Migrate Yandex SpeechKit from API v1 to v3 (REST)

build.gradle.kts

@@ -211,6 +211,7 @@ tasks.withType<JacocoReport> {
                             "com/epam/brn/exception/**",
                             "com/epam/brn/Application*",
                             "com/epam/brn/service/azure/tts/config/**",
+                            "com/epam/brn/service/yandex/tts/config/**",
                             "com/epam/brn/webclient/customizer/**",
                             "com/epam/brn/webclient/model/**",
                         )
@@ -251,6 +252,7 @@ sonarqube {
                 "**/com/epam/brn/service/load/FirebaseUserDataLoader*," +
                 "**/com/epam/brn/service/azure/tts/AzureVoiceLoader*," +
                 "**/com/epam/brn/service/azure/tts/config/**," +
+                "**/com/epam/brn/service/yandex/tts/config/**," +
                 "**/com/epam/brn/webclient/customizer/**," +
                 "**/com/epam/brn/webclient/model/**",
         )

frontend/make-words.js

@@ -1,8 +1,7 @@
 /* eslint-disable @typescript-eslint/no-var-requires */
 /* global require */
 const fs = require('fs');
-const request = require('request');
-const qs = require('querystring');
+const https = require('https');
 
 const words = `бам,сам,дам,зал,бум`;
 const token = '';
@@ -12,31 +11,77 @@ const folderId = '';
 // install ffmpeg
 
 // yc iam create-token
-// https://cloud.yandex.ru/docs/speechkit/tts/request
+// https://cloud.yandex.ru/docs/speechkit/tts/v3/api-ref/grpc/
 
-const yandex_tts_url =
-  'https://tts.api.cloud.yandex.net/speech/v1/tts:synthesize?';
+const yandex_tts_url = '/tts/v3/utteranceSynthesis';
 
 function YandexTTS(options, callback) {
-  var params = {};
-
-  params['text'] = options['text'];
-  params['folderId'] = folderId;
-  params['format'] = 'oggopus';
-  params['lang'] = 'ru-RU';
-  params['voice'] = 'filipp';
-  params['emotion'] = 'good';
-
-  var full_url = yandex_tts_url + qs.stringify(params);
+  const body = JSON.stringify({
+    text: options['text'],
+    outputAudioSpec: {
+      containerAudio: {
+        containerAudioType: 'OGG_OPUS',
+      },
+    },
+    hints: [
+      { voice: 'filipp' },
+      { role: 'neutral' },
+    ],
+    loudnessNormalizationType: 'LUFS',
+  });
 
-  var file = fs.createWriteStream(options['file']);
-  file.on('finish', callback);
-  request({
-    url: full_url,
+  const reqOptions = {
+    hostname: 'tts.api.cloud.yandex.net',
+    port: 443,
+    path: yandex_tts_url,
+    method: 'POST',
     headers: {
-      Authorization: `Bearer ${token}`,
+      'Authorization': `Bearer ${token}`,
+      'x-folder-id': folderId,
+      'Content-Type': 'application/json',
+      'Content-Length': Buffer.byteLength(body),
     },
-  }).pipe(file);
+  };
+
+  const file = fs.createWriteStream(options['file']);
+  const req = https.request(reqOptions, (res) => {
+    let responseData = '';
+    res.on('data', (chunk) => {
+      responseData += chunk;
+    });
+    res.on('end', () => {
+      if (res.statusCode !== 200) {
+        console.error(`HTTP ${res.statusCode}: ${responseData}`);
+        file.end();
+        callback();
+        return;
+      }
+
+      const lines = responseData.split('\n').filter((line) => line.trim());
+      for (const line of lines) {
+        try {
+          const parsed = JSON.parse(line);
+          if (parsed.result && parsed.result.audioChunk && parsed.result.audioChunk.data) {
+            const audioBuffer = Buffer.from(parsed.result.audioChunk.data, 'base64');
+            file.write(audioBuffer);
+          }
+        } catch (e) {
+          // Ignore non-JSON transport lines.
+        }
+      }
+
+      file.end(callback);
+    });
+  });
+
+  req.on('error', (e) => {
+    console.error(`Request error: ${e.message}`);
+    file.end();
+    callback();
+  });
+
+  req.write(body);
+  req.end();
 }
 
 const execSync = require('child_process').execSync;
@@ -72,8 +117,3 @@ async function makeFiles() {
 }
 
 makeFiles();
-// stack.forEach((word)=>{
-//     let file = word.trim();
-//     execSync(`gtts-cli "${word}." -lang_check --lang ru --output ${file}.mp3`);
-//     execSync(`ffmpeg-normalize ${file}.mp3 --normalization-type peak --target-level 0 -c:a libmp3lame -b:a 320k -o ${file}_n.mp3`)
-// });

src/main/kotlin/com/epam/brn/controller/YandexAudioSettingsController.kt

@@ -0,0 +1,66 @@
+package com.epam.brn.controller
+
+import com.epam.brn.dto.request.audio.AudioVoiceOverrideRequest
+import com.epam.brn.dto.response.BrnResponse
+import com.epam.brn.dto.response.audio.AudioVoiceOptionResponse
+import com.epam.brn.dto.response.audio.AudioVoiceSettingsResponse
+import com.epam.brn.enums.BrnRole
+import com.epam.brn.service.WordsService
+import io.swagger.v3.oas.annotations.Operation
+import io.swagger.v3.oas.annotations.tags.Tag
+import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty
+import org.springframework.http.ResponseEntity
+import org.springframework.web.bind.annotation.GetMapping
+import org.springframework.web.bind.annotation.PostMapping
+import org.springframework.web.bind.annotation.RequestBody
+import org.springframework.web.bind.annotation.RequestMapping
+import org.springframework.web.bind.annotation.RequestParam
+import org.springframework.web.bind.annotation.RestController
+import javax.annotation.security.RolesAllowed
+
+@RestController
+@RequestMapping("/audio")
+@Tag(name = "Audio", description = "Contains actions for getting audio file for words")
+@ConditionalOnProperty(name = ["default.tts.provider"], havingValue = "yandex")
+@RolesAllowed(BrnRole.USER)
+class YandexAudioSettingsController(
+    private val wordsService: WordsService,
+) {
+    @GetMapping("/voices")
+    @Operation(summary = "Get available Yandex voices and the current runtime default for a locale")
+    fun getVoices(
+        @RequestParam(required = false, defaultValue = "ru-ru") locale: String,
+    ): ResponseEntity<BrnResponse<AudioVoiceSettingsResponse>> = ResponseEntity
+        .ok()
+        .body(BrnResponse(data = buildVoiceSettingsResponse(locale)))
+
+    @PostMapping("/default-voice")
+    @Operation(summary = "Set the runtime default Yandex voice for a locale until the server restarts")
+    @RolesAllowed(BrnRole.ADMIN)
+    fun setDefaultVoice(
+        @RequestBody request: AudioVoiceOverrideRequest,
+    ): ResponseEntity<BrnResponse<AudioVoiceSettingsResponse>> {
+        wordsService.setDefaultVoiceForLocale(request.locale, request.voice)
+        return ResponseEntity.ok().body(BrnResponse(data = buildVoiceSettingsResponse(request.locale)))
+    }
+
+    private fun buildVoiceSettingsResponse(locale: String): AudioVoiceSettingsResponse {
+        val defaultVoice = wordsService.getDefaultVoiceForLocale(locale)
+        val voiceOptions =
+            wordsService.getAvailableVoicesForLocale(locale).map { voice ->
+                AudioVoiceOptionResponse(
+                    name = voice.name,
+                    apiValue = voice.apiValue,
+                    gender = voice.gender.name.lowercase(),
+                    roles = voice.supportedRoles.map { it.apiValue },
+                    isDefault = voice.name == defaultVoice,
+                )
+            }
+
+        return AudioVoiceSettingsResponse(
+            locale = locale.lowercase(),
+            defaultVoice = defaultVoice,
+            voices = voiceOptions,
+        )
+    }
+}

src/main/kotlin/com/epam/brn/dto/request/audio/AudioVoiceOverrideRequest.kt

@@ -0,0 +1,6 @@
+package com.epam.brn.dto.request.audio
+
+data class AudioVoiceOverrideRequest(
+    val locale: String,
+    val voice: String,
+)

src/main/kotlin/com/epam/brn/dto/response/audio/AudioVoiceSettingsResponse.kt

@@ -0,0 +1,15 @@
+package com.epam.brn.dto.response.audio
+
+data class AudioVoiceSettingsResponse(
+    val locale: String,
+    val defaultVoice: String,
+    val voices: List<AudioVoiceOptionResponse>,
+)
+
+data class AudioVoiceOptionResponse(
+    val name: String,
+    val apiValue: String,
+    val gender: String,
+    val roles: List<String>,
+    val isDefault: Boolean,
+)

src/main/kotlin/com/epam/brn/dto/yandex/tts/YandexTtsRequest.kt

@@ -0,0 +1,25 @@
+package com.epam.brn.dto.yandex.tts
+
+import com.fasterxml.jackson.annotation.JsonInclude
+
+data class YandexTtsRequest(
+    val text: String,
+    val outputAudioSpec: OutputAudioSpec,
+    val hints: List<Hint>,
+    val loudnessNormalizationType: String = "LUFS",
+)
+
+data class OutputAudioSpec(
+    val containerAudio: ContainerAudio,
+)
+
+data class ContainerAudio(
+    val containerAudioType: String = "OGG_OPUS",
+)
+
+@JsonInclude(JsonInclude.Include.NON_NULL)
+data class Hint(
+    val voice: String? = null,
+    val speed: String? = null,
+    val role: String? = null,
+)

src/main/kotlin/com/epam/brn/dto/yandex/tts/YandexTtsResponse.kt

@@ -0,0 +1,18 @@
+package com.epam.brn.dto.yandex.tts
+
+import com.fasterxml.jackson.annotation.JsonIgnoreProperties
+
+@JsonIgnoreProperties(ignoreUnknown = true)
+data class YandexTtsResponse(
+    val result: YandexTtsResult? = null,
+)
+
+@JsonIgnoreProperties(ignoreUnknown = true)
+data class YandexTtsResult(
+    val audioChunk: AudioChunk? = null,
+)
+
+@JsonIgnoreProperties(ignoreUnknown = true)
+data class AudioChunk(
+    val data: String? = null,
+)

src/main/kotlin/com/epam/brn/enums/BrnLocale.kt

@@ -5,5 +5,4 @@ enum class BrnLocale(
 ) {
     RU("ru-ru"),
     EN("en-us"),
-    TR("tr-tr"),
 }

src/main/kotlin/com/epam/brn/enums/Voice.kt

@@ -1,20 +1,61 @@
 package com.epam.brn.enums
 
-// from docs: https://cloud.yandex.ru/docs/speechkit/tts/voices
-enum class Voice {
-    FILIPP, // old Russian man voice
-    ALEXANDER, // new Russian man voice: neutral, good
-    KIRILL, // new Russian man voice: neutral, strict, good
-
-    OKSANA, // old Russian woman voice
-    MARINA, // new Russian woman voice
-    LERA, // new Russian woman voice
-    DASHA, // new Russian woman voice
-
-    NICK, // old English man voice
-    JOHN, // new English man voice
-    ALYSS, // old English woman voice
-
-    ERKANYAVAS, // old Turkish man voice
-    SILAERKAN, // old Turkish voice
+// Based on the current Yandex SpeechKit TTS voices docs.
+enum class Voice(
+    val locale: String,
+    val gender: VoiceGender,
+    vararg supportedRoles: VoiceRole,
+) {
+    FILIPP(BrnLocale.RU.locale, VoiceGender.MALE, VoiceRole.NEUTRAL),
+    ERMIL(BrnLocale.RU.locale, VoiceGender.MALE, VoiceRole.NEUTRAL, VoiceRole.GOOD),
+    ZAHAR(BrnLocale.RU.locale, VoiceGender.MALE),
+    ALEXANDER(BrnLocale.RU.locale, VoiceGender.MALE, VoiceRole.NEUTRAL, VoiceRole.GOOD),
+    KIRILL(BrnLocale.RU.locale, VoiceGender.MALE, VoiceRole.NEUTRAL, VoiceRole.STRICT, VoiceRole.GOOD),
+
+    ALENA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    OKSANA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    MARINA(BrnLocale.RU.locale, VoiceGender.FEMALE, VoiceRole.FRIENDLY),
+    DASHA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    LERA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    JULIA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    MASHA(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    MADI_RU(BrnLocale.RU.locale, VoiceGender.FEMALE),
+    OMAZH(BrnLocale.RU.locale, VoiceGender.FEMALE),
+
+    JOHN(BrnLocale.EN.locale, VoiceGender.MALE),
+    NICK(BrnLocale.EN.locale, VoiceGender.MALE),
+    JANE(BrnLocale.EN.locale, VoiceGender.FEMALE),
+    ALYSS(BrnLocale.EN.locale, VoiceGender.FEMALE),
+    ;
+
+    val supportedRoles: List<VoiceRole> = supportedRoles.toList()
+
+    val apiValue: String
+        get() = name.lowercase()
+
+    companion object {
+        fun getVoicesForLocale(locale: String): List<Voice> = values().filter { it.locale == locale.lowercase() }
+
+        fun findByValue(value: String): Voice? = values().firstOrNull { it.name.equals(value, ignoreCase = true) }
+    }
+}
+
+enum class VoiceGender {
+    MALE,
+    FEMALE,
+}
+
+enum class VoiceRole {
+    NEUTRAL,
+    GOOD,
+    FRIENDLY,
+    STRICT,
+    ;
+
+    val apiValue: String
+        get() = name.lowercase()
+
+    companion object {
+        fun findByValue(value: String): VoiceRole? = values().firstOrNull { it.name.equals(value, ignoreCase = true) }
+    }
 }