claude-voice-experiment/lib/pending-query.mjs

const LOUD_STREAK_NEEDED = 3

export class Pending_Query {
	constructor({
		silence_timeout = 6000,
		use_timer = false,
		use_classifier = false,
		classifier = null,
		send_words = new Set(['go', 'done', 'send']),
		cancel_words = new Set(['abort', 'cancel']),
		use_wake_word = false,
		wake_words = new Set(['computer']),
		always_listen_words = new Set(['always listen']),
		stop_listening_words = new Set(['stop listening']),
		mode_query_words = new Set(['mode query']),
		instant_dispatch = new Map([['yes please', 'Yes please.']]),
		pause_words = new Set(['pause']),
		resume_words = new Set(['resume', 'continue']),
		on_submit = null,
		on_cancel = null,
		on_empty_submit = null,
		on_activate = null,
		on_mode_change = null,
		on_mode_query = null,
	} = {}) {
		this.accumulated = ''
		this.silence_timeout = silence_timeout
		this.use_timer = use_timer
		this.use_classifier = use_classifier
		this.classifier = classifier
		this.send_words = send_words
		this.cancel_words = cancel_words
		this.use_wake_word = use_wake_word
		this.wake_words = wake_words
		this.always_listen_words = always_listen_words
		this.stop_listening_words = stop_listening_words
		this.mode_query_words = mode_query_words
		this.instant_dispatch = instant_dispatch
		this.pause_words = pause_words
		this.resume_words = resume_words
		this._paused = false
		this.on_submit = on_submit
		this.on_cancel = on_cancel
		this.on_empty_submit = on_empty_submit
		this.on_activate = on_activate
		this.on_mode_change = on_mode_change
		this.on_mode_query = on_mode_query
		this._active = !use_wake_word
		this._silence_timer = null
		this._loud_streak = 0
	}

	get is_empty() {
		return !this.accumulated.trim()
	}

	_has_real_words(text) {
		return (text.match(/\w/g)?.length ?? 0) >= 1
	}

	append(text) {
		this.accumulated = this.accumulated ? `${this.accumulated}\n${text}` : text
		if (this.use_timer) {
			this._reset_silence_timer()
		}
	}

	async check_classifier() {
		if (!this.use_classifier || !this.classifier || this.is_empty) {
			return false
		}
		const last_line = this.accumulated.split('\n').at(-1)
		return await this.classifier(last_line)
	}

	on_audio(chunk) {
		if (!this.use_timer || this.is_empty) {
			return
		}
		const aligned = (chunk.byteOffset % 2 === 0) ? chunk : Buffer.from(chunk)
		const samples = new Int16Array(aligned.buffer, aligned.byteOffset, aligned.byteLength >> 1)
		let sum = 0
		for (let i = 0; i < samples.length; i++) {
			sum += samples[i] * samples[i]
		}
		const rms = Math.sqrt(sum / samples.length) / 32768
		if (rms > 0.02) {
			this._loud_streak++
			if (this._loud_streak >= LOUD_STREAK_NEEDED) {
				process.stderr.write(`[pending-query] audio reset timer (rms=${rms.toFixed(3)})\n`)
				this._reset_silence_timer()
			}
		} else {
			this._loud_streak = 0
		}
	}

	_collapse_repeated_words(norm) {
		return norm.replace(/\b(\w+)(?:\s+\1){2,}\b/g, '$1')
	}

	async process_utterance(text) {
		const norm = this._collapse_repeated_words(
			text.trim().toLowerCase().replace(/[^a-z ]/g, '').replace(/\s+/g, ' ').trim()
		)

		process.stderr.write(`[pending-query] utterance norm: ${JSON.stringify(norm)}\n`)

		if (this.mode_query_words.has(norm)) {
			process.stderr.write(`[pending-query] mode query — wake_word=${this.use_wake_word}, active=${this._active}\n`)
			this.on_mode_query?.({ use_wake_word: this.use_wake_word, active: this._active })
			return
		}

		if (this.always_listen_words.has(norm)) {
			this.use_wake_word = false
			this.use_timer = false
			this._active = true
			clearTimeout(this._silence_timer)
			this._silence_timer = null
			process.stderr.write('[pending-query] mode: always listening\n')
			this.on_mode_change?.({ use_wake_word: false })
			return
		}

		if (this.stop_listening_words.has(norm)) {
			this.use_wake_word = true
			this.use_timer = true
			this._active = false
			process.stderr.write('[pending-query] mode: wake word\n')
			this.on_mode_change?.({ use_wake_word: true })
			return
		}

		const norm_compact_early = norm.replace(/ /g, '')

		if (this.resume_words.has(norm_compact_early)) {
			if (this._paused) {
				this._paused = false
				process.stderr.write('[pending-query] resumed\n')
				if (this._active && this.use_timer && !this.is_empty) {
					this._reset_silence_timer()
				}
			}
			return
		}

		if (this.pause_words.has(norm_compact_early)) {
			this._paused = true
			clearTimeout(this._silence_timer)
			this._silence_timer = null
			process.stderr.write('[pending-query] paused\n')
			return
		}

		if (this._paused) {
			process.stderr.write(`[pending-query] paused, ignored: ${JSON.stringify(text)}\n`)
			return
		}

		if (this.instant_dispatch.has(norm)) {
			const query = this.instant_dispatch.get(norm)
			process.stderr.write(`[pending-query] instant dispatch: ${JSON.stringify(query)}\n`)
			clearTimeout(this._silence_timer)
			this._silence_timer = null
			this.accumulated = ''
			this._loud_streak = 0
			if (this.use_wake_word) {
				this._active = false
			}
			this.on_submit?.(query)
			return
		}

		const norm_compact = norm_compact_early

		if (!this._active) {
			if (this.wake_words.has(norm_compact)) {
				this._active = true
				process.stderr.write('[pending-query] wake word detected — active\n')
				this.on_activate?.()
			} else {
				process.stderr.write(`[pending-query] idle, ignored: ${JSON.stringify(text)}\n`)
			}
			return
		}

		if (this.cancel_words.has(norm_compact)) {
			this.cancel()
			return
		}
		if (this.send_words.has(norm_compact)) {
			this.submit()
			return
		}
		if (!this._has_real_words(text)) {
			process.stderr.write(`[pending-query] noise fragment dropped: ${JSON.stringify(text)}\n`)
			return
		}
		this.append(text)
		process.stderr.write(`[pending-query] fragment: ${JSON.stringify(this.accumulated)}\n`)
		if (await this.check_classifier()) {
			this.submit()
		}
	}

	submit() {
		clearTimeout(this._silence_timer)
		this._silence_timer = null
		const text = this.accumulated.trim()
		this.accumulated = ''
		this._loud_streak = 0
		if (this.use_wake_word) {
			this._active = false
		}
		if (text) {
			this.on_submit?.(text)
		} else {
			this.on_empty_submit?.()
		}
	}

	cancel() {
		clearTimeout(this._silence_timer)
		this._silence_timer = null
		this.accumulated = ''
		this._loud_streak = 0
		if (this.use_wake_word) {
			this._active = false
		}
		this.on_cancel?.()
	}

	_reset_silence_timer() {
		clearTimeout(this._silence_timer)
		this._silence_timer = setTimeout(() => {
			if (this.is_empty) {
				return
			}
			process.stderr.write('[pending-query] silence timeout — submitting\n')
			this.submit()
		}, this.silence_timeout)
	}
}