@iobroker/assistant-satellite 0.0.1

package/build/wakeword.js

@@ -0,0 +1,179 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.WakeWord = void 0;
+/**
+ * OpenWakeWord inference in Node via onnxruntime-node.
+ *
+ * Pipeline (streaming): 16 kHz mono int16 audio → melspectrogram model → sliding embedding model
+ * (Google speech embeddings) → wake-word classifier → score in [0,1].
+ *
+ * Shapes follow the openWakeWord models:
+ *   melspec:   in  [1, samples] (int16 values as float32) → out [1,1,F,32]; then mel = mel/10 + 2
+ *   embedding: in  [1,76,32,1]  → out [1,1,1,96]; window of 76 mel frames, hop 8
+ *   wakeword:  in  [1,16,96]    → out [1,1]     ; last 16 embeddings
+ *
+ * ⚠️ The exact frame counts (mel frames per chunk, warm-up) can vary by model version — validate the
+ * detection threshold on the target device and adjust `wakewordThreshold`.
+ */
+const ort = __importStar(require("onnxruntime-node"));
+const MEL_BINS = 32;
+const EMB_WINDOW = 76; // mel frames per embedding
+const EMB_HOP = 8; // mel frames advanced per embedding
+const WW_FRAMES = 16; // embeddings per wake-word inference
+const EMB_DIM = 96;
+class WakeWord {
+    models;
+    threshold;
+    log;
+    melspec;
+    embedding;
+    classifier;
+    melBuffer = []; // rows of 32 mel bins
+    melProcessed = 0; // mel frames already turned into embeddings
+    embBuffer = []; // rows of 96 embedding dims
+    melShapeLogged = false;
+    embShapeLogged = false;
+    constructor(models, threshold, log) {
+        this.models = models;
+        this.threshold = threshold;
+        this.log = log;
+    }
+    async load() {
+        this.melspec = await ort.InferenceSession.create(this.models.melspec);
+        this.embedding = await ort.InferenceSession.create(this.models.embedding);
+        this.classifier = await ort.InferenceSession.create(this.models.wakeword);
+        this.log.info('Wake-word models loaded.');
+        this.log.info(`  melspec   IO: in=[${this.melspec.inputNames}] out=[${this.melspec.outputNames}]`);
+        this.log.info(`  embedding IO: in=[${this.embedding.inputNames}] out=[${this.embedding.outputNames}]`);
+        this.log.info(`  classifier IO: in=[${this.classifier.inputNames}] out=[${this.classifier.outputNames}]`);
+    }
+    /** Reset the streaming buffers (call after a detection so it re-arms cleanly). */
+    reset() {
+        this.melBuffer = [];
+        this.melProcessed = 0;
+        this.embBuffer = [];
+    }
+    /**
+     * Feed one audio chunk (16-bit signed LE PCM, ideally ~1280 samples = 80 ms) and return the
+     * current wake-word score, or null if not enough context yet.
+     */
+    async process(pcm) {
+        await this.appendMel(pcm);
+        await this.appendEmbeddings();
+        if (this.embBuffer.length < WW_FRAMES) {
+            return null;
+        }
+        return this.classify();
+    }
+    /** True when the given score crosses the configured detection threshold. */
+    triggered(score) {
+        return score !== null && score >= this.threshold;
+    }
+    async appendMel(pcm) {
+        const n = Math.floor(pcm.length / 2);
+        const audio = new Float32Array(n);
+        for (let i = 0; i < n; i++) {
+            audio[i] = pcm.readInt16LE(i * 2); // int16 value as float (openWakeWord convention)
+        }
+        const inName = this.melspec.inputNames[0];
+        const outName = this.melspec.outputNames[0];
+        const out = await this.melspec.run({ [inName]: new ort.Tensor('float32', audio, [1, n]) });
+        const t = out[outName];
+        const data = t.data;
+        if (!this.melShapeLogged) {
+            this.log.info(`melspec output dims for ${n} samples: [${t.dims.join(',')}] (expected [1,1,F,32])`);
+            this.melShapeLogged = true;
+        }
+        // dims [1,1,F,32]
+        const frames = Number(t.dims[2]);
+        for (let f = 0; f < frames; f++) {
+            const row = new Array(MEL_BINS);
+            for (let b = 0; b < MEL_BINS; b++) {
+                row[b] = data[f * MEL_BINS + b] / 10 + 2; // openWakeWord mel transform
+            }
+            this.melBuffer.push(row);
+        }
+    }
+    async appendEmbeddings() {
+        const inName = this.embedding.inputNames[0];
+        const outName = this.embedding.outputNames[0];
+        while (this.melProcessed + EMB_WINDOW <= this.melBuffer.length) {
+            const flat = new Float32Array(EMB_WINDOW * MEL_BINS);
+            for (let f = 0; f < EMB_WINDOW; f++) {
+                const row = this.melBuffer[this.melProcessed + f];
+                for (let b = 0; b < MEL_BINS; b++) {
+                    flat[f * MEL_BINS + b] = row[b];
+                }
+            }
+            const out = await this.embedding.run({
+                [inName]: new ort.Tensor('float32', flat, [1, EMB_WINDOW, MEL_BINS, 1]),
+            });
+            if (!this.embShapeLogged) {
+                this.log.info(`embedding output dims: [${out[outName].dims.join(',')}] (expected [1,1,1,96])`);
+                this.embShapeLogged = true;
+            }
+            const emb = out[outName].data; // [1,1,1,96]
+            this.embBuffer.push(Array.from(emb.slice(0, EMB_DIM)));
+            this.melProcessed += EMB_HOP;
+        }
+        // Trim consumed mel frames to keep memory bounded.
+        if (this.melProcessed > EMB_WINDOW) {
+            const drop = this.melProcessed - EMB_WINDOW;
+            this.melBuffer.splice(0, drop);
+            this.melProcessed -= drop;
+        }
+        if (this.embBuffer.length > WW_FRAMES * 4) {
+            this.embBuffer.splice(0, this.embBuffer.length - WW_FRAMES * 4);
+        }
+    }
+    async classify() {
+        const inName = this.classifier.inputNames[0];
+        const outName = this.classifier.outputNames[0];
+        const window = this.embBuffer.slice(-WW_FRAMES);
+        const flat = new Float32Array(WW_FRAMES * EMB_DIM);
+        for (let f = 0; f < WW_FRAMES; f++) {
+            for (let d = 0; d < EMB_DIM; d++) {
+                flat[f * EMB_DIM + d] = window[f][d];
+            }
+        }
+        const out = await this.classifier.run({
+            [inName]: new ort.Tensor('float32', flat, [1, WW_FRAMES, EMB_DIM]),
+        });
+        return out[outName].data[0];
+    }
+}
+exports.WakeWord = WakeWord;
+//# sourceMappingURL=wakeword.js.map

package/build/wakeword.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"wakeword.js","sourceRoot":"","sources":["../src/wakeword.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;;;;;;;;;;;;;GAaG;AACH,sDAAwC;AAIxC,MAAM,QAAQ,GAAG,EAAE,CAAC;AACpB,MAAM,UAAU,GAAG,EAAE,CAAC,CAAC,2BAA2B;AAClD,MAAM,OAAO,GAAG,CAAC,CAAC,CAAC,oCAAoC;AACvD,MAAM,SAAS,GAAG,EAAE,CAAC,CAAC,qCAAqC;AAC3D,MAAM,OAAO,GAAG,EAAE,CAAC;AAEnB,MAAa,QAAQ;IAYI;IACA;IACA;IAbb,OAAO,CAAwB;IAC/B,SAAS,CAAwB;IACjC,UAAU,CAAwB;IAElC,SAAS,GAAe,EAAE,CAAC,CAAC,sBAAsB;IAClD,YAAY,GAAG,CAAC,CAAC,CAAC,4CAA4C;IAC9D,SAAS,GAAe,EAAE,CAAC,CAAC,4BAA4B;IACxD,cAAc,GAAG,KAAK,CAAC;IACvB,cAAc,GAAG,KAAK,CAAC;IAE/B,YACqB,MAAkB,EAClB,SAAiB,EACjB,GAAW;QAFX,WAAM,GAAN,MAAM,CAAY;QAClB,cAAS,GAAT,SAAS,CAAQ;QACjB,QAAG,GAAH,GAAG,CAAQ;IAC7B,CAAC;IAEJ,KAAK,CAAC,IAAI;QACN,IAAI,CAAC,OAAO,GAAG,MAAM,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QACtE,IAAI,CAAC,SAAS,GAAG,MAAM,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QAC1E,IAAI,CAAC,UAAU,GAAG,MAAM,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC1E,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC;QAC1C,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,uBAAuB,IAAI,CAAC,OAAO,CAAC,UAAU,UAAU,IAAI,CAAC,OAAO,CAAC,WAAW,GAAG,CAAC,CAAC;QACnG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,uBAAuB,IAAI,CAAC,SAAS,CAAC,UAAU,UAAU,IAAI,CAAC,SAAS,CAAC,WAAW,GAAG,CAAC,CAAC;QACvG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,wBAAwB,IAAI,CAAC,UAAU,CAAC,UAAU,UAAU,IAAI,CAAC,UAAU,CAAC,WAAW,GAAG,CAAC,CAAC;IAC9G,CAAC;IAED,kFAAkF;IAClF,KAAK;QACD,IAAI,CAAC,SAAS,GAAG,EAAE,CAAC;QACpB,IAAI,CAAC,YAAY,GAAG,CAAC,CAAC;QACtB,IAAI,CAAC,SAAS,GAAG,EAAE,CAAC;IACxB,CAAC;IAED;;;OAGG;IACH,KAAK,CAAC,OAAO,CAAC,GAAW;QACrB,MAAM,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QAC1B,MAAM,IAAI,CAAC,gBAAgB,EAAE,CAAC;QAC9B,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;YACpC,OAAO,IAAI,CAAC;QAChB,CAAC;QACD,OAAO,IAAI,CAAC,QAAQ,EAAE,CAAC;IAC3B,CAAC;IAED,4EAA4E;IAC5E,SAAS,CAAC,KAAoB;QAC1B,OAAO,KAAK,KAAK,IAAI,IAAI,KAAK,IAAI,IAAI,CAAC,SAAS,CAAC;IACrD,CAAC;IAEO,KAAK,CAAC,SAAS,CAAC,GAAW;QAC/B,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACrC,MAAM,KAAK,GAAG,IAAI,YAAY,CAAC,CAAC,CAAC,CAAC;QAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YACzB,KAAK,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,WAAW,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,iDAAiD;QACxF,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC5C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,MAAM,CAAC,EAAE,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,EAAE,KAAK,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAC3F,MAAM,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,CAAC;QACvB,MAAM,IAAI,GAAG,CAAC,CAAC,IAAoB,CAAC;QACpC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,CAAC;YACvB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,2BAA2B,CAAC,cAAc,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,yBAAyB,CAAC,CAAC;YACnG,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC;QAC/B,CAAC;QACD,kBAAkB;QAClB,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,QAAQ,CAAC,CAAC;YACxC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;gBAChC,GAAG,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,QAAQ,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,6BAA6B;YAC3E,CAAC;YACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC7B,CAAC;IACL,CAAC;IAEO,KAAK,CAAC,gBAAgB;QAC1B,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAC5C,MAAM,OAAO,GAAG,IAAI,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC9C,OAAO,IAAI,CAAC,YAAY,GAAG,UAAU,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,CAAC;YAC7D,MAAM,IAAI,GAAG,IAAI,YAAY,CAAC,UAAU,GAAG,QAAQ,CAAC,CAAC;YACrD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,EAAE,CAAC,EAAE,EAAE,CAAC;gBAClC,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,YAAY,GAAG,CAAC,CAAC,CAAC;gBAClD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;oBAChC,IAAI,CAAC,CAAC,GAAG,QAAQ,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;gBACpC,CAAC;YACL,CAAC;YACD,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;gBACjC,CAAC,MAAM,CAAC,EAAE,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,UAAU,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC;aAC1E,CAAC,CAAC;YACH,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,CAAC;gBACvB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,2BAA2B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,yBAAyB,CAAC,CAAC;gBAC/F,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC;YAC/B,CAAC;YACD,MAAM,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,CAAC,IAAoB,CAAC,CAAC,aAAa;YAC5D,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC;YACvD,IAAI,CAAC,YAAY,IAAI,OAAO,CAAC;QACjC,CAAC;QACD,mDAAmD;QACnD,IAAI,IAAI,CAAC,YAAY,GAAG,UAAU,EAAE,CAAC;YACjC,MAAM,IAAI,GAAG,IAAI,CAAC,YAAY,GAAG,UAAU,CAAC;YAC5C,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;YAC/B,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC;QAC9B,CAAC;QACD,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC;QACpE,CAAC;IACL,CAAC;IAEO,KAAK,CAAC,QAAQ;QAClB,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAC7C,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC/C,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,SAAS,CAAC,CAAC;QAChD,MAAM,IAAI,GAAG,IAAI,YAAY,CAAC,SAAS,GAAG,OAAO,CAAC,CAAC;QACnD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/B,IAAI,CAAC,CAAC,GAAG,OAAO,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACzC,CAAC;QACL,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAClC,CAAC,MAAM,CAAC,EAAE,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,CAAC;SACrE,CAAC,CAAC;QACH,OAAQ,GAAG,CAAC,OAAO,CAAC,CAAC,IAAqB,CAAC,CAAC,CAAC,CAAC;IAClD,CAAC;CACJ;AA9HD,4BA8HC"}

package/config.example.json

@@ -0,0 +1,23 @@
+{
+    "logLevel": "info",
+    "device": "wohnzimmer",
+    "room": "Wohnzimmer",
+
+    "host": "192.168.1.129",
+    "port": 7775,
+    "listenPort": 7776,
+
+    "audioBackend": "auto",
+    "micDevice": "plughw:2,0",
+    "speakerDevice": "plughw:2,0",
+
+    "wakewordModel": "hey_jarvis",
+    "wakewordThreshold": 0.5,
+    "modelsDir": "models",
+
+    "silenceThreshold": 300,
+    "silenceMs": 800,
+    "minRecordMs": 800,
+    "maxRecordMs": 8000,
+    "preBufferChunks": 5
+}

package/package.json

@@ -0,0 +1,35 @@
+{
+    "name": "@iobroker/assistant-satellite",
+    "version": "0.0.1",
+    "description": "Standalone voice satellite for the ioBroker.assistant adapter (wake word + mic streaming + TTS playback over the Hannah UDP protocol). Runs on a bare Pi via npx, no ioBroker required.",
+    "author": "ioBroker",
+    "publishConfig": {
+        "access": "public"
+    },
+    "license": "MIT",
+    "keywords": ["iobroker", "assistant", "voice", "satellite", "wakeword", "openwakeword"],
+    "type": "commonjs",
+    "main": "build/index.js",
+    "types": "build/index.d.ts",
+    "bin": {
+        "assistant-satellite": "build/cli.js"
+    },
+    "engines": {
+        "node": ">=22"
+    },
+    "files": ["build/", "config.example.json", "README.md", "LICENSE"],
+    "scripts": {
+        "build": "tsc -p tsconfig.json",
+        "watch": "tsc -w -p tsconfig.json",
+        "start": "node build/cli.js",
+        "prepublishOnly": "npm run build"
+    },
+    "dependencies": {
+        "mqtt": "^5.10.1",
+        "onnxruntime-node": "^1.20.1"
+    },
+    "devDependencies": {
+        "@types/node": "^20.19.43",
+        "typescript": "^5.7.2"
+    }
+}