react-native-davoice-tts 1.0.410 → 1.0.411

package/ios/TTSRNBridge/DavoiceTTS.xcframework/ios-arm64_x86_64-simulator/DavoiceTTS.framework/Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.private.swiftinterface

@@ -161,6 +161,10 @@ public enum AudioPlaybackHook {
   }
   @objc final public func getAllVoices() -> [Swift.String]
   @objc final public func getVoiceAbilities(_ name: Swift.String) -> [Swift.String]
+  #if compiler(>=5.3) && $NonescapableTypes
+  @objc(setExternalVoicesPath:completion:) final public func setExternalVoicesPath(_ path: Swift.String?, completion: (([Swift.String]) -> Swift.Void)? = nil)
+  #endif
+  @objc public static func createVoiceClone(_ inputAudioPath: Swift.String, outputPath: Swift.String, embedderPath: Swift.String) throws -> [Swift.String : Any]
   @objc(changeVoice:completion:) final public func changeVoice(_ name: Swift.String, completion: @escaping (Swift.Int) -> Swift.Void)
   @objc final public var onLastUtteranceFinished: (() -> Swift.Void)?
   @objc deinit

package/ios/TTSRNBridge/DavoiceTTS.xcframework/ios-arm64_x86_64-simulator/DavoiceTTS.framework/Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftinterface

@@ -161,6 +161,10 @@ public enum AudioPlaybackHook {
   }
   @objc final public func getAllVoices() -> [Swift.String]
   @objc final public func getVoiceAbilities(_ name: Swift.String) -> [Swift.String]
+  #if compiler(>=5.3) && $NonescapableTypes
+  @objc(setExternalVoicesPath:completion:) final public func setExternalVoicesPath(_ path: Swift.String?, completion: (([Swift.String]) -> Swift.Void)? = nil)
+  #endif
+  @objc public static func createVoiceClone(_ inputAudioPath: Swift.String, outputPath: Swift.String, embedderPath: Swift.String) throws -> [Swift.String : Any]
   @objc(changeVoice:completion:) final public func changeVoice(_ name: Swift.String, completion: @escaping (Swift.Int) -> Swift.Void)
   @objc final public var onLastUtteranceFinished: (() -> Swift.Void)?
   @objc deinit

package/ios/TTSRNBridge/DavoiceTTS.xcframework/ios-arm64_x86_64-simulator/DavoiceTTS.framework/_CodeSignature/CodeResources

@@ -6,7 +6,7 @@
 	<dict>
 		<key>Headers/DavoiceTTS-Swift.h</key>
 		<data>
-		aXTVWBRj3rHjoLELB1SKEpSKlVI=
+		sBeBhBzj/r1GrbXqcLsAirfQIKQ=
 		</data>
 		<key>Info.plist</key>
 		<data>
@@ -14,43 +14,43 @@
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.abi.json</key>
 		<data>
-		GWj4rqzBJFHhghEkrYmW4vHZBTQ=
+		NZOQSk+oyJ0wjXl83FP/8CfQY9Q=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.private.swiftinterface</key>
 		<data>
-		ksSF6xaCJF6G3MwPOHiS/nveH7o=
+		+SjLLdC9wzcmmVKdK3OspIIfl4c=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftdoc</key>
 		<data>
-		+fcSy/bltR5s4BKnsFh0ZW8LoDg=
+		5Ssm9whRQvwOrYUSWUeqAmJlBCk=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftinterface</key>
 		<data>
-		ksSF6xaCJF6G3MwPOHiS/nveH7o=
+		+SjLLdC9wzcmmVKdK3OspIIfl4c=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftmodule</key>
 		<data>
-		mH9jIRkv1+Mb2wr01gTzH/Lq4KI=
+		IUrwKe5bYNqrgkU5fQDMrzwCF2A=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.abi.json</key>
 		<data>
-		GWj4rqzBJFHhghEkrYmW4vHZBTQ=
+		NZOQSk+oyJ0wjXl83FP/8CfQY9Q=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.private.swiftinterface</key>
 		<data>
-		ZT1wxKJLUV/5yDXgkr1hiFY6Kkw=
+		CCVbMysPRrLBmlDQ1ev8miGvbX4=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftdoc</key>
 		<data>
-		fI/NjOu5fRs+DuF5Nsf2eQwCFZg=
+		mGZKyR1CWLd+ITTHoFSKKVlTcxQ=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftinterface</key>
 		<data>
-		ZT1wxKJLUV/5yDXgkr1hiFY6Kkw=
+		CCVbMysPRrLBmlDQ1ev8miGvbX4=
 		</data>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftmodule</key>
 		<data>
-		WKflcCZyLnNxAl4KY0kr/rkp3Yw=
+		cA4vZqE2K7PeK7Oc4Josd+Vt/Ls=
 		</data>
 		<key>Modules/module.modulemap</key>
 		<data>
@@ -63,121 +63,121 @@
 		<dict>
 			<key>hash</key>
 			<data>
-			aXTVWBRj3rHjoLELB1SKEpSKlVI=
+			sBeBhBzj/r1GrbXqcLsAirfQIKQ=
 			</data>
 			<key>hash2</key>
 			<data>
-			8AwLIvn9Cc6MkqCoTSwxxt0PdUyRSH33aZgxBH42d2c=
+			WIZa1mRKiVZ+bP5gKEKczEKDF8R9zEhTU3SFieC0fO8=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.abi.json</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			GWj4rqzBJFHhghEkrYmW4vHZBTQ=
+			NZOQSk+oyJ0wjXl83FP/8CfQY9Q=
 			</data>
 			<key>hash2</key>
 			<data>
-			IUrltZXg4lQ05e1Tp7k/jFjfoqqHkai2OOvbmlkaAs4=
+			dImQrwOEdDV/E4EJqL/m2QqEGw0dPMKHybTSk2TKAkQ=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.private.swiftinterface</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			ksSF6xaCJF6G3MwPOHiS/nveH7o=
+			+SjLLdC9wzcmmVKdK3OspIIfl4c=
 			</data>
 			<key>hash2</key>
 			<data>
-			SCUrsTaAxwGLnaArK6cNSD5hYMgb029jIt3i0+iOelg=
+			e9/3r3eXLtlmcPkmnRU0xYdk/hzgUy0Rpv8GzYPXK2o=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftdoc</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			+fcSy/bltR5s4BKnsFh0ZW8LoDg=
+			5Ssm9whRQvwOrYUSWUeqAmJlBCk=
 			</data>
 			<key>hash2</key>
 			<data>
-			HMDb7i6UKre/ujb7siIgBLS8+Fz0defkXEApVC4vLIM=
+			dtLnnJQR8LoDEVnD1tg06i/kKRwcaXwxJ/jHdnDmhEI=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftinterface</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			ksSF6xaCJF6G3MwPOHiS/nveH7o=
+			+SjLLdC9wzcmmVKdK3OspIIfl4c=
 			</data>
 			<key>hash2</key>
 			<data>
-			SCUrsTaAxwGLnaArK6cNSD5hYMgb029jIt3i0+iOelg=
+			e9/3r3eXLtlmcPkmnRU0xYdk/hzgUy0Rpv8GzYPXK2o=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/arm64-apple-ios-simulator.swiftmodule</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			mH9jIRkv1+Mb2wr01gTzH/Lq4KI=
+			IUrwKe5bYNqrgkU5fQDMrzwCF2A=
 			</data>
 			<key>hash2</key>
 			<data>
-			n3EI7Sk02U7mWJOkNn1NaMtD4zbkEEFIN9TXL19EBYM=
+			S/nm0VmvBX9ZvqiP8S6Y9LBt62/or63J0jrUsMBV5X0=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.abi.json</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			GWj4rqzBJFHhghEkrYmW4vHZBTQ=
+			NZOQSk+oyJ0wjXl83FP/8CfQY9Q=
 			</data>
 			<key>hash2</key>
 			<data>
-			IUrltZXg4lQ05e1Tp7k/jFjfoqqHkai2OOvbmlkaAs4=
+			dImQrwOEdDV/E4EJqL/m2QqEGw0dPMKHybTSk2TKAkQ=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.private.swiftinterface</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			ZT1wxKJLUV/5yDXgkr1hiFY6Kkw=
+			CCVbMysPRrLBmlDQ1ev8miGvbX4=
 			</data>
 			<key>hash2</key>
 			<data>
-			nywtFQgfcnZrO0U914JYYA2M7Do7Fpqy+mFpzteM43w=
+			zs5TSJY3M0yp73wG3kcfebXGiEGjaBbjDvBJN1eDcEg=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftdoc</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			fI/NjOu5fRs+DuF5Nsf2eQwCFZg=
+			mGZKyR1CWLd+ITTHoFSKKVlTcxQ=
 			</data>
 			<key>hash2</key>
 			<data>
-			kedZaxsmXt+0P9wn0DUIyKHC6sUL1OmlaOtGVaL3k3k=
+			Iqp+5eTycIlB+l9STROeFAxcZ/W9bSUDLqAMYDkMbCw=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftinterface</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			ZT1wxKJLUV/5yDXgkr1hiFY6Kkw=
+			CCVbMysPRrLBmlDQ1ev8miGvbX4=
 			</data>
 			<key>hash2</key>
 			<data>
-			nywtFQgfcnZrO0U914JYYA2M7Do7Fpqy+mFpzteM43w=
+			zs5TSJY3M0yp73wG3kcfebXGiEGjaBbjDvBJN1eDcEg=
 			</data>
 		</dict>
 		<key>Modules/DavoiceTTS.swiftmodule/x86_64-apple-ios-simulator.swiftmodule</key>
 		<dict>
 			<key>hash</key>
 			<data>
-			WKflcCZyLnNxAl4KY0kr/rkp3Yw=
+			cA4vZqE2K7PeK7Oc4Josd+Vt/Ls=
 			</data>
 			<key>hash2</key>
 			<data>
-			89+0RSxFC9KOOhvMzQN3ina8oFsBJ1uob3XKWjnaNfc=
+			QAjZ2CvyRKa72Kv5Ki3OOS/Q6EuERdLQwHK9/M4yO6Y=
 			</data>
 		</dict>
 		<key>Modules/module.modulemap</key>

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "react-native-davoice-tts",
-    "version": "1.0.410",
+    "version": "1.0.411",
     "description": "tts library for React Native",
     "main": "tts/index.js",
     "types": "tts/index.d.ts",

package/speech/index.d.ts

@@ -37,6 +37,18 @@ export type NewSpeechWAVEvent = {
 };
 export type RNAssetLike = any;
 export type ModelRef = string | RNAssetLike;
+/** createVoiceClone() result. */
+export type VoiceCloneResult = {
+    /** The .emd written (its file name without .emd is the voice name). */
+    path: string;
+    voice: string;
+    /** Length of the reference clip that was embedded. */
+    seconds: number;
+    /** Non-empty when less than 4 s of speech remained or the cut fell inside speech. */
+    warnings: string[];
+    /** What the preparation did (silence trimmed, pauses shortened, cut point). */
+    report: string[];
+};
 /** iOS AVAudioSession routing entry shared by STT and TTS. */
 export type PreferredAudioInput = 'bluetoothHighQualityMic' | 'builtInMic' | 'none';
 export type RouteConfigEntry = {
@@ -155,6 +167,32 @@ declare class Speech {
      * Resolves [] for voices without variants or when unavailable.
      */
     getVoiceAbilities(name: string): Promise<string[]>;
+    /** Also look for voices in the folder `path`: <name>.emd files made by
+     * createVoiceClone() (or .npy). They are listed by getAllVoices() after
+     * the bundled voices and selected with changeVoice() like them; a bundled
+     * voice wins over a file with the same name. Call again to pick up new files;
+     * null or "" removes them. Resolves the external voice names found.
+     */
+    setExternalVoicesPath(path: string | null): Promise<string[]>;
+    /** Creates a voice clone from reference audio and saves it as a
+     * scrambled .emd at `outputPath`.
+     *
+     * - inputAudio: path/URI of a wav, mp3, m4a/aac, ogg, flac, ... recording. The
+     *   best ~6 s of speech is used: silence trimmed, long pauses shortened, cut at
+     *   a gap between words (never inside a word).
+     * - outputPath: the .emd file to write (".emd" is added when missing), or a
+     *   folder ending in "/" (-> <folder>/<input name>.emd). The file name is the
+     *   voice name.
+     * - embedder: the language's voice encoder embedder_<code>.dm, e.g.
+     *   require('./assets/models/embedder_en.dm') or a path; it is unscrambled in
+     *   memory only.
+     *
+     * Resolves { path, voice, seconds, warnings, report }; warnings is non-empty
+     * when less than 4 s of speech remained or the cut had to fall inside speech.
+     * When the .emd lands in the folder given to setExternalVoicesPath(), the new
+     * voice is usable right away.
+     */
+    createVoiceClone(inputAudio: ModelRef, outputPath: string, embedder: ModelRef): Promise<VoiceCloneResult>;
     /** Reference voice, matched case-insensitively without .wav.
      * Returns 0 on match; -1 on fallback/unavailability. Missing names select the
      * first WAV (alphabetical order). Already generated audio keeps its voice.
@@ -176,6 +214,7 @@ declare class Speech {
     private hasCompletedFullInit;
     private hasCompletedTtsInit;
     private iosTtsOnly;
+    private externalVoicesPath;
     private logCall;
     private resolveModelExt;
     private resolveModelToPath;

package/speech/index.ts

@@ -137,6 +137,19 @@ export type NewSpeechWAVEvent = { path: string };
 export type RNAssetLike = any; // keep permissive (Metro asset numbers/objects vary)
 export type ModelRef = string | RNAssetLike;
 
+/** createVoiceClone() result. */
+export type VoiceCloneResult = {
+  /** The .emd written (its file name without .emd is the voice name). */
+  path: string;
+  voice: string;
+  /** Length of the reference clip that was embedded. */
+  seconds: number;
+  /** Non-empty when less than 4 s of speech remained or the cut fell inside speech. */
+  warnings: string[];
+  /** What the preparation did (silence trimmed, pauses shortened, cut point). */
+  report: string[];
+};
+
 /** iOS AVAudioSession routing entry shared by STT and TTS. */
 export type PreferredAudioInput =
   | 'bluetoothHighQualityMic'
@@ -337,6 +350,65 @@ class Speech {
     }
   }
 
+  /** Also look for voices in the folder `path`: <name>.emd files made by
+   * createVoiceClone() (or .npy). They are listed by getAllVoices() after
+   * the bundled voices and selected with changeVoice() like them; a bundled
+   * voice wins over a file with the same name. Call again to pick up new files;
+   * null or "" removes them. Resolves the external voice names found.
+   */
+  async setExternalVoicesPath(path: string | null): Promise<string[]> {
+    const bridge = Platform.OS === 'ios' && !this.iosTtsOnly && this.hasCompletedFullInit
+      ? NativeSpeech : NativeTTS;
+    if ((Platform.OS !== 'android' && Platform.OS !== 'ios') || typeof bridge?.setExternalVoicesPath !== 'function') return [];
+    this.externalVoicesPath = path || null;
+    try {
+      const names = await bridge.setExternalVoicesPath(path || null);
+      return Array.isArray(names) ? names.filter((v: unknown): v is string => typeof v === 'string') : [];
+    } catch {
+      return [];
+    }
+  }
+
+  /** Creates a voice clone from reference audio and saves it as a
+   * scrambled .emd at `outputPath`.
+   *
+   * - inputAudio: path/URI of a wav, mp3, m4a/aac, ogg, flac, ... recording. The
+   *   best ~6 s of speech is used: silence trimmed, long pauses shortened, cut at
+   *   a gap between words (never inside a word).
+   * - outputPath: the .emd file to write (".emd" is added when missing), or a
+   *   folder ending in "/" (-> <folder>/<input name>.emd). The file name is the
+   *   voice name.
+   * - embedder: the language's voice encoder embedder_<code>.dm, e.g.
+   *   require('./assets/models/embedder_en.dm') or a path; it is unscrambled in
+   *   memory only.
+   *
+   * Resolves { path, voice, seconds, warnings, report }; warnings is non-empty
+   * when less than 4 s of speech remained or the cut had to fall inside speech.
+   * When the .emd lands in the folder given to setExternalVoicesPath(), the new
+   * voice is usable right away.
+   */
+  async createVoiceClone(inputAudio: ModelRef, outputPath: string, embedder: ModelRef): Promise<VoiceCloneResult> {
+    const bridge = (typeof NativeTTS?.createVoiceClone === 'function') ? NativeTTS : NativeSpeech;
+    if (typeof bridge?.createVoiceClone !== 'function') {
+      throw new Error('createVoiceClone is not available on this platform');
+    }
+    const result = await bridge.createVoiceClone(
+      this.resolveModelToPath(inputAudio), String(outputPath), this.resolveModelToPath(embedder));
+    const folder = this.externalVoicesPath;
+    if (folder && typeof result?.path === 'string') {
+      const strip = (p: string) => p.replace(/^file:\/\//, '').replace(/\/+$/, '');
+      const parent = result.path.substring(0, result.path.lastIndexOf('/'));
+      if (strip(parent) === strip(folder)) await this.setExternalVoicesPath(folder);
+    }
+    return {
+      path: String(result?.path ?? ''),
+      voice: String(result?.voice ?? ''),
+      seconds: Number(result?.seconds ?? 0),
+      warnings: Array.isArray(result?.warnings) ? result.warnings.map(String) : [],
+      report: Array.isArray(result?.report) ? result.report.map(String) : [],
+    };
+  }
+
   /** Reference voice, matched case-insensitively without .wav.
    * Returns 0 on match; -1 on fallback/unavailability. Missing names select the
    * first WAV (alphabetical order). Already generated audio keeps its voice.
@@ -423,6 +495,7 @@ class Speech {
   private hasCompletedFullInit = false;
   private hasCompletedTtsInit = false;
   private iosTtsOnly = false; // when true, use NativeTTS directly on iOS
+  private externalVoicesPath: string | null = null;  // setExternalVoicesPath()
 
 
   private logCall(name: string, payload?: any) {