speechflow 0.9.0

package/src/speechflow-logo.svg

@@ -0,0 +1,46 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg id="Layer_1" xmlns="http://www.w3.org/2000/svg" version="1.1" viewBox="0 0 580 286">
+  <!-- Generator: Adobe Illustrator 29.4.0, SVG Export Plug-In . SVG Version: 2.1.0 Build 152)  -->
+  <defs>
+    <style>
+      .st0 {
+        fill: #fff;
+      }
+
+      .st0, .st1 {
+        fill-rule: evenodd;
+      }
+
+      .st1 {
+        fill: #b06820;
+      }
+
+      .st2 {
+        fill: #369;
+      }
+    </style>
+  </defs>
+  <g>
+    <path d="M23.1,90.8c4.1,2.8,9.8,4.8,15.9,4.8,10.8,0,17.5-6.1,17.5-15.1s-4.2-12.9-14.7-17c-11.7-4.1-18.7-10.1-18.7-20s8.8-18.4,21.1-18.4,12,1.8,14.4,3.4l-1.9,4.3c-1.9-1.4-6.6-3.4-12.8-3.4-11.7,0-15.7,7.5-15.7,13.3,0,8.1,4.6,12.2,15.1,16.3,12,4.8,18.3,10.2,18.3,21.1s-7.5,19.9-23.1,19.9-13.6-2-17.3-4.8l1.9-4.3Z"/>
+    <path d="M70,63.7c0-6.5-.2-11.6-.4-16.5h4.8l.4,9.8h.2c3.7-6.9,10-11,19.1-11,13.2,0,22.2,11,22.2,26.2s-11,27.9-23.5,27.9-14-3.3-17.6-9.6h-.2v29.7h-5.1v-56.6ZM75,78.5c0,1.4.2,2.8.4,4.1,2.2,8,9,13.1,17,13.1,11.9,0,18.6-9.6,18.6-23.4s-6.5-22.2-18.2-22.2-14.9,5.3-17.1,13.8c-.3,1.4-.8,3-.8,4.4v10.1Z"/>
+    <path d="M124.8,72.5c0,16.6,9,23.1,19.7,23.1s11.4-1.5,14.3-2.9l1.3,4.1c-1.9,1.1-7.2,3.2-16.3,3.2-14.9,0-24.1-10.8-24.1-26s9.9-28,23.3-28,20.1,15.9,20.1,23.4,0,2.3-.2,3.1h-38.1ZM157.8,68.4c.1-7.2-2.9-18.2-15.4-18.2s-16.4,10.2-17.2,18.2h32.6Z"/>
+    <path d="M171.7,72.5c0,16.6,9,23.1,19.7,23.1s11.4-1.5,14.3-2.9l1.3,4.1c-1.9,1.1-7.2,3.2-16.3,3.2-14.9,0-24.1-10.8-24.1-26s9.9-28,23.3-28,20.1,15.9,20.1,23.4,0,2.3-.2,3.1h-38.1ZM204.6,68.4c.1-7.2-2.9-18.2-15.4-18.2s-16.4,10.2-17.2,18.2h32.6Z"/>
+    <path d="M253.4,96.6c-2.3,1.2-7.6,3.3-15.1,3.3-15,0-24.9-10.8-24.9-26.4s11.2-27.6,26.7-27.6,11.5,1.8,13.6,3.2l-1.8,4.2c-2.6-1.4-6.6-3-12.3-3-13.9,0-21,10.7-21,22.8s8.5,22.3,20.5,22.3,10.1-1.7,12.9-2.9l1.4,4Z"/>
+    <path d="M260.4,22.4h5.1v34h.2c1.5-2.9,4-5.7,7.1-7.5,2.9-1.8,6.5-2.9,10.4-2.9s17.9,2.5,17.9,21.7v31.2h-5.1v-30.6c0-9.4-3.7-17.9-14.2-17.9s-13.5,5.3-15.6,11.7c-.5,1.5-.8,3-.8,5.1v31.7h-5.1V22.4Z"/>
+  </g>
+  <g>
+    <path class="st2" d="M102.1,122.7h71.1v21.6h-44.7v26.6h41.8v21.4h-41.8v46.8h-26.4v-116.4Z"/>
+    <path class="st2" d="M184.5,122.7h26.4v94.3h46.3v22.1h-72.7v-116.4Z"/>
+  </g>
+  <path class="st1" d="M329.8,110.5c-43.9,0-79.7,31.5-79.7,70.3s6.5,32.2,18.3,44.8c2.3,9.6-.5,19.8-7.5,26.9-2.9,2.9-.9,8,3.3,8,13.4,0,26.2-5.2,35.8-14.5,9.5,3.4,19.5,5.1,29.8,5.1,43.9,0,80.3-31.5,80.3-70.3s-36.4-70.3-80.3-70.3Z"/>
+  <g id="Icon">
+    <path class="st0" d="M318.3,174.4v9c0,2.7,1.1,5.2,3,7.1s4.5,3,7.1,3h9c2.7,0,5.2-1.1,7.1-3s3-4.5,3-7.1v-9c0-2.7-1.1-5.2-3-7.1-1.9-1.9-4.5-3-7.1-3h-9c-2.7,0-5.2,1.1-7.1,3s-3,4.5-3,7.1Z"/>
+    <path class="st0" d="M284.7,140.8v9c0,2.7,1.1,5.2,3,7.1s4.5,3,7.1,3h9c2.7,0,5.2-1.1,7.1-3s3-4.5,3-7.1v-9c0-2.7-1.1-5.2-3-7.1s-4.5-3-7.1-3h-9c-2.7,0-5.2,1.1-7.1,3s-3,4.5-3,7.1Z"/>
+    <path class="st0" d="M351.9,208v9c0,2.7,1.1,5.2,3,7.1,1.9,1.9,4.5,3,7.1,3h9c2.7,0,5.2-1.1,7.1-3,1.9-1.9,3-4.5,3-7.1v-9c0-2.7-1.1-5.2-3-7.1s-4.5-3-7.1-3h-9c-2.7,0-5.2,1.1-7.1,3s-3,4.5-3,7.1Z"/>
+    <path class="st0" d="M310.4,148.7h58.3c1.5,0,2.9.6,4,1.6,1,1.1,1.6,2.5,1.6,4v15.7c0,1.5-.6,2.9-1.6,4-1.1,1-2.5,1.6-4,1.6h-13.4c-1.9,0-3.4,1.5-3.4,3.4s1.5,3.4,3.4,3.4h13.4c3.3,0,6.4-1.3,8.7-3.6,2.3-2.3,3.6-5.4,3.6-8.7v-15.7c0-3.3-1.3-6.4-3.6-8.7-2.3-2.3-5.4-3.6-8.7-3.6h-58.3c-1.9,0-3.4,1.5-3.4,3.4s1.5,3.4,3.4,3.4Z"/>
+    <path class="st0" d="M337.3,209.2h-40.3c-1.5,0-2.9-.6-4-1.6-1-1.1-1.6-2.5-1.6-4v-15.7c0-1.5.6-2.9,1.6-4,1.1-1,2.5-1.6,4-1.6h24.6c1.9,0,3.4-1.5,3.4-3.4s-1.5-3.4-3.4-3.4h-24.6c-3.3,0-6.4,1.3-8.7,3.6-2.3,2.3-3.6,5.4-3.6,8.7v15.7c0,3.3,1.3,6.4,3.6,8.7,2.3,2.3,5.4,3.6,8.7,3.6h40.3c1.9,0,3.4-1.5,3.4-3.4s-1.5-3.4-3.4-3.4Z"/>
+    <path class="st0" d="M366.6,185.5l-6.6-6.6,6.6-6.6c1.3-1.3,1.3-3.4,0-4.8-1.3-1.3-3.4-1.3-4.8,0l-9,9c-1.3,1.3-1.3,3.4,0,4.8l9,9c1.3,1.3,3.4,1.3,4.8,0,1.3-1.3,1.3-3.4,0-4.8h0Z"/>
+    <path class="st0" d="M335.2,223.9l9-9c1.3-1.3,1.3-3.4,0-4.8l-9-9c-1.3-1.3-3.4-1.3-4.8,0-1.3,1.3-1.3,3.4,0,4.8l6.6,6.6-6.6,6.6c-1.3,1.3-1.3,3.4,0,4.8,1.3,1.3,3.4,1.3,4.8,0h0Z"/>
+  </g>
+  <path class="st2" d="M432.3,239.1l-27.6-116.4h28.1l8.8,48c2.6,13.8,5,28.8,6.9,40.6h.3c1.9-12.6,4.7-26.6,7.6-40.9l9.8-47.7h28l9.3,49c2.6,13.6,4.5,26.1,6.2,39h.3c1.7-13,4.3-26.6,6.7-40.4l9.5-47.7h26.8l-30,116.4h-28.5l-9.8-50.1c-2.2-11.7-4.1-22.6-5.5-35.9h-.3c-2.1,13.1-4,24.2-6.7,35.9l-11.1,50.1h-28.8Z"/>
+</svg>

package/src/speechflow-node-deepgram.ts

@@ -0,0 +1,102 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import { EventEmitter } from "node:events"
+import Stream           from "node:stream"
+import * as Deepgram    from "@deepgram/sdk"
+import SpeechFlowNode   from "./speechflow-node"
+
+export default class SpeechFlowNodeDevice extends SpeechFlowNode {
+    private dg: Deepgram.LiveClient | null = null
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+        this.configure({
+            key:      { type: "string", val: process.env.SPEECHFLOW_KEY_DEEPGRAM },
+            model:    { type: "string", val: "nova-2", pos: 0 }, /* FIXME: nova-3 multiligual */
+            version:  { type: "string", val: "latest", pos: 1 },
+            language: { type: "string", val: "de",     pos: 2 }
+        })
+    }
+    async open () {
+        this.input  = "audio"
+        this.output = "text"
+        this.stream = null
+
+        /*  sanity check situation  */
+        if (this.config.audioBitDepth !== 16 || !this.config.audioLittleEndian)
+            throw new Error("Deepgram node currently supports PCM-S16LE audio only")
+
+        /*  connect to Deepgram API  */
+        const queue = new EventEmitter()
+        const deepgram = Deepgram.createClient(this.params.key)
+        this.dg = deepgram.listen.live({
+            model:            this.params.model,
+            version:          this.params.version,
+            language:         this.params.language,
+            channels:         this.config.audioChannels,
+            sample_rate:      this.config.audioSampleRate,
+            encoding:         "linear16",
+            multichannel:     false,
+            // endpointing:      false,  /* FIXME: ? */
+            interim_results:  false,
+            smart_format:     true,
+            punctuate:        true,
+            filler_words:     true,
+            diarize:          true,
+            numerals:         true,
+            paragraphs:       true,
+            profanity_filter: true,
+            utterances:       false,
+        })
+        await new Promise((resolve) => {
+            this.dg!.on(Deepgram.LiveTranscriptionEvents.Open, () => {
+                this.log("info", "Deepgram: connection open")
+                resolve(true)
+            })
+        })
+
+        /*  hooks onto Deepgram API events  */
+        this.dg.on(Deepgram.LiveTranscriptionEvents.Close, () => {
+            this.log("info", "Deepgram: connection close")
+        })
+        this.dg.on(Deepgram.LiveTranscriptionEvents.Transcript, async (data) => {
+            const text = data.channel?.alternatives[0].transcript ?? ""
+            if (text === "")
+                return
+            queue.emit("text", text)
+        })
+        this.dg.on(Deepgram.LiveTranscriptionEvents.Error, (error: Error) => {
+            this.log("error", `Deepgram: ${error}`)
+        })
+
+        /*  provide Duplex stream and internally attach to Deepgram API  */
+        const dg = this.dg
+        this.stream = new Stream.Duplex({
+            write (chunk: Buffer, encoding: BufferEncoding, callback: (error?: Error | null | undefined) => void) {
+                const data = chunk.buffer.slice(chunk.byteOffset, chunk.byteOffset + chunk.byteLength)
+                if (data.byteLength === 0)
+                    queue.emit("text", "")
+                else
+                    dg.send(data)
+                callback()
+            },
+            read (size: number) {
+                queue.once("text", (text: string) => {
+                    if (text !== "")
+                        this.push(text)
+                })
+            }
+        })
+    }
+    async close () {
+        if (this.stream !== null) {
+            this.stream.destroy()
+            this.stream = null
+        }
+        if (this.dg !== null)
+            this.dg.requestClose()
+    }
+}

package/src/speechflow-node-deepl.ts

@@ -0,0 +1,76 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import Stream           from "node:stream"
+import { EventEmitter } from "node:events"
+import SpeechFlowNode   from "./speechflow-node"
+import * as DeepL       from "deepl-node"
+
+export default class SpeechFlowNodeDeepL extends SpeechFlowNode {
+    private translator: DeepL.Translator | null = null
+
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+
+        this.input  = "text"
+        this.output = "text"
+        this.stream = null
+
+        this.configure({
+            key: { type: "string", val: process.env.SPEECHFLOW_KEY_DEEPL },
+            src: { type: "string", pos: 0, val: "de",    match: /^(?:de|en-US)$/ },
+            dst: { type: "string", pos: 1, val: "en-US", match: /^(?:de|en-US)$/ }
+        })
+    }
+
+    async open () {
+        /*  instantiate DeepL API SDK  */
+        this.translator = new DeepL.Translator(this.params.key)
+
+        /*  provide text-to-text translation  */
+        const translate = async (text: string) => {
+            const result = await this.translator!.translateText(text, this.params.src, this.params.dst, {
+                splitSentences: "off"
+            })
+            return (result?.text ?? text)
+        }
+
+        /*  establish a duplex stream and connect it to the translation  */
+        const queue = new EventEmitter()
+        this.stream = new Stream.Duplex({
+            write (chunk: Buffer, encoding: BufferEncoding, callback: (error?: Error | null | undefined) => void) {
+                const data = chunk.toString()
+                if (data === "") {
+                    queue.emit("result", "")
+                    callback()
+                }
+                else {
+                    translate(data).then((result) => {
+                        queue.emit("result", result)
+                        callback()
+                    }).catch((err) => {
+                        callback(err)
+                    })
+                }
+            },
+            read (size: number) {
+                queue.once("result", (result: string) => {
+                    this.push(result)
+                })
+            }
+        })
+    }
+
+    async close () {
+        if (this.stream !== null) {
+            this.stream.destroy()
+            this.stream = null
+        }
+        if (this.translator !== null)
+            this.translator = null
+    }
+}
+

package/src/speechflow-node-device.ts

@@ -0,0 +1,96 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import Stream           from "node:stream"
+import PortAudio        from "@gpeng/naudiodon"
+import SpeechFlowNode   from "./speechflow-node"
+import SpeechFlowUtil   from "./speechflow-util"
+
+export default class SpeechFlowNodeDevice extends SpeechFlowNode {
+    private io: PortAudio.IoStreamRead | PortAudio.IoStreamWrite | PortAudio.IoStreamDuplex | null = null
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+        this.configure({
+            device: { type: "string", pos: 0,            match: /^(.+?):(.+)$/ },
+            mode:   { type: "string", pos: 1, val: "rw", match: /^(?:r|w|rw)$/ }
+        })
+    }
+    async open () {
+        /*  determine device  */
+        const device = SpeechFlowUtil.audioDeviceFromURL(this.params.mode, this.params.device)
+
+        /*  sanity check sample rate compatibility
+            (we still do not resample in input/output for simplification reasons)  */
+        if (device.defaultSampleRate !== this.config.audioSampleRate)
+            throw new Error(`device audio sample rate ${device.defaultSampleRate} is ` +
+                `incompatible with required sample rate ${this.config.audioSampleRate}`)
+
+        /*  establish device connection
+            Notice: "naudion" actually implements Stream.{Readable,Writable,Duplex}, but
+            declares just its sub-interface NodeJS.{Readable,Writable,Duplex}Stream,
+            so it is correct to cast it back to Stream.{Readable,Writable,Duplex}  */
+        if (device.maxInputChannels > 0 && device.maxOutputChannels > 0) {
+            this.log("info", `resolved "${this.params.device}" to duplex device "${device.id}"`)
+            this.input  = "audio"
+            this.output = "audio"
+            this.io = PortAudio.AudioIO({
+                inOptions: {
+                    deviceId:     device.id,
+                    channelCount: this.config.audioChannels,
+                    sampleRate:   this.config.audioSampleRate,
+                    sampleFormat: this.config.audioBitDepth
+                },
+                outOptions: {
+                    deviceId:     device.id,
+                    channelCount: this.config.audioChannels,
+                    sampleRate:   this.config.audioSampleRate,
+                    sampleFormat: this.config.audioBitDepth
+                }
+            })
+            this.stream = this.io as unknown as Stream.Duplex
+        }
+        else if (device.maxInputChannels > 0 && device.maxOutputChannels === 0) {
+            this.log("info", `resolved "${this.params.device}" to input device "${device.id}"`)
+            this.input  = "none"
+            this.output = "audio"
+            this.io = PortAudio.AudioIO({
+                inOptions: {
+                    deviceId:     device.id,
+                    channelCount: this.config.audioChannels,
+                    sampleRate:   this.config.audioSampleRate,
+                    sampleFormat: this.config.audioBitDepth
+                }
+            })
+            this.stream = this.io as unknown as Stream.Readable
+        }
+        else if (device.maxInputChannels === 0 && device.maxOutputChannels > 0) {
+            this.log("info", `resolved "${this.params.device}" to output device "${device.id}"`)
+            this.input  = "audio"
+            this.output = "none"
+            this.io = PortAudio.AudioIO({
+                outOptions: {
+                    deviceId:     device.id,
+                    channelCount: this.config.audioChannels,
+                    sampleRate:   this.config.audioSampleRate,
+                    sampleFormat: this.config.audioBitDepth
+                }
+            })
+            this.stream = this.io as unknown as Stream.Writable
+        }
+        else
+            throw new Error(`device "${device.id}" does not have any input or output channels`)
+
+        /*  pass-through errors  */
+        this.io.on("error", (err) => {
+            this.emit("error", err)
+        })
+    }
+    async close () {
+        if (this.io !== null)
+            this.io.quit()
+    }
+}
+

package/src/speechflow-node-elevenlabs.ts

@@ -0,0 +1,99 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import Stream                from "node:stream"
+import { EventEmitter }      from "node:events"
+
+import * as ElevenLabs       from "elevenlabs"
+import { getStreamAsBuffer } from "get-stream"
+
+import SpeechFlowNode        from "./speechflow-node"
+
+/*
+const elevenlabsVoices = {
+    "drew":    { name: "Drew",    model: "eleven_multilingual_v2", lang: [ "en", "de" ] },
+    "george":  { name: "George",  model: "eleven_multilingual_v2", lang: [ "en", "de" ] },
+    "bill":    { name: "Bill",    model: "eleven_multilingual_v2", lang: [ "en", "de" ] },
+    "daniel":  { name: "Daniel",  model: "eleven_multilingual_v1", lang: [ "en", "de" ] },
+    "brian":   { name: "Brian",   model: "eleven_turbo_v2",        lang: [ "en"       ] },
+    "sarah":   { name: "Sarah",   model: "eleven_multilingual_v2", lang: [ "en", "de" ] },
+    "racel":   { name: "Racel",   model: "eleven_multilingual_v2", lang: [ "en", "de" ] },
+    "grace":   { name: "Grace",   model: "eleven_multilingual_v1", lang: [ "en", "de" ] },
+    "matilda": { name: "Matilda", model: "eleven_multilingual_v1", lang: [ "en", "de" ] },
+    "alice":   { name: "Alice",   model: "eleven_turbo_v2",        lang: [ "en"       ] }
+}
+*/
+
+export default class SpeechFlowNodeElevenlabs extends SpeechFlowNode {
+    private elevenlabs: ElevenLabs.ElevenLabsClient | null = null
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+        this.configure({
+            key:      { type: "string", val: process.env.SPEECHFLOW_KEY_ELEVENLABS },
+            voice:    { type: "string", val: "Brian",  pos: 0 },
+            language: { type: "string", val: "de",     pos: 1 }
+        })
+    }
+    async open () {
+        this.input  = "text"
+        this.output = "audio"
+
+        this.elevenlabs = new ElevenLabs.ElevenLabsClient({
+            apiKey: this.params.key
+        })
+        const voices = await this.elevenlabs.voices.getAll()
+        const voice = voices.voices.find((voice) => voice.name === this.params.voice)
+        if (voice === undefined)
+            throw new Error(`invalid ElevenLabs voice "${this.params.voice}"`)
+        const speechStream = (text: string) => {
+            return this.elevenlabs!.textToSpeech.convert(voice.voice_id, {
+                text,
+                optimize_streaming_latency: 2,
+                output_format: "pcm_16000", // S16LE
+                model_id: "eleven_flash_v2_5",
+                /*
+                voice_settings: {
+                    stability: 0,
+                    similarity_boost: 0
+                }
+                */
+            }, {
+                timeoutInSeconds: 30,
+                maxRetries: 10
+            })
+        }
+        const queue = new EventEmitter()
+        this.stream = new Stream.Duplex({
+            write (chunk: Buffer, encoding: BufferEncoding, callback: (error?: Error | null | undefined) => void) {
+                if (encoding !== "utf8" && encoding !== "utf-8")
+                    callback(new Error("only text input supported by Elevenlabs node"))
+                const data = chunk.toString()
+                speechStream(data).then((stream) => {
+                    getStreamAsBuffer(stream).then((buffer) => {
+                        queue.emit("audio", buffer)
+                        callback()
+                    }).catch((error) => {
+                        callback(error)
+                    })
+                }).catch((error) => {
+                    callback(error)
+                })
+            },
+            read (size: number) {
+                queue.once("audio", (buffer: Buffer) => {
+                    this.push(buffer, "binary")
+                })
+            }
+        })
+    }
+    async close () {
+        if (this.stream !== null) {
+            this.stream.destroy()
+            this.stream = null
+        }
+    }
+}
+

package/src/speechflow-node-file.ts

@@ -0,0 +1,46 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import fs               from "node:fs"
+import SpeechFlowNode   from "./speechflow-node"
+
+export default class SpeechFlowNodeDevice extends SpeechFlowNode {
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+        this.configure({
+            path: { type: "string", pos: 0 },
+            mode: { type: "string", pos: 1, val: "r",     match: /^(?:r|w|rw)$/ },
+            type: { type: "string", pos: 2, val: "audio", match: /^(?:audio|text)$/ }
+        })
+    }
+    async open () {
+        if (this.params.mode === "r") {
+            this.output = this.params.type
+            if (this.params.path === "-")
+                this.stream = process.stdin
+            else
+                this.stream = fs.createReadStream(this.params.path,
+                    { encoding: this.params.type === "text" ? this.config.textEncoding : "binary" })
+        }
+        else if (this.params.mode === "w") {
+            this.input = this.params.type
+            if (this.params.path === "-")
+                this.stream = process.stdout
+            else
+                this.stream = fs.createWriteStream(this.params.path,
+                    { encoding: this.params.type === "text" ? this.config.textEncoding : "binary" })
+        }
+        else
+            throw new Error(`invalid file mode "${this.params.mode}"`)
+    }
+    async close () {
+        if (this.stream !== null && this.params.path !== "-") {
+            this.stream.destroy()
+            this.stream = null
+        }
+    }
+}
+

package/src/speechflow-node-websocket.ts

@@ -0,0 +1,140 @@
+/*
+**  SpeechFlow - Speech Processing Flow Graph
+**  Copyright (c) 2024-2025 Dr. Ralf S. Engelschall <rse@engelschall.com>
+**  Licensed under GPL 3.0 <https://spdx.org/licenses/GPL-3.0-only>
+*/
+
+import Stream           from "node:stream"
+import ws               from "ws"
+import ReconnWebsocket, { ErrorEvent }  from "@opensumi/reconnecting-websocket"
+import SpeechFlowNode   from "./speechflow-node"
+
+export default class SpeechFlowNodeWebsocket extends SpeechFlowNode {
+    private server: ws.WebSocketServer | null = null
+    private client: WebSocket | null = null
+    constructor (id: string, opts: { [ id: string ]: any }, args: any[]) {
+        super(id, opts, args)
+        this.configure({
+            listen:  { type: "string", val: "",     match: /^(?:|ws:\/\/(.+?):(\d+))$/ },
+            connect: { type: "string", val: "",     match: /^(?:|ws:\/\/(.+?):(\d+)(?:\/.*)?)$/ },
+            type:    { type: "string", val: "text", match: /^(?:audio|text)$/ }
+        })
+    }
+    async open () {
+        this.input  = this.params.type
+        this.output = this.params.type
+        if (this.params.listen !== "") {
+            const url = new URL(this.params.listen)
+            let websocket: ws.WebSocket | null = null
+            const server = new ws.WebSocketServer({
+                host: url.hostname,
+                port: Number.parseInt(url.port),
+                path: url.pathname
+            })
+            server.on("listening", () => {
+                this.log("info", `listening on URL ${this.params.listen}`)
+            })
+            server.on("connection", (ws, request) => {
+                this.log("info", `connection opened on URL ${this.params.listen}`)
+                websocket = ws
+            })
+            server.on("close", () => {
+                this.log("info", `connection closed on URL ${this.params.listen}`)
+                websocket = null
+            })
+            server.on("error", (error) => {
+                this.log("error", `error on URL ${this.params.listen}: ${error.message}`)
+                websocket = null
+            })
+            this.stream = new Stream.Duplex({
+                write (chunk: Buffer, encoding: BufferEncoding, callback: (error?: Error | null | undefined) => void) {
+                    const data = chunk.buffer.slice(chunk.byteOffset, chunk.byteOffset + chunk.byteLength)
+                    if (websocket !== null) {
+                        websocket.send(data, (error) => {
+                            if (error) callback(error)
+                            else       callback()
+                        })
+                    }
+                    else
+                        callback(new Error("still no Websocket connection available"))
+                },
+                read (size: number) {
+                    if (websocket !== null) {
+                        websocket.once("message", (data, isBinary) => {
+                            this.push(data, isBinary ? "binary" : "utf8")
+                        })
+                    }
+                    else
+                        throw new Error("still no Websocket connection available")
+                }
+            })
+        }
+        else if (this.params.connect !== "") {
+            this.client = new ReconnWebsocket(this.params.connect, [], {
+                WebSocket:                   ws,
+                WebSocketOptions:            {},
+                reconnectionDelayGrowFactor: 1.3,
+                maxReconnectionDelay:        4000,
+                minReconnectionDelay:        1000,
+                connectionTimeout:           4000,
+                minUptime:                   5000
+            })
+            this.client.addEventListener("open", (ev: Event) => {
+                this.log("info", `connection opened on URL ${this.params.connect}`)
+            })
+            this.client.addEventListener("close", (ev: Event) => {
+                this.log("info", `connection closed on URL ${this.params.connect}`)
+            })
+            this.client.addEventListener("error", (ev: ErrorEvent) => {
+                this.log("error", `error on URL ${this.params.connect}: ${ev.error.message}`)
+            })
+            const client = this.client
+            client.binaryType = "arraybuffer"
+            this.stream = new Stream.Duplex({
+                write (chunk: Buffer, encoding: BufferEncoding, callback: (error?: Error | null | undefined) => void) {
+                    const data = chunk.buffer.slice(chunk.byteOffset, chunk.byteOffset + chunk.byteLength)
+                    if (client.OPEN) {
+                        client.send(data)
+                        callback()
+                    }
+                    else
+                        callback(new Error("still no Websocket connection available"))
+                },
+                read (size: number) {
+                    if (client.OPEN) {
+                        client.addEventListener("message", (ev: MessageEvent) => {
+                            if (ev.data instanceof ArrayBuffer)
+                                this.push(ev.data, "binary")
+                            else
+                                this.push(ev.data, "utf8")
+                        }, { once: true })
+                    }
+                    else
+                        throw new Error("still no Websocket connection available")
+                }
+            })
+        }
+        else
+            throw new Error("neither listen nor connect mode requested")
+    }
+    async close () {
+        if (this.server !== null) {
+            await new Promise<void>((resolve, reject) => {
+                this.server!.close((error) => {
+                    if (error) reject(error)
+                    else       resolve()
+                })
+            })
+            this.server = null
+        }
+        if (this.client !== null) {
+            this.client!.close()
+            this.client = null
+        }
+        if (this.stream !== null) {
+            this.stream.destroy()
+            this.stream = null
+        }
+    }
+}
+