@storyteller-platform/ghost-story 0.0.1

package/dist/vad/Silero.cjs

@@ -0,0 +1,174 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var Silero_exports = {};
+__export(Silero_exports, {
+  detectVoiceActivity: () => detectVoiceActivity,
+  ensureVadInstalled: () => ensureVadInstalled,
+  segmentsToTimeline: () => segmentsToTimeline
+});
+module.exports = __toCommonJS(Silero_exports);
+var import_node_child_process = require("node:child_process");
+var import_node_fs = require("node:fs");
+var import_fs_extra = require("fs-extra");
+var import_config = require("../cli/config.cjs");
+var import_install = require("../cli/install.cjs");
+const defaultOptions = {
+  printOutput: false,
+  autoInstall: true,
+  threshold: 0.5,
+  minSpeechDurationMs: 250,
+  minSilenceDurationMs: 100,
+  speechPadMs: 30
+};
+async function ensureVadInstalled(options = {}) {
+  const opts = { ...defaultOptions, ...options };
+  const modelDir = opts.modelDir ?? (0, import_config.getModelDir)();
+  const installDir = opts.installDir ?? (0, import_config.getInstallDir)();
+  await (0, import_fs_extra.ensureDir)(modelDir);
+  if (opts.autoInstall) {
+    await (0, import_install.installBinary)({ printOutput: opts.printOutput });
+    await (0, import_install.installVadModel)({ modelDir, printOutput: opts.printOutput });
+  }
+  const vadModelPath = (0, import_config.getVadModelPath)(modelDir);
+  if (!(0, import_node_fs.existsSync)(vadModelPath)) {
+    throw new Error(
+      `VAD model not found at ${vadModelPath}. Run 'ghost-story install --vad' to install.`
+    );
+  }
+  return { installDir, vadModelPath };
+}
+async function detectVoiceActivity(inputPath, options = {}) {
+  const opts = { ...defaultOptions, ...options };
+  if (!(0, import_node_fs.existsSync)(inputPath)) {
+    throw new Error(`Input file does not exist: ${inputPath}`);
+  }
+  const { installDir, vadModelPath } = await ensureVadInstalled(opts);
+  const executablePath = (0, import_config.getVadExecutablePath)(installDir);
+  if (!(0, import_node_fs.existsSync)(executablePath)) {
+    throw new Error(
+      `VAD executable not found at ${executablePath}. Run 'storywhisper install binary' to install.`
+    );
+  }
+  const pathToInput = inputPath;
+  const args = [
+    "-f",
+    pathToInput,
+    "-vm",
+    vadModelPath,
+    "-vt",
+    String(opts.threshold),
+    // vspd is not supported for some reason
+    // "-vspd",
+    // String(opts.minSpeechDurationMs),
+    "-vsd",
+    String(opts.minSilenceDurationMs),
+    "-vp",
+    String(opts.speechPadMs)
+  ];
+  if (opts.printOutput) {
+    console.log(`Running VAD: ${executablePath} ${args.join(" ")}`);
+  }
+  const segments = await runVadProcess(
+    executablePath,
+    args,
+    installDir,
+    opts.printOutput
+  );
+  return segments;
+}
+async function runVadProcess(executable, args, cwd, printOutput) {
+  if (printOutput) {
+    console.log(`Running VAD: ${executable} ${args.join(" ")}`);
+  }
+  return new Promise((resolve, reject) => {
+    const proc = (0, import_node_child_process.spawn)(executable, args, { cwd });
+    let stdout = "";
+    let stderr = "";
+    proc.stdout.on("data", (data) => {
+      const str = data.toString("utf-8");
+      stdout += str;
+      if (printOutput) {
+        process.stdout.write(data);
+      }
+    });
+    proc.stderr.on("data", (data) => {
+      const str = data.toString("utf-8");
+      stderr += str;
+      if (printOutput) {
+        process.stderr.write(data);
+      }
+    });
+    proc.on("exit", (code, signal) => {
+      if (printOutput) {
+        console.log(`VAD process exited with code ${code} and signal ${signal}`);
+      }
+      const segments = parseVadOutput(stdout + stderr);
+      if (printOutput) {
+        console.log(segments);
+      }
+      if (code !== 0 && segments.length === 0) {
+        reject(
+          new Error(
+            `VAD process exited with code ${code}: ${stderr || stdout}`
+          )
+        );
+        return;
+      }
+      resolve(segments);
+    });
+    proc.on("error", (err) => {
+      if (printOutput) {
+        console.error(err);
+      }
+      reject(new Error(`Failed to start VAD process: ${err.message}`));
+    });
+  });
+}
+function parseVadOutput(output) {
+  const segments = [];
+  const segmentRegex = /Speech segment (\d+): start = (\d+\.\d+), end = (\d+\.\d+)/g;
+  let match;
+  while ((match = segmentRegex.exec(output)) !== null) {
+    if (!match[2] || !match[3]) {
+      throw new Error(`Invalid segment format: ${match[0]}`);
+    }
+    const startTime = parseFloat(match[2]);
+    const endTime = parseFloat(match[3]);
+    segments.push({
+      startTime,
+      endTime,
+      isSpeech: true
+    });
+  }
+  return segments;
+}
+function segmentsToTimeline(segments) {
+  return segments.map((seg) => ({
+    type: "segment",
+    text: seg.isSpeech ? "speech" : "silence",
+    startTime: seg.startTime,
+    endTime: seg.endTime
+  }));
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  detectVoiceActivity,
+  ensureVadInstalled,
+  segmentsToTimeline
+});

package/dist/vad/Silero.d.cts

@@ -0,0 +1,25 @@
+import { Timeline } from '../utilities/Timeline.cjs';
+
+interface SileroOptions {
+    modelDir?: string;
+    installDir?: string;
+    printOutput?: boolean;
+    autoInstall?: boolean;
+    threshold?: number;
+    minSpeechDurationMs?: number;
+    minSilenceDurationMs?: number;
+    speechPadMs?: number;
+}
+interface VadSegment {
+    startTime: number;
+    endTime: number;
+    isSpeech: boolean;
+}
+declare function ensureVadInstalled(options?: SileroOptions): Promise<{
+    installDir: string;
+    vadModelPath: string;
+}>;
+declare function detectVoiceActivity(inputPath: string, options?: SileroOptions): Promise<VadSegment[]>;
+declare function segmentsToTimeline(segments: VadSegment[]): Timeline;
+
+export { type SileroOptions, type VadSegment, detectVoiceActivity, ensureVadInstalled, segmentsToTimeline };

package/dist/vad/Silero.d.ts

@@ -0,0 +1,25 @@
+import { Timeline } from '../utilities/Timeline.js';
+
+interface SileroOptions {
+    modelDir?: string;
+    installDir?: string;
+    printOutput?: boolean;
+    autoInstall?: boolean;
+    threshold?: number;
+    minSpeechDurationMs?: number;
+    minSilenceDurationMs?: number;
+    speechPadMs?: number;
+}
+interface VadSegment {
+    startTime: number;
+    endTime: number;
+    isSpeech: boolean;
+}
+declare function ensureVadInstalled(options?: SileroOptions): Promise<{
+    installDir: string;
+    vadModelPath: string;
+}>;
+declare function detectVoiceActivity(inputPath: string, options?: SileroOptions): Promise<VadSegment[]>;
+declare function segmentsToTimeline(segments: VadSegment[]): Timeline;
+
+export { type SileroOptions, type VadSegment, detectVoiceActivity, ensureVadInstalled, segmentsToTimeline };

package/dist/vad/Silero.js

@@ -0,0 +1,153 @@
+import { spawn } from "node:child_process";
+import { existsSync } from "node:fs";
+import { ensureDir } from "fs-extra";
+import {
+  getInstallDir,
+  getModelDir,
+  getVadExecutablePath,
+  getVadModelPath
+} from "../cli/config.js";
+import { installBinary, installVadModel } from "../cli/install.js";
+const defaultOptions = {
+  printOutput: false,
+  autoInstall: true,
+  threshold: 0.5,
+  minSpeechDurationMs: 250,
+  minSilenceDurationMs: 100,
+  speechPadMs: 30
+};
+async function ensureVadInstalled(options = {}) {
+  const opts = { ...defaultOptions, ...options };
+  const modelDir = opts.modelDir ?? getModelDir();
+  const installDir = opts.installDir ?? getInstallDir();
+  await ensureDir(modelDir);
+  if (opts.autoInstall) {
+    await installBinary({ printOutput: opts.printOutput });
+    await installVadModel({ modelDir, printOutput: opts.printOutput });
+  }
+  const vadModelPath = getVadModelPath(modelDir);
+  if (!existsSync(vadModelPath)) {
+    throw new Error(
+      `VAD model not found at ${vadModelPath}. Run 'ghost-story install --vad' to install.`
+    );
+  }
+  return { installDir, vadModelPath };
+}
+async function detectVoiceActivity(inputPath, options = {}) {
+  const opts = { ...defaultOptions, ...options };
+  if (!existsSync(inputPath)) {
+    throw new Error(`Input file does not exist: ${inputPath}`);
+  }
+  const { installDir, vadModelPath } = await ensureVadInstalled(opts);
+  const executablePath = getVadExecutablePath(installDir);
+  if (!existsSync(executablePath)) {
+    throw new Error(
+      `VAD executable not found at ${executablePath}. Run 'storywhisper install binary' to install.`
+    );
+  }
+  const pathToInput = inputPath;
+  const args = [
+    "-f",
+    pathToInput,
+    "-vm",
+    vadModelPath,
+    "-vt",
+    String(opts.threshold),
+    // vspd is not supported for some reason
+    // "-vspd",
+    // String(opts.minSpeechDurationMs),
+    "-vsd",
+    String(opts.minSilenceDurationMs),
+    "-vp",
+    String(opts.speechPadMs)
+  ];
+  if (opts.printOutput) {
+    console.log(`Running VAD: ${executablePath} ${args.join(" ")}`);
+  }
+  const segments = await runVadProcess(
+    executablePath,
+    args,
+    installDir,
+    opts.printOutput
+  );
+  return segments;
+}
+async function runVadProcess(executable, args, cwd, printOutput) {
+  if (printOutput) {
+    console.log(`Running VAD: ${executable} ${args.join(" ")}`);
+  }
+  return new Promise((resolve, reject) => {
+    const proc = spawn(executable, args, { cwd });
+    let stdout = "";
+    let stderr = "";
+    proc.stdout.on("data", (data) => {
+      const str = data.toString("utf-8");
+      stdout += str;
+      if (printOutput) {
+        process.stdout.write(data);
+      }
+    });
+    proc.stderr.on("data", (data) => {
+      const str = data.toString("utf-8");
+      stderr += str;
+      if (printOutput) {
+        process.stderr.write(data);
+      }
+    });
+    proc.on("exit", (code, signal) => {
+      if (printOutput) {
+        console.log(`VAD process exited with code ${code} and signal ${signal}`);
+      }
+      const segments = parseVadOutput(stdout + stderr);
+      if (printOutput) {
+        console.log(segments);
+      }
+      if (code !== 0 && segments.length === 0) {
+        reject(
+          new Error(
+            `VAD process exited with code ${code}: ${stderr || stdout}`
+          )
+        );
+        return;
+      }
+      resolve(segments);
+    });
+    proc.on("error", (err) => {
+      if (printOutput) {
+        console.error(err);
+      }
+      reject(new Error(`Failed to start VAD process: ${err.message}`));
+    });
+  });
+}
+function parseVadOutput(output) {
+  const segments = [];
+  const segmentRegex = /Speech segment (\d+): start = (\d+\.\d+), end = (\d+\.\d+)/g;
+  let match;
+  while ((match = segmentRegex.exec(output)) !== null) {
+    if (!match[2] || !match[3]) {
+      throw new Error(`Invalid segment format: ${match[0]}`);
+    }
+    const startTime = parseFloat(match[2]);
+    const endTime = parseFloat(match[3]);
+    segments.push({
+      startTime,
+      endTime,
+      isSpeech: true
+    });
+  }
+  return segments;
+}
+function segmentsToTimeline(segments) {
+  return segments.map((seg) => ({
+    type: "segment",
+    text: seg.isSpeech ? "speech" : "silence",
+    startTime: seg.startTime,
+    endTime: seg.endTime
+  }));
+}
+export {
+  detectVoiceActivity,
+  ensureVadInstalled,
+  segmentsToTimeline
+};

package/package.json

@@ -0,0 +1,125 @@
+{
+  "name": "@storyteller-platform/ghost-story",
+  "version": "0.0.1",
+  "description": "An easy-to-use speech toolset. Fork of the original echogarden project.",
+  "author": "",
+  "license": "GPL-3.0",
+  "keywords": [
+    "text-to-speech"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/storyteller-platform/storyteller"
+  },
+  "bugs": {
+    "url": "https://github.com/storyteller-platform/storyteller/issues"
+  },
+  "engines": {
+    "node": ">=20"
+  },
+  "os": [
+    "win32",
+    "darwin",
+    "linux"
+  ],
+  "type": "module",
+  "files": [
+    "dist",
+    "README.md",
+    "LICENSE.md"
+  ],
+  "exports": {
+    ".": {
+      "import": {
+        "types": "./dist/index.d.ts",
+        "default": "./dist/index.js"
+      },
+      "require": {
+        "types": "./dist/index.d.cts",
+        "default": "./dist/index.cjs"
+      }
+    },
+    "./server": {
+      "import": {
+        "types": "./dist/cli/whisper-server.d.ts",
+        "default": "./dist/cli/whisper-server.js"
+      },
+      "require": {
+        "types": "./dist/cli/whisper-server.d.cts",
+        "default": "./dist/cli/whisper-server.cjs"
+      }
+    },
+    "./recognition": {
+      "types": "./dist/api/Recognition.d.ts",
+      "default": "./dist/api/Recognition.js"
+    },
+    "./vad": {
+      "types": "./dist/api/VoiceActivityDetection.d.ts",
+      "default": "./dist/api/VoiceActivityDetection.js"
+    }
+  },
+  "scripts": {
+    "build": "tsup",
+    "prepack": "yarn build",
+    "test": "tsx -C @storyteller --test test/whisper-server.test.ts &&  tsx -C @storyteller --test test/whispercpp.test.ts &&  tsx -C @storyteller --test test/vad.test.ts"
+  },
+  "bin": "./dist/cli/bin.js",
+  "dependencies": {
+    "@aws-sdk/client-transcribe-streaming": "~3.817.0",
+    "@robingenz/zli": "^0.2.0",
+    "chalk": "^5.4.1",
+    "ffmpeg-stream": "^1.0.1",
+    "fs-extra": "~11.3.0",
+    "lodash.merge": "^4.6.2",
+    "microsoft-cognitiveservices-speech-sdk": "~1.44.0",
+    "openai": "~4.103.0",
+    "tar": "^7.4.3",
+    "zod": "^3.25.0"
+  },
+  "devDependencies": {
+    "@storyteller-platform/eslint": "0.1.0",
+    "@storyteller-platform/tsup": "^0.1.0",
+    "@tsconfig/strictest": "^2.0.5",
+    "@types/fs-extra": "~11.0.4",
+    "@types/lodash.merge": "^4",
+    "@types/node": "^24.0.0",
+    "@types/tar": "^6.1.13",
+    "eslint": "^8.0.0",
+    "tsup": "^8.5.0",
+    "tsx": "^4.19.2",
+    "typescript": "~5.8.3"
+  },
+  "publishConfig": {
+    "access": "public",
+    "exports": {
+      ".": {
+        "import": {
+          "types": "./dist/index.d.ts",
+          "default": "./dist/index.js"
+        },
+        "require": {
+          "types": "./dist/index.d.cts",
+          "default": "./dist/index.cjs"
+        }
+      },
+      "./server": {
+        "import": {
+          "types": "./dist/cli/whisper-server.d.ts",
+          "default": "./dist/cli/whisper-server.js"
+        },
+        "require": {
+          "types": "./dist/cli/whisper-server.d.cts",
+          "default": "./dist/cli/whisper-server.cjs"
+        }
+      },
+      "./recognition": {
+        "types": "./dist/api/Recognition.d.ts",
+        "default": "./dist/api/Recognition.js"
+      },
+      "./vad": {
+        "types": "./dist/api/VoiceActivityDetection.d.ts",
+        "default": "./dist/api/VoiceActivityDetection.js"
+      }
+    }
+  }
+}