	return currentWindowsFileTimeTicks
+}

package/src/synthesis/OpenAICloudTTS.ts

@@ -1,10 +1,10 @@
 import { ensureRawAudio } from '../audio/AudioUtilities.js';
-import { SynthesisVoice } from '../api/Synthesis.js';
+import { SynthesisCallbacks, SynthesisVoice } from '../api/Synthesis.js';
 import { Logger } from '../utilities/Logger.js';
 import { extendDeep } from '../utilities/ObjectUtilities.js';
 
-export async function synthesize(text: string, voice: string, speed: number, options: OpenAICloudTTSOptions) {
-	const logger = new Logger()
+export async function synthesize(text: string, voice: string, speed: number, options: OpenAICloudTTSOptions, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 
 	logger.start('Request synthesis from OpenAI Cloud API')
 
@@ -18,7 +18,7 @@ export async function synthesize(text: string, voice: string, speed: number, opt
 
 	const openai = new OpenAI(options)
 
-	const result = await openai.audio.speech.create({
+	const response = await openai.audio.speech.create({
 		input: text,
 		model: options.model!,
 		voice: voice as any,
@@ -29,10 +29,11 @@ export async function synthesize(text: string, voice: string, speed: number, opt
 		maxRetries: 10
 	})
 
-	const resultBuffer = await result.buffer()
+	const responseBodyAsArrayBuffer = await response.arrayBuffer()
+	const responseBodyAsBytes = new Uint8Array(responseBodyAsArrayBuffer)
 
 	logger.start('Decode returned audio')
-	const resultRawAudio = ensureRawAudio(resultBuffer)
+	const resultRawAudio = ensureRawAudio(responseBodyAsBytes, undefined, undefined, callbacks)
 
 	logger.end()
 

package/src/synthesis/SamTTS.ts

@@ -1,3 +1,4 @@
+import { SynthesisCallbacks } from '../api/API.js'
 import { RawAudio } from '../audio/AudioUtilities.js'
 import { Logger } from '../utilities/Logger.js'
 
@@ -7,8 +8,8 @@ import { Logger } from '../utilities/Logger.js'
 //
 // https://habr-com.translate.goog/ru/post/500764/?_x_tr_sl=auto&_x_tr_tl=en
 
-export async function synthesize(text: string, pitch = 64, speed = 72, mouth = 128, throat = 128) {
-	const logger = new Logger()
+export async function synthesize(text: string, pitch = 64, speed = 72, mouth = 128, throat = 128, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 	logger.start('Initialize sam module')
 
 	const { default: SamJs } = await import('sam-js')

package/src/synthesis/SapiTTS.ts

@@ -1,4 +1,4 @@
-import { SynthesisVoice } from '../api/API.js'
+import { SynthesisCallbacks, SynthesisVoice } from '../api/API.js'
 import { decodeToChannels, SampleFormat } from '../audio/AudioBufferConversion.js'
 import { RawAudio } from '../audio/AudioUtilities.js'
 import { getShortLanguageCode, lcidToIsoLanguageCode } from '../utilities/Locale.js'
@@ -8,9 +8,9 @@ import { logToStderr } from '../utilities/Utilities.js'
 
 const log = logToStderr
 
-export function synthesize(text: string, voiceName: string, rate = 0, useSpeechPlatform = false) {
+export function synthesize(text: string, voiceName: string, rate = 0, useSpeechPlatform = false, callbacks: SynthesisCallbacks) {
 	return new Promise<{ rawAudio: RawAudio, timeline: Timeline }>(async (resolve, reject) => {
-		const logger = new Logger()
+		const logger = new Logger(callbacks.logLevel)
 		logger.start('Initialize winax module')
 
 		const { default: WinAX } = await import('winax')

package/src/synthesis/SvoxPicoTTS.ts

@@ -1,4 +1,4 @@
-import { SynthesisVoice } from '../api/API.js'
+import { SynthesisCallbacks, SynthesisVoice } from '../api/API.js'
 import { decodeToChannels, SampleFormat } from '../audio/AudioBufferConversion.js'
 import { bandwidthToQFactor } from '../dsp/BiquadFilter.js'
 import { Logger } from '../utilities/Logger.js'
@@ -6,18 +6,17 @@ import { RawAudio } from '../audio/AudioUtilities.js'
 import { readFileAsBinary } from '../utilities/FileSystem.js'
 import { concatUint8Arrays } from '../utilities/Utilities.js'
 
-import { wrapEmscriptenModuleHeap } from 'wasm-heap-manager'
-
 let svoxPicoInstance: any
 
-export async function synthesize(text: string, textAnalysisFilePath: string, signalGenerationFilePath: string, postprocessOutput = true) {
-	const logger = new Logger()
+export async function synthesize(text: string, textAnalysisFilePath: string, signalGenerationFilePath: string, postprocessOutput = true, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 	logger.start('Get pico WASM instance')
 
 	const m = await getInstance()
 
 	logger.start('Initialize pico engine')
 
+	const { wrapEmscriptenModuleHeap } = await import('wasm-heap-manager')
 	const wasmHeap = wrapEmscriptenModuleHeap(m)
 
 	const pico_initialize = m._pico_initialize

package/src/synthesis/VitsTTS.ts

@@ -1,5 +1,5 @@
 import type * as Onnx from 'onnxruntime-node'
-import { SynthesisVoice } from '../api/API.js'
+import { SynthesisCallbacks, SynthesisVoice } from '../api/API.js'
 import { Logger } from '../utilities/Logger.js'
 import { RawAudio, getEmptyRawAudio, getRawAudioDuration } from '../audio/AudioUtilities.js'
 import { Lexicon } from '../nlp/Lexicon.js'
@@ -18,7 +18,8 @@ export async function synthesizeSentence(
 	lengthScale: number,
 	speakerId: number,
 	lexicons: Lexicon[],
-	executionProviders: OnnxExecutionProvider[]) {
+	executionProviders: OnnxExecutionProvider[],
+	callbacks: SynthesisCallbacks) {
 
 	const cacheLookupKey = modelPath
 
@@ -31,7 +32,13 @@ export async function synthesizeSentence(
 		cachedInstanceLookup.set(cacheLookupKey, vitsTTS)
 	}
 
-	const result = await vitsTTS.synthesizeSentence(text, lengthScale, speakerId, lexicons)
+	const result = await vitsTTS.synthesizeSentence(
+		text,
+		lengthScale,
+		speakerId,
+		lexicons,
+		callbacks,
+	)
 
 	return result
 }
@@ -47,8 +54,10 @@ export class VitsTTS {
 		public readonly executionProviders: OnnxExecutionProvider[]) {
 	}
 
-	async synthesizeSentence(sentence: string, lengthScale: number, speakerId = 0, lexicons?: Lexicon[]) {
-		const logger = new Logger()
+	async synthesizeSentence(sentence: string, lengthScale: number, speakerId = 0, lexicons: Lexicon[], callbacks: SynthesisCallbacks) {
+		const logger = new Logger(callbacks.logLevel)
+
+		await logger.startAsync('Initialize VITS ONNX synthesis model')
 
 		await this.initializeIfNeeded()
 
@@ -72,6 +81,7 @@ export class VitsTTS {
 		const Espeak = await import('../synthesis/EspeakTTS.js')
 
 		logger.end()
+		callbacks.abortSignal?.throwIfAborted()
 
 		const espeakOptions: EspeakOptions = {
 			...defaultEspeakOptions,
@@ -79,11 +89,23 @@ export class VitsTTS {
 			useKlatt: false
 		}
 
-		const { referenceSynthesizedAudio, referenceTimeline, fragments, phonemizedFragmentsSubstitutions, phonemizedSentence } = await Espeak.preprocessAndSynthesize(sentence, languageCode, espeakOptions, lexicons)
+		const {
+			referenceSynthesizedAudio,
+			referenceTimeline,
+			fragments,
+			phonemizedFragmentsSubstitutions,
+			phonemizedSentence
+		} = await Espeak.preprocessAndSynthesize(
+			sentence,
+			languageCode,
+			espeakOptions,
+			lexicons,
+			callbacks,
+		)
 
-		if (phonemizedSentence.length == 0) {
-			logger.end()
+		logger.end()
 
+		if (phonemizedSentence.length == 0) {
 			return {
 				rawAudio: getEmptyRawAudio(1, outputSampleRate),
 				timeline: [],
@@ -147,7 +169,9 @@ export class VitsTTS {
 		const bigIntIds = new BigInt64Array(ids.map(id => BigInt(id)))
 		const idLengths = new BigInt64Array([BigInt(bigIntIds.length)])
 
-		await logger.startAsync('Generate audio using synthesis model')
+		logger.end()
+		callbacks.abortSignal?.throwIfAborted()
+		await logger.startAsync('Synthesize audio using model')
 
 		const Onnx = await import('onnxruntime-node')
 
@@ -170,6 +194,8 @@ export class VitsTTS {
 
 		const synthesizedAudio: RawAudio = { audioChannels: [modelOutputAudioSamples], sampleRate: outputSampleRate }
 
+		logger.end()
+		callbacks.abortSignal?.throwIfAborted()
 		await logger.startAsync('Align with reference synthesized audio')
 
 		const { alignUsingDtw } = await import('../alignment/SpeechAlignment.js')
@@ -177,7 +203,14 @@ export class VitsTTS {
 		const referenceWordTimeline = referenceTimeline.flatMap(phrase => phrase.timeline!)
 
 		const dtwWindowDuration = Math.max(5, Math.ceil(0.2 * getRawAudioDuration(synthesizedAudio)))
-		const mappedTimeline = await alignUsingDtw(synthesizedAudio, referenceSynthesizedAudio, referenceWordTimeline, ['high'], [dtwWindowDuration])
+		const mappedTimeline = await alignUsingDtw(
+			synthesizedAudio,
+			referenceSynthesizedAudio,
+			referenceWordTimeline,
+			['high'],
+			[dtwWindowDuration],
+			{ ...callbacks, logLevel: 'warning' },
+		)
 
 		logger.end()
 
@@ -189,9 +222,6 @@ export class VitsTTS {
 			return
 		}
 
-		const logger = new Logger()
-		await logger.startAsync('Initialize VITS ONNX synthesis model')
-
 		const Onnx = await import('onnxruntime-node')
 
 		const filesInModelPath = await readdir(this.modelPath)
@@ -213,8 +243,6 @@ export class VitsTTS {
 		for (const key in this.metadata.phoneme_id_map) {
 			this.phonemeMap.set(key, this.metadata.phoneme_id_map[key])
 		}
-
-		logger.end()
 	}
 }
 

package/src/text-translation/DeepLTextTranslation.ts

@@ -1,11 +1,11 @@
-import { request } from "gaxios"
-import { Logger } from "../utilities/Logger.js"
-import { parseText } from "../nlp/Segmentation.js"
-import { TranslationPair } from "../api/TextTranslation.js"
-import { getChromeOnWindowsHeaders } from "../utilities/BrowserRequestHeaders.js"
+import { Logger } from '../utilities/Logger.js'
+import { parseText } from '../nlp/Segmentation.js'
+import { TextTranslationCallbacks, TranslationPair } from '../api/TextTranslation.js'
+import { getChromeOnWindowsHeaders } from '../utilities/BrowserRequestHeaders.js'
+import { requestHttp } from 'easier-http-request'
 
-export async function translateText(text: string, sourceLanguage: string, targetLanguage: string): Promise<TranslationPair[]> {
-	const logger = new Logger()
+export async function translateText(text: string, sourceLanguage: string, targetLanguage: string, callbacks: TextTranslationCallbacks): Promise<TranslationPair[]> {
+	const logger = new Logger(callbacks.logLevel)
 
 	logger.start(`Prepare request`)
 
@@ -58,7 +58,7 @@ export async function translateText(text: string, sourceLanguage: string, target
 		id: 756456347,
 	}
 
-	const response = await request<any>({
+	const response = await requestHttp({
 		method: 'POST',
 
 		url: `https://www2.deepl.com/jsonrpc`,
@@ -76,11 +76,13 @@ export async function translateText(text: string, sourceLanguage: string, target
 			'Content-Type': 'application/json',
 		},
 
-		body: JSON.stringify(requestBodyObject),
+		body: requestBodyObject,
 
-		responseType: 'json'
+		abortSignal: callbacks?.abortSignal
 	})
 
+	const responseObject = await response.json()
+
 	logger.start('Parse response')
 
 	logger.end()