echogarden 2.10.2 → 3.0.0

package/src/subtitles/Subtitles.ts

@@ -28,7 +28,7 @@ export function subtitlesToTimeline(subtitles: string, removeMarkup = true) {
 	for (let line of lines) {
 		line = line.trim()
 
-		if (line.length == 0) {
+		if (line.length === 0) {
 			isWithinCue = false
 
 			continue
@@ -52,7 +52,7 @@ export function subtitlesToTimeline(subtitles: string, removeMarkup = true) {
 		} else if (isWithinCue && timeline.length > 0) {
 			const lastEntry = timeline[timeline.length - 1]
 
-			if (lastEntry.text == '') {
+			if (lastEntry.text === '') {
 				lastEntry.text = line
 			} else {
 				lastEntry.text += ' ' + line
@@ -86,7 +86,7 @@ export function timelineToSubtitles(timeline: Timeline, subtitlesConfig?: Subtit
 
 	let config = subtitlesConfig || {}
 
-	if (config.format && config.format == 'webvtt') {
+	if (config.format && config.format === 'webvtt') {
 		config = { ...defaultSubtitlesBaseConfig, ...webVttConfigExtension, ...config }
 	} else {
 		config = { ...defaultSubtitlesBaseConfig, ...srtConfigExtension, ...config }
@@ -97,7 +97,7 @@ export function timelineToSubtitles(timeline: Timeline, subtitlesConfig?: Subtit
 
 	let outText = ''
 
-	if (config.format == 'webvtt') {
+	if (config.format === 'webvtt') {
 		outText += `WEBVTT${lineBreakString}Kind: captions${lineBreakString}`
 
 		if (config.language) {
@@ -110,11 +110,11 @@ export function timelineToSubtitles(timeline: Timeline, subtitlesConfig?: Subtit
 	// Generate the cues from the given timeline
 	let cues: Cue[]
 
-	if (config.mode == 'segment' || config.mode == 'sentence') {
+	if (config.mode === 'segment' || config.mode === 'sentence') {
 		cues = getCuesFromTimeline_IsolateSegmentSentence(timeline, config)
-	} else if (config.mode == 'word' || config.mode == 'phone' || config.mode == 'word+phone') {
+	} else if (config.mode === 'word' || config.mode === 'phone' || config.mode === 'word+phone') {
 		cues = getCuesFromTimeline_IsolateWordPhone(timeline, config)
-	} else if (config.mode == 'line') {
+	} else if (config.mode === 'line') {
 		cues = getCuesFromTimeline_IsolateLines(timeline, config)
 	} else {
 		throw new Error('Invalid subtitles mode.')
@@ -213,7 +213,7 @@ function getCuesFromTimeline_IsolateSegmentSentence(timeline: Timeline, config:
 		for (const wordEntry of wordTimeline) {
 			const wordStartOffset = entryText.indexOf(wordEntry.text, lastWordEndOffset)
 
-			if (wordStartOffset == -1) {
+			if (wordStartOffset === -1) {
 				throw new Error(`Couldn't find word '${wordEntry.text}' in its parent entry text`)
 			}
 
@@ -235,13 +235,13 @@ function getCuesFromTimeline_IsolateSegmentSentence(timeline: Timeline, config:
 		let lineStartOffset = 0
 
 		for (let wordIndex = 0; wordIndex < wordTimeline.length; wordIndex++) {
-			const isLastWord = wordIndex == wordTimeline.length - 1
+			const isLastWord = wordIndex === wordTimeline.length - 1
 
 			const wordEntry = wordTimeline[wordIndex]
 			const wordEndOffset = wordEntry.endOffsetUtf16!
 
 			function getExtendedEndOffset(offset: number | undefined) {
-				if (offset == undefined) {
+				if (offset === undefined) {
 					return entryText.length
 				}
 
@@ -300,14 +300,14 @@ function getCuesFromTimeline_IsolateSegmentSentence(timeline: Timeline, config:
 				currentCue.lines.push(lineText)
 
 				// Update cue start and end times
-				if (currentCue.startTime == -1) {
+				if (currentCue.startTime === -1) {
 					currentCue.startTime = lineStartTime
 				}
 
 				currentCue.endTime = lineEndTime
 
 				// Finalize cue if needed
-				if (isLastWord || currentCue.lines.length == config.maxLineCount) {
+				if (isLastWord || currentCue.lines.length === config.maxLineCount) {
 					cues.push(currentCue)
 
 					currentCue = {
@@ -329,7 +329,7 @@ function getCuesFromTimeline_IsolateSegmentSentence(timeline: Timeline, config:
 
 // Generates cues from timeline. Isolates words or phones in individual cues.
 function getCuesFromTimeline_IsolateWordPhone(timeline: Timeline, config: SubtitlesConfig) {
-	if (timeline.length == 0) {
+	if (timeline.length === 0) {
 		return []
 	}
 
@@ -338,12 +338,12 @@ function getCuesFromTimeline_IsolateWordPhone(timeline: Timeline, config: Subtit
 	const cues: Cue[] = []
 
 	for (const entry of timeline) {
-		const entryIsWord = entry.type == 'word'
-		const entryIsPhone = entry.type == 'phone'
+		const entryIsWord = entry.type === 'word'
+		const entryIsPhone = entry.type === 'phone'
 
 		const shouldIncludeEntry =
-			(entryIsWord && (mode == 'word' || mode == 'word+phone')) ||
-			(entryIsPhone && (mode == 'phone' || mode == 'word+phone'))
+			(entryIsWord && (mode === 'word' || mode === 'word+phone')) ||
+			(entryIsPhone && (mode === 'phone' || mode === 'word+phone'))
 
 		if (shouldIncludeEntry) {
 			cues.push({
@@ -363,7 +363,7 @@ function getCuesFromTimeline_IsolateWordPhone(timeline: Timeline, config: Subtit
 
 // Generates cues from timeline. Isolates lines in individual cues.
 function getCuesFromTimeline_IsolateLines(timeline: Timeline, config: SubtitlesConfig) {
-	if (timeline.length == 0) {
+	if (timeline.length === 0) {
 		return []
 	}
 
@@ -390,7 +390,7 @@ function getCuesFromTimeline_IsolateLines(timeline: Timeline, config: SubtitlesC
 	let currentCueWords: Timeline = []
 
 	function addCueFromCurrentWords() {
-		if (currentCueWords.length == 0) {
+		if (currentCueWords.length === 0) {
 			return
 		}
 
@@ -411,7 +411,7 @@ function getCuesFromTimeline_IsolateLines(timeline: Timeline, config: SubtitlesC
 
 	function addCuesFrom(timeline: Timeline) {
 		for (const entry of timeline) {
-			if (entry.type == 'word') {
+			if (entry.type === 'word') {
 				const currentWordLineNumber = charOffsetToLineNumber[entry.startOffsetUtf16!]
 				const previousWordEntry = currentCueWords[currentCueWords.length - 1]
 
@@ -481,7 +481,7 @@ export function tryParseTimeRangePatternWithoutHours(line: string) {
 }
 
 function cueObjectToText(cue: Cue, cueIndex: number, config: SubtitlesConfig) {
-	if (!cue || !cue.lines || cue.lines.length == 0) {
+	if (!cue || !cue.lines || cue.lines.length === 0) {
 		throw new Error(`Cue is empty`)
 	}
 
@@ -496,7 +496,7 @@ function cueObjectToText(cue: Cue, cueIndex: number, config: SubtitlesConfig) {
 	let formattedStartTime: string
 	let formattedEndTime: string
 
-	if (config.includeHours == true) {
+	if (config.includeHours === true) {
 		formattedStartTime = formatHMS(secondsToHMS(cue.startTime), config.decimalSeparator)
 		formattedEndTime = formatHMS(secondsToHMS(cue.endTime), config.decimalSeparator)
 	} else {

package/src/synthesis/AwsPollyTTS.ts

@@ -4,9 +4,11 @@ import * as FFMpegTranscoder from '../codecs/FFMpegTranscoder.js'
 import { Logger } from '../utilities/Logger.js'
 
 import { readBinaryIncomingMessage } from '../utilities/Utilities.js'
+import { SynthesisCallbacks, VoiceListRequestCallbacks } from '../api/API.js'
+
+export async function synthesize(text: string, language: string | undefined, voice: string, region: string, accessKeyId: string, secretAccessKey: string, engine: 'standard' | 'neural' = 'standard', ssmlEnabled = false, lexiconNames: string[] | undefined, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 
-export async function synthesize(text: string, language: string | undefined, voice: string, region: string, accessKeyId: string, secretAccessKey: string, engine: 'standard' | 'neural' = 'standard', ssmlEnabled = false, lexiconNames?: string[]) {
-	const logger = new Logger()
 	logger.start('Load AWS SDK client module')
 
 	const polly = await import('@aws-sdk/client-polly')
@@ -45,13 +47,18 @@ export async function synthesize(text: string, language: string | undefined, voi
 
 	logger.end()
 
-	const rawAudio = await FFMpegTranscoder.decodeToChannels(audioData as any)
+	const rawAudio = await FFMpegTranscoder.decodeToChannels(
+		audioData as any,
+		undefined,
+		undefined,
+		{ abortSignal: callbacks.abortSignal, logLevel: 'warning' }
+	)
 
 	return { rawAudio }
 }
 
-export async function getVoiceList(region: string, accessKeyId: string, secretAccessKey: string) {
-	const logger = new Logger()
+export async function getVoiceList(region: string, accessKeyId: string, secretAccessKey: string, callbacks: VoiceListRequestCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 	logger.start('Load AWS SDK client module')
 
 	const polly = await import('@aws-sdk/client-polly')

package/src/synthesis/AzureCognitiveServicesTTS.ts

@@ -7,6 +7,7 @@ import { Timeline } from '../utilities/Timeline.js'
 import { RawAudio, getRawAudioDuration } from '../audio/AudioUtilities.js'
 import { concatUint8Arrays } from '../utilities/Utilities.js'
 import { escapeHtml } from '../encodings/HtmlEscape.js'
+import { SynthesisCallbacks } from '../api/Synthesis.js'
 
 export async function synthesize(
 	text: string,
@@ -16,10 +17,12 @@ export async function synthesize(
 	voice = 'Microsoft Server Speech Text to Speech Voice (en-US, AvaNeural)',
 	ssmlEnabled = false,
 	ssmlPitchString = '+0Hz',
-	ssmlRateString = '+0%') {
+	ssmlRateString = '+0%',
+	callbacks: SynthesisCallbacks) {
 
 	return new Promise<{ rawAudio: RawAudio, timeline: Timeline }>((resolve, reject) => {
-		const logger = new Logger()
+		const logger = new Logger(callbacks.logLevel)
+
 		logger.start('Request synthesis from Azure Cognitive Services')
 
 		const speechConfig = SpeechSDK.SpeechConfig.fromSubscription(subscriptionKey, serviceRegion)
@@ -71,7 +74,12 @@ export async function synthesize(
 
 			logger.end()
 
-			const rawAudio = await FFMpegTranscoder.decodeToChannels(encodedAudio, 24000, 1)
+			const rawAudio = await FFMpegTranscoder.decodeToChannels(
+				encodedAudio,
+				24000,
+				1,
+				{ abortSignal: callbacks.abortSignal, logLevel: 'warning' }
+			)
 
 			logger.start('Convert boundary events to a timeline')
 

package/src/synthesis/CoquiServerTTS.ts

@@ -1,14 +1,14 @@
-import { request } from 'gaxios'
+import { SynthesisCallbacks } from '../api/API.js'
 import { decodeWaveToRawAudio } from '../audio/AudioUtilities.js'
+import { requestHttp } from 'easier-http-request'
 import { Logger } from '../utilities/Logger.js'
 import { logToStderr } from '../utilities/Utilities.js'
-const log = logToStderr
 
-export async function synthesize(text: string, speakerId: string | null, serverURL = 'http://[::1]:5002') {
-	const logger = new Logger()
+export async function synthesize(text: string, speakerId: string | null, serverURL = 'http://[::1]:5002', callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
 	logger.start('Request synthesis from Coqui Server')
 
-	const response = await request<Uint8Array>({
+	const response = await requestHttp({
 		url: `${serverURL}/api/tts`,
 
 		params: {
@@ -16,10 +16,11 @@ export async function synthesize(text: string, speakerId: string | null, serverU
 			'speaker_id': speakerId
 		},
 
-		responseType: 'arraybuffer'
+		abortSignal: callbacks?.abortSignal
 	})
 
-	const waveData = new Uint8Array(response.data)
+	const responseBody = await response.arrayBuffer()
+	const waveData = new Uint8Array(responseBody)
 
 	const { rawAudio } = decodeWaveToRawAudio(waveData)
 

package/src/synthesis/DeepgramTTS.ts

@@ -1,22 +1,23 @@
-import { GaxiosResponse, request } from 'gaxios'
-import { SynthesisVoice } from '../api/API.js'
+import { SynthesisCallbacks, SynthesisVoice } from '../api/API.js'
 import * as FFMpegTranscoder from '../codecs/FFMpegTranscoder.js'
 import { Logger } from '../utilities/Logger.js'
 import { logToStderr } from '../utilities/Utilities.js'
 import { extendDeep } from '../utilities/ObjectUtilities.js'
+import { EasierHttpRequestError, requestHttp } from 'easier-http-request'
 
 const log = logToStderr
 
-export async function synthesize(text: string, modelId: string, options: DeepgramTTSOptions) {
-	const logger = new Logger()
+export async function synthesize(text: string, modelId: string, options: DeepgramTTSOptions, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
+
 	logger.start('Request synthesis from Deepgram')
 
 	options = extendDeep(defaultDeepgramTTSOptions, options)
 
-	let response: GaxiosResponse<any>
+	let responseBody: ArrayBuffer
 
 	try {
-		response = await request<any>({
+		const response = await requestHttp({
 			url: `https://api.deepgram.com/v1/speak`,
 
 			params: {
@@ -32,21 +33,21 @@ export async function synthesize(text: string, modelId: string, options: Deepgra
 				'Authorization': `Token ${options.apiKey}`,
 			},
 
-			data: {
+			body: {
 				text,
 			},
 
-			responseType: 'arraybuffer'
+			abortSignal: callbacks?.abortSignal,
 		})
-	} catch (e: any) {
-		const response = e.response
 
-		if (response) {
-			logger.log(`Request failed with status code ${response.status}`)
+		responseBody = await response.arrayBuffer()
+	} catch (e: any) {
+		if (e instanceof EasierHttpRequestError) {
+			logger.log(`Request failed with status code ${e.statusCode}: ${e.statusText}.`)
 
-			if (response.data) {
+			if (e.errorBody) {
 				logger.log(`Server responded with:`)
-				logger.log(response.data)
+				logger.log(e.errorBody)
 			}
 		}
 
@@ -54,7 +55,12 @@ export async function synthesize(text: string, modelId: string, options: Deepgra
 	}
 
 	logger.start('Decode synthesized audio')
-	const rawAudio = await FFMpegTranscoder.decodeToChannels(new Uint8Array(response.data))
+	const rawAudio = await FFMpegTranscoder.decodeToChannels(
+		new Uint8Array(responseBody),
+		undefined,
+		undefined,
+		{ abortSignal: callbacks.abortSignal, logLevel: 'warning' }
+	)
 
 	logger.end()
 

package/src/synthesis/ElevenLabsTTS.ts

@@ -1,5 +1,4 @@
-import { GaxiosResponse, request } from 'gaxios'
-import { SynthesisVoice, VoiceGender } from '../api/API.js'
+import { VoiceListRequestCallbacks, SynthesisCallbacks, SynthesisVoice, VoiceGender } from '../api/API.js'
 import * as FFMpegTranscoder from '../codecs/FFMpegTranscoder.js'
 import { Logger } from '../utilities/Logger.js'
 import { logToStderr } from '../utilities/Utilities.js'
@@ -7,19 +6,21 @@ import { extendDeep } from '../utilities/ObjectUtilities.js'
 import { decodeBase64 } from '../encodings/Base64.js'
 import { splitToWords } from '../nlp/Segmentation.js'
 import { Timeline } from '../utilities/Timeline.js'
+import { EasierHttpRequestError, requestHttp } from 'easier-http-request'
 
 const log = logToStderr
 
-export async function synthesize(text: string, voiceId: string, language: string, options: ElevenLabsTTSOptions) {
-	const logger = new Logger()
+export async function synthesize(text: string, voiceId: string, language: string, options: ElevenLabsTTSOptions, callbacks: SynthesisCallbacks) {
+	const logger = new Logger(callbacks.logLevel)
+
 	logger.start('Request synthesis from ElevenLabs')
 
 	options = extendDeep(defaultElevenLabsTTSOptions, options)
 
-	let response: GaxiosResponse<any>
+	let responseObject: any
 
 	try {
-		response = await request<any>({
+		const response = await requestHttp({
 			url: `https://api.elevenlabs.io/v1/text-to-speech/${voiceId}/with-timestamps`,
 
 			method: 'POST',
@@ -33,7 +34,7 @@ export async function synthesize(text: string, voiceId: string, language: string
 				output_format: 'mp3_44100_64',
 			},
 
-			data: {
+			body: {
 				text,
 
 				model_id: options.modelId,
@@ -48,17 +49,17 @@ export async function synthesize(text: string, voiceId: string, language: string
 				seed: options.seed
 			},
 
-			responseType: 'json'
+			abortSignal: callbacks?.abortSignal,
 		})
-	} catch (e: any) {
-		const response = e.response
 
-		if (response) {
-			logger.log(`Request failed with status code ${response.status}`)
+		responseObject = await response.json()
+	} catch (e: any) {
+		if (e instanceof EasierHttpRequestError) {
+			logger.log(`Request failed with status code ${e.statusCode}: ${e.statusText}.`)
 
-			if (response.data) {
+			if (e.errorBody) {
 				logger.log(`Server responded with:`)
-				logger.log(response.data)
+				logger.log(e.errorBody)
 			}
 		}
 
@@ -66,14 +67,21 @@ export async function synthesize(text: string, voiceId: string, language: string
 	}
 
 	logger.start('Decode synthesized audio')
-	const audioData = decodeBase64(response.data.audio_base64)
-	const rawAudio = await FFMpegTranscoder.decodeToChannels(audioData)
+
+	const audioData = decodeBase64(responseObject.audio_base64)
+
+	const rawAudio = await FFMpegTranscoder.decodeToChannels(
+		audioData,
+		undefined,
+		undefined,
+		{ abortSignal: callbacks.abortSignal, logLevel: 'warning' }
+	)
 
 	let timeline: Timeline | undefined
 
-	const characters: string[] = response.data.alignment?.characters
-	const characterStartTimes: number[] = response.data.alignment?.character_start_times_seconds
-	const characterEndTimes: number[] = response.data.alignment?.character_end_times_seconds
+	const characters: string[] = responseObject.alignment?.characters
+	const characterStartTimes: number[] = responseObject.alignment?.character_start_times_seconds
+	const characterEndTimes: number[] = responseObject.alignment?.character_end_times_seconds
 
 	if (characters && characterStartTimes && characterEndTimes) {
 		logger.start('Create timeline from returned character timings')
@@ -106,8 +114,8 @@ export async function synthesize(text: string, voiceId: string, language: string
 	return { rawAudio, timeline }
 }
 
-export async function getVoiceList(apiKey: string) {
-	const response = await request<any>({
+export async function getVoiceList(apiKey: string, callbacks: VoiceListRequestCallbacks) {
+	const response = await requestHttp({
 		method: 'GET',
 
 		url: 'https://api.elevenlabs.io/v1/voices',
@@ -117,10 +125,12 @@ export async function getVoiceList(apiKey: string) {
 			'xi-api-key': apiKey
 		},
 
-		responseType: 'json'
+		abortSignal: callbacks?.abortSignal
 	})
 
-	const elevenLabsVoices: any[] = response.data.voices
+	const responseObject = await response.json()
+
+	const elevenLabsVoices: any[] = responseObject.voices
 
 	const voices: SynthesisVoice[] = elevenLabsVoices.map(elevenLabsVoice => {
 		const gender: VoiceGender = elevenLabsVoice?.labels?.gender ?? 'unknown'