Implement FFmpeg converter and enable mp3 conversion for /sonos

README.md

@@ -41,21 +41,22 @@ old PC-speakers you are able to connect your device to them via line port and
 get a complete speech-based notification system.
 
 
-## The `/wave` endpoint
+## The `/{wave,acc,mp3,m4a,wma,flac}` endpoint
 ```
-POST /wave
+POST /{wave,acc,mp3,m4a,wma,flac}
 {
 "text": "The text to be spoken",
 "language": "en_US"
 }
 ```
-*Returns:* `200 OK` with wave file (`Content-Type: audio/x-wav`)
+*Returns:* `200 OK` with proper audio file
 
-The `/wave` endpoint enables you to download a wav file containing speech of the
+This endpoint enables you to download an audio file file containing speech of the
 provided text. The goal of this endpoint is to provide interface allowing you establishment
 of the connection between the TTS Server and some other kind of already running TTS system,
 which can invoke the HTTP request to your Android device and do something with returned
-wav file.
+audio file. Note that all files but wav originates right from the wav file through FFmpeg
+converter.
 
 ## The `/sonos` endpoint
 ```

app/build.gradle

@@ -37,6 +37,7 @@ dependencies {
     implementation 'com.bartlomiejpluta:sonos-controller:v.0.2'
     implementation 'com.google.dagger:dagger-android:2.15'
     implementation 'com.google.dagger:dagger-android-support:2.15'
+    implementation 'com.github.adrielcafe:AndroidAudioConverter:0.0.8'
     kapt 'com.google.dagger:dagger-android-processor:2.15'
     kapt 'com.google.dagger:dagger-compiler:2.15'
     testImplementation 'junit:junit:4.12'

app/src/main/assets/help/help.html

@@ -58,21 +58,22 @@ POST /say
     get a complete speech-based notification system.
 </p>
 
-<h2>The <code>/wave</code> endpoint</h2>
+<h2>The <code>/{wave,acc,mp3,m4a,wma,flac}</code> endpoint</h2>
 <pre>
-POST /wave
+POST /{wave,acc,mp3,m4a,wma,flac}
 {
     "text": "The text to be spoken",
     "language": "en_US"
 }
 </pre>
-<p><b>Returns:</b> <code>200 OK</code> with wave file (<code>Content-Type: audio/x-wav</code>)</p>
+<p><b>Returns:</b> <code>200 OK</code> with proper audio file</p>
 <p>
-    The <code>/wave</code> endpoint enables you to download a wav file containing speech of the
+    This endpoint enables you to download an audio file file containing speech of the
     provided text. The goal of this endpoint is to provide interface allowing you establishment
     of the connection between the TTS Server and some other kind of already running TTS system,
     which can invoke the HTTP request to your Android device and do something with returned
-    wav file.
+    audio file. Note that all files but wav originates right from the wav file through FFmpeg
+    converter.
 </p>
 
 <h2>The <code>/sonos</code> endpoint</h2>

app/src/main/java/com/bartlomiejpluta/ttsserver/core/sonos/worker/SonosWorker.kt

@@ -1,6 +1,7 @@
 package com.bartlomiejpluta.ttsserver.core.sonos.worker
 
 import android.content.SharedPreferences
+import cafe.adriel.androidaudioconverter.model.AudioFormat
 import com.bartlomiejpluta.ttsserver.core.tts.engine.TTSEngine
 import com.bartlomiejpluta.ttsserver.core.web.dto.SonosDTO
 import com.bartlomiejpluta.ttsserver.core.web.endpoint.Endpoint
@@ -40,7 +41,7 @@ class SonosWorker(
       }
 
    private fun prepareTTSFile(data: SonosDTO): String {
-      val filename = tts.createTTSFile(data.text, data.language).name
+      val filename = tts.createTTSFile(data.text, data.language, AudioFormat.MP3).name
       return "$announcementUrl$filename"
    }
 

app/src/main/java/com/bartlomiejpluta/ttsserver/core/tts/engine/TTSEngine.kt

@@ -6,12 +6,14 @@ import android.media.AudioAttributes
 import android.media.MediaPlayer
 import android.net.Uri
 import android.speech.tts.TextToSpeech
+import cafe.adriel.androidaudioconverter.model.AudioFormat
 import com.bartlomiejpluta.ttsserver.core.tts.exception.TTSException
 import com.bartlomiejpluta.ttsserver.core.tts.listener.GongListener
 import com.bartlomiejpluta.ttsserver.core.tts.listener.TTSProcessListener
 import com.bartlomiejpluta.ttsserver.core.tts.model.TTSStream
 import com.bartlomiejpluta.ttsserver.core.tts.status.TTSStatus
 import com.bartlomiejpluta.ttsserver.core.tts.status.TTSStatusHolder
+import com.bartlomiejpluta.ttsserver.core.util.AudioConverter
 import com.bartlomiejpluta.ttsserver.ui.preference.PreferenceKey
 import java.io.BufferedInputStream
 import java.io.File
@@ -23,38 +25,50 @@ class TTSEngine(
    private val context: Context,
    private val tts: TextToSpeech,
    private val ttsStatusHolder: TTSStatusHolder,
-   private val preferences: SharedPreferences
+   private val preferences: SharedPreferences,
+   private val audioConverter: AudioConverter
 ) {
    private val messageDigest = MessageDigest.getInstance("SHA-256")
 
    val status: TTSStatus
       get() = ttsStatusHolder.status
 
-   fun createTTSFile(text: String, language: Locale): File {
+   fun createTTSFile(text: String, language: Locale, audioFormat: AudioFormat = AudioFormat.WAV): File {
       val digest = hash(text, language)
-      val filename = "tts_$digest.wav"
-      val file = File(context.cacheDir, filename)
+      val targetFilename = "tts_$digest.${audioFormat.format}"
+      val wavFilename = "tts_$digest.wav"
+      val wavFile = File(context.cacheDir, wavFilename)
+      val targetFile = File(context.cacheDir, targetFilename)
 
-      file.takeIf { it.exists() }?.let { return it }
+      targetFile.takeIf { it.exists() }?.let { return it }
 
       val uuid = UUID.randomUUID().toString()
       val listener = TTSProcessListener(uuid)
       tts.setOnUtteranceProgressListener(listener)
 
       tts.language = language
-      tts.synthesizeToFile(text, null, file, uuid)
+      tts.synthesizeToFile(text, null, wavFile, uuid)
       listener.await()
 
+      return convertFile(wavFile, audioFormat)
+   }
+
+   private fun convertFile(file: File, audioFormat: AudioFormat): File {
+      if (audioFormat == AudioFormat.WAV) {
          return file
       }
 
+
+      return audioConverter.convert(file, audioFormat).also { file.delete() }
+   }
+
    private fun hash(text: String, language: Locale): String {
       val bytes = "$text$language".toByteArray()
       val digest = messageDigest.digest(bytes)
       return digest.fold("", { str, it -> str + "%02x".format(it) })
    }
 
-   fun fetchTTSStream(text: String, language: Locale): TTSStream {
+   fun fetchTTSStream(text: String, language: Locale, audioFormat: AudioFormat = AudioFormat.WAV): TTSStream {
       val file = createTempFile("tmp_tts_server", ".wav")
 
       val uuid = UUID.randomUUID().toString()
@@ -65,8 +79,9 @@ class TTSEngine(
       tts.synthesizeToFile(text, null, file, uuid)
       listener.await()
 
-      val stream = BufferedInputStream(FileInputStream(file))
-      val length = file.length()
+      val converted = convertFile(file, audioFormat)
+      val stream = BufferedInputStream(FileInputStream(converted))
+      val length = converted.length()
 
       file.delete()
 

app/src/main/java/com/bartlomiejpluta/ttsserver/core/tts/exception/AudioConversionException.kt

@@ -0,0 +1,6 @@
+package com.bartlomiejpluta.ttsserver.core.tts.exception
+
+class AudioConversionException : Exception {
+   constructor(message: String) : super(message)
+   constructor(message: String, cause: Throwable) : super(message, cause)
+}

app/src/main/java/com/bartlomiejpluta/ttsserver/core/tts/listener/ConverterListener.kt

@@ -0,0 +1,20 @@
+package com.bartlomiejpluta.ttsserver.core.tts.listener
+
+import cafe.adriel.androidaudioconverter.callback.IConvertCallback
+import com.bartlomiejpluta.ttsserver.core.tts.exception.AudioConversionException
+import java.io.File
+import java.util.concurrent.LinkedBlockingQueue
+
+class ConverterListener : IConvertCallback {
+   private val queue = LinkedBlockingQueue<File>()
+
+   fun await() = queue.take()
+
+   override fun onSuccess(convertedFile: File?) {
+      queue.add(convertedFile)
+   }
+
+   override fun onFailure(error: Exception?) {
+      error?.let { throw AudioConversionException("Conversion failed", error) }
+   }
+}

app/src/main/java/com/bartlomiejpluta/ttsserver/core/util/AudioConverter.kt

@@ -0,0 +1,47 @@
+package com.bartlomiejpluta.ttsserver.core.util
+
+import android.content.Context
+import cafe.adriel.androidaudioconverter.AndroidAudioConverter
+import cafe.adriel.androidaudioconverter.callback.ILoadCallback
+import cafe.adriel.androidaudioconverter.model.AudioFormat
+import com.bartlomiejpluta.ttsserver.core.tts.exception.AudioConversionException
+import com.bartlomiejpluta.ttsserver.core.tts.listener.ConverterListener
+import java.io.File
+
+class AudioConverter(private val context: Context) {
+   var state: State = State.UNLOADED
+      private set
+
+   enum class State {
+      READY,
+      ERROR,
+      UNLOADED
+   }
+
+   init {
+      AndroidAudioConverter.load(context, object : ILoadCallback {
+         override fun onSuccess() {
+            state = State.READY
+         }
+
+         override fun onFailure(error: Exception?) {
+            state = State.ERROR
+         }
+      })
+   }
+
+   fun convert(file: File, format: AudioFormat): File {
+      if(state != State.READY) {
+         throw AudioConversionException("Converter is not ready")
+      }
+
+      val listener = ConverterListener()
+      AndroidAudioConverter.with(context)
+         .setFile(file)
+         .setFormat(format)
+         .setCallback(listener)
+         .convert()
+
+      return listener.await()
+   }
+}

app/src/main/java/com/bartlomiejpluta/ttsserver/core/web/endpoint/Endpoint.kt

@@ -4,14 +4,19 @@ enum class Endpoint(val uri: String, val id: Int) {
    UNKNOWN("/", 1),
    SAY("/say", 2),
    WAVE("/wave", 3),
-   SONOS("/sonos", 4),
-   SONOS_CACHE("/sonos/*", 5),
-   GONG("/gong.wav", 6);
+   AAC("/aac", 4),
+   MP3("/mp3", 5),
+   M4A("/m4a", 6),
+   WMA("/wma", 7),
+   FLAC("/flac", 8),
+   SONOS("/sonos", 9),
+   SONOS_CACHE("/sonos/*", 10),
+   GONG("/gong.wav", 11);
 
    val trimmedUri: String
       get() = uri.replace("*", "")
 
    companion object {
-      fun of(id: Int) = values().firstOrNull { it.id == id } ?: UNKNOWN
+      fun of(ordinal: Int) = values().firstOrNull { it.ordinal == ordinal } ?: UNKNOWN
    }
 }

app/src/main/java/com/bartlomiejpluta/ttsserver/core/web/endpoint/EndpointMatcher.kt

@@ -8,7 +8,7 @@ object EndpointMatcher {
 
    init {
       Endpoint.values().forEach {
-         uriMatcher.addURI("", it.uri, it.id)
+         uriMatcher.addURI("", it.uri, it.ordinal)
       }
    }
 

app/src/main/java/com/bartlomiejpluta/ttsserver/core/web/mime/MimeType.kt

@@ -0,0 +1,28 @@
+package com.bartlomiejpluta.ttsserver.core.web.mime
+
+import cafe.adriel.androidaudioconverter.model.AudioFormat
+import java.io.File
+import java.util.*
+
+enum class MimeType(val mimeType: String) {
+   AAC("audio/aac"),
+   MP3("audio/mpeg"),
+   M4A("audio/m4a"),
+   WMA("audio/x-ms-wma"),
+   WAV("audio/x-wav"),
+   FLAC("audio/x-wav"),
+   JSON("application/json");
+
+   companion object {
+      fun forAudioFormat(audioFormat: AudioFormat) = when(audioFormat) {
+         AudioFormat.AAC -> AAC
+         AudioFormat.MP3 -> MP3
+         AudioFormat.M4A -> M4A
+         AudioFormat.WMA -> WMA
+         AudioFormat.WAV -> WAV
+         AudioFormat.FLAC -> FLAC
+      }
+
+      fun forFile(file: File) = valueOf(file.extension.toUpperCase(Locale.ROOT))
+   }
+}

app/src/main/java/com/bartlomiejpluta/ttsserver/core/web/server/WebServer.kt

@@ -5,6 +5,7 @@ import android.content.Intent
 import android.content.SharedPreferences
 import android.net.Uri
 import androidx.localbroadcastmanager.content.LocalBroadcastManager
+import cafe.adriel.androidaudioconverter.model.AudioFormat
 import com.bartlomiejpluta.ttsserver.core.sonos.queue.SonosQueue
 import com.bartlomiejpluta.ttsserver.core.tts.engine.TTSEngine
 import com.bartlomiejpluta.ttsserver.core.tts.exception.TTSException
@@ -14,6 +15,7 @@ import com.bartlomiejpluta.ttsserver.core.web.dto.SonosDTO
 import com.bartlomiejpluta.ttsserver.core.web.endpoint.Endpoint
 import com.bartlomiejpluta.ttsserver.core.web.endpoint.EndpointMatcher
 import com.bartlomiejpluta.ttsserver.core.web.exception.WebException
+import com.bartlomiejpluta.ttsserver.core.web.mime.MimeType
 import com.bartlomiejpluta.ttsserver.service.foreground.ForegroundService
 import com.bartlomiejpluta.ttsserver.service.state.ServiceState
 import com.bartlomiejpluta.ttsserver.ui.preference.PreferenceKey
@@ -57,7 +59,12 @@ class WebServer(
    private fun dispatch(it: IHTTPSession): Response {
       return when (EndpointMatcher.match(it.uri)) {
          Endpoint.SAY -> say(it)
-         Endpoint.WAVE -> wave(it)
+         Endpoint.WAVE -> file(it, AudioFormat.WAV)
+         Endpoint.AAC -> file(it, AudioFormat.AAC)
+         Endpoint.MP3 -> file(it, AudioFormat.MP3)
+         Endpoint.M4A -> file(it, AudioFormat.M4A)
+         Endpoint.WMA -> file(it, AudioFormat.WMA)
+         Endpoint.FLAC -> file(it, AudioFormat.FLAC)
          Endpoint.SONOS -> sonos(it)
          Endpoint.SONOS_CACHE -> sonosCache(it)
          Endpoint.GONG -> gong(it)
@@ -103,8 +110,8 @@ class WebServer(
       }
    }
 
-   private fun wave(session: IHTTPSession): Response {
-      if (!preferences.getBoolean(PreferenceKey.ENABLE_WAVE_ENDPOINT, true)) {
+   private fun file(session: IHTTPSession, audioFormat: AudioFormat): Response {
+      if (!preferences.getBoolean(PreferenceKey.ENABLE_FILE_ENDPOINTS, true)) {
          throw WebException(NOT_FOUND)
       }
 
@@ -118,8 +125,8 @@ class WebServer(
 
       val dto = extractBody(session) { BaseDTO(it) }
 
-      val (stream, size) = tts.fetchTTSStream(dto.text, dto.language)
-      return newFixedLengthResponse(OK, MIME_WAVE, stream, size)
+      val (stream, size) = tts.fetchTTSStream(dto.text, dto.language, audioFormat)
+      return newFixedLengthResponse(OK, MimeType.forAudioFormat(audioFormat).mimeType, stream, size)
    }
 
    private fun sonos(session: IHTTPSession): Response {
@@ -160,7 +167,7 @@ class WebServer(
 
       val stream = BufferedInputStream(FileInputStream(file))
       val size = file.length()
-      return newFixedLengthResponse(OK, MIME_WAVE, stream, size)
+      return newFixedLengthResponse(OK, MimeType.forFile(file).mimeType, stream, size)
    }
 
    private fun gong(session: IHTTPSession): Response {
@@ -183,7 +190,7 @@ class WebServer(
          context.contentResolver.openInputStream(uri) ?: throw TTSException()
       )
 
-      return newFixedLengthResponse(OK, MIME_WAVE, stream, size)
+      return newFixedLengthResponse(OK, MimeType.WAV.mimeType, stream, size)
    }
 
    override fun start() {
@@ -208,7 +215,6 @@ class WebServer(
 
    companion object {
       private const val MIME_JSON = "application/json"
-      private const val MIME_WAVE = "audio/x-wav"
       private const val CONTENT_TYPE = "content-type"
       private val SUCCESS_RESPONSE = response("Request has been completed")
       private val QUEUED_RESPONSE = response("Request has been queued")

app/src/main/java/com/bartlomiejpluta/ttsserver/di/module/TTSModule.kt

@@ -7,6 +7,7 @@ import androidx.preference.PreferenceManager
 import com.bartlomiejpluta.ttsserver.core.sonos.queue.SonosQueue
 import com.bartlomiejpluta.ttsserver.core.tts.engine.TTSEngine
 import com.bartlomiejpluta.ttsserver.core.tts.status.TTSStatusHolder
+import com.bartlomiejpluta.ttsserver.core.util.AudioConverter
 import com.bartlomiejpluta.ttsserver.core.util.NetworkUtil
 import com.bartlomiejpluta.ttsserver.core.web.server.WebServerFactory
 import com.bartlomiejpluta.ttsserver.service.notification.ForegroundNotificationFactory
@@ -32,8 +33,9 @@ class TTSModule {
       context: Context,
       textToSpeech: TextToSpeech,
       ttsStatusHolder: TTSStatusHolder,
-      preferences: SharedPreferences
-   ) = TTSEngine(context, textToSpeech, ttsStatusHolder, preferences)
+      preferences: SharedPreferences,
+      converter: AudioConverter
+   ) = TTSEngine(context, textToSpeech, ttsStatusHolder, preferences, converter)
 
    @Provides
    @Singleton
@@ -58,6 +60,10 @@ class TTSModule {
    fun networkUtil(context: Context, preferences: SharedPreferences) =
       NetworkUtil(context, preferences)
 
+   @Provides
+   @Singleton
+   fun adudioConverter(context: Context) = AudioConverter(context)
+
    @Provides
    @Singleton
    fun sonosQueue(tts: TTSEngine, preferences: SharedPreferences, networkUtil: NetworkUtil) =

app/src/main/java/com/bartlomiejpluta/ttsserver/ui/preference/Preference.kt

@@ -5,7 +5,7 @@ object PreferenceKey {
    const val PORT = "preference_port"
    const val ENABLE_HTTP_DEBUG = "preference_http_debug"
    const val ENABLE_SAY_ENDPOINT = "preference_enable_say_endpoint"
-   const val ENABLE_WAVE_ENDPOINT = "preference_enable_wave_endpoint"
+   const val ENABLE_FILE_ENDPOINTS = "preference_enable_file_endpoints"
    const val ENABLE_SONOS_ENDPOINT = "preference_enable_sonos_endpoint"
    const val ENABLE_GONG = "preference_enable_gong"
    const val GONG = "preference_gong"

app/src/main/java/com/bartlomiejpluta/ttsserver/ui/preference/PreferencesFragment.kt

@@ -12,6 +12,7 @@ import androidx.preference.Preference
 import androidx.preference.PreferenceFragmentCompat
 import androidx.preference.SwitchPreference
 import com.bartlomiejpluta.R
+import com.bartlomiejpluta.ttsserver.core.web.mime.MimeType
 import com.bartlomiejpluta.ttsserver.service.foreground.ForegroundService
 import com.bartlomiejpluta.ttsserver.service.state.ServiceState
 
@@ -19,7 +20,7 @@ import com.bartlomiejpluta.ttsserver.service.state.ServiceState
 class PreferencesFragment : PreferenceFragmentCompat() {
    private lateinit var portPreference: IntEditTextPreference
    private lateinit var sayEndpointPreference: SwitchPreference
-   private lateinit var waveEndpointPreference: SwitchPreference
+   private lateinit var fileEndpointPreference: SwitchPreference
    private lateinit var sonosEndpointPreference: SwitchPreference
    private lateinit var httpDebugPreference: SwitchPreference
    private lateinit var enableGongPreference: SwitchPreference
@@ -61,7 +62,7 @@ class PreferencesFragment : PreferenceFragmentCompat() {
       portPreference.setOnBindEditTextListener { it.inputType = InputType.TYPE_CLASS_NUMBER }
       httpDebugPreference = findPreference(PreferenceKey.ENABLE_HTTP_DEBUG)!!
       sayEndpointPreference = findPreference(PreferenceKey.ENABLE_SAY_ENDPOINT)!!
-      waveEndpointPreference = findPreference(PreferenceKey.ENABLE_WAVE_ENDPOINT)!!
+      fileEndpointPreference = findPreference(PreferenceKey.ENABLE_FILE_ENDPOINTS)!!
       sonosEndpointPreference = findPreference(PreferenceKey.ENABLE_SONOS_ENDPOINT)!!
       enableGongPreference = findPreference(PreferenceKey.ENABLE_GONG)!!
       enableGongPreference.setOnPreferenceClickListener { preference ->
@@ -90,7 +91,7 @@ class PreferencesFragment : PreferenceFragmentCompat() {
    private fun openFilePicker(preference: Preference?) {
       if ((preference as SwitchPreference).isChecked) {
          val intent = Intent(Intent.ACTION_OPEN_DOCUMENT)
-            .apply { type = "audio/x-wav" }
+            .apply { type = MimeType.WAV.mimeType }
             .let { Intent.createChooser(it, getString(R.string.preference_gong_picker_prompt)) }
 
          startActivityForResult(intent, PICKFILE_RESULT_CODE)

app/src/main/res/values/strings.xml

@@ -19,8 +19,8 @@
     <string name="preference_port_title">Server port</string>
     <string name="preference_enable_say_endpoint_summary">Allow HTTP clients to use /say endpoint which enables them to say message using builtin speakers or external ones connected to the device</string>
     <string name="preference_enable_say_endpoint_title">Enable /say endpoint</string>
-    <string name="preference_enable_wave_endpoint_summary">Allow HTTP clients to use /wave endpoint which enables them to retrieve TTS message as wave file</string>
-    <string name="preference_enable_wave_endpoint_title">Enable /wave endpoint</string>
+    <string name="preference_enable_file_endpoints_summary">Allow HTTP clients to use file endpoints (/wav, /mp3, /flac etc.) which enables them to retrieve TTS message as audio file</string>
+    <string name="preference_enable_file_endpoints_title">Enable file endpoints</string>
     <string name="preference_enable_sonos_endpoint_summary">Allow HTTP clients to use /sonos endpoint which enables them to send TTS messages directly to Sonos devices</string>
     <string name="preference_enable_sonos_endpoint_title">Enable /sonos endpoint</string>
     <string name="preference_http_debug_summary">Attach the stacktrace to each error HTTP response if available</string>

app/src/main/res/xml/preferences.xml

@@ -27,9 +27,9 @@
 
         <SwitchPreference
             android:defaultValue="true"
-            android:key="preference_enable_wave_endpoint"
-            android:summary="@string/preference_enable_wave_endpoint_summary"
-            android:title="@string/preference_enable_wave_endpoint_title"
+            android:key="preference_enable_file_endpoints"
+            android:summary="@string/preference_enable_file_endpoints_summary"
+            android:title="@string/preference_enable_file_endpoints_title"
             app:iconSpaceReserved="false" />
 
         <SwitchPreference