[whisper] Add OpenAI API compatibility

Apply PR comments

Signed-off-by: Gwendal Roulleau <[email protected]>

Author: dalgwen

bundles/org.openhab.voice.whisperstt/src/main/java/org/openhab/voice/whisperstt/internal/WhisperSTTService.java

@@ -286,7 +286,6 @@ public Set<AudioFormat> getSupportedFormats() {
     @Override
     public STTServiceHandle recognize(STTListener sttListener, AudioStream audioStream, Locale locale, Set<String> set)
             throws STTException {
-
         AtomicBoolean aborted = new AtomicBoolean(false);
         try {
             logger.debug("Creating VAD instance...");
@@ -607,7 +606,6 @@ private String recognizeLocal(int audioSamplesOffset, short[] audioSamples, Stri
     }
 
     private String recognizeAPI(int audioSamplesOffset, short[] audioStream, String language) throws STTException {
-
         // convert to byte array, Each short has 2 bytes
         int size = audioSamplesOffset * 2;
         ByteBuffer byteArrayBuffer = ByteBuffer.allocate(size).order(ByteOrder.LITTLE_ENDIAN);
@@ -621,7 +619,7 @@ private String recognizeAPI(int audioSamplesOffset, short[] audioStream, String
 
         try {
             AudioInputStream audioInputStream = new AudioInputStream(new ByteArrayInputStream(byteArray), jAudioFormat,
-                    size);
+                    audioSamplesOffset);
 
             // write stream as a WAV file, in a byte array stream :
             ByteArrayInputStream byteArrayInputStream = null;

bundles/org.openhab.voice.whisperstt/src/main/resources/OH-INF/i18n/whisperstt.properties

@@ -3,6 +3,12 @@
 addon.whisperstt.name = Whisper Speech-to-Text
 addon.whisperstt.description = Whisper STT Service uses the whisper.cpp library to transcript audio data to text.
 
+voice.config.whisperstt.apiKey.label = API Key
+voice.config.whisperstt.apiKey.description = Key to access the API
+voice.config.whisperstt.apiModelName.label = API Model
+voice.config.whisperstt.apiModelName.description = Model name to use (API only). Default to OpenAI only available model (whisper-1).
+voice.config.whisperstt.apiUrl.label = API Url
+voice.config.whisperstt.apiUrl.description = OpenAI compatible API URL. Default to OpenAI transcription service.
 voice.config.whisperstt.audioContext.label = Audio Context
 voice.config.whisperstt.audioContext.description = Overwrite the audio context size. (0 to use whisper default context size)
 voice.config.whisperstt.beamSize.label = Beam Size
@@ -24,27 +30,35 @@ voice.config.whisperstt.greedyBestOf.description = Best Of configuration for sam
 voice.config.whisperstt.group.developer.label = Developer
 voice.config.whisperstt.group.developer.description = Options added for developers.
 voice.config.whisperstt.group.grammar.label = Grammar
-voice.config.whisperstt.group.grammar.description = Define a grammar to improve transcrptions.
+voice.config.whisperstt.group.grammar.description = Define a grammar to improve transcriptions.
 voice.config.whisperstt.group.messages.label = Info Messages
 voice.config.whisperstt.group.messages.description = Configure service information messages.
+voice.config.whisperstt.group.openaiapi.label = API Configuration Options
+voice.config.whisperstt.group.openaiapi.description = Configure OpenAI compatible API, if you don't want to use the local model.
 voice.config.whisperstt.group.stt.label = STT Configuration
 voice.config.whisperstt.group.stt.description = Configure Speech to Text.
 voice.config.whisperstt.group.vad.label = Voice Activity Detection
-voice.config.whisperstt.group.vad.description = Configure the VAD mechanisim used to isolate single phrases to feed whisper with.
+voice.config.whisperstt.group.vad.description = Configure the VAD mechanism used to isolate single phrases to feed whisper with.
 voice.config.whisperstt.group.whisper.label = Whisper Options
 voice.config.whisperstt.group.whisper.description = Configure the whisper.cpp transcription options.
 voice.config.whisperstt.initSilenceSeconds.label = Initial Silence Seconds
 voice.config.whisperstt.initSilenceSeconds.description = Max initial seconds of silence to discard transcription.
 voice.config.whisperstt.initialPrompt.label = Initial Prompt
 voice.config.whisperstt.initialPrompt.description = Initial prompt to feed whisper with.
+voice.config.whisperstt.language.label = Language
+voice.config.whisperstt.language.description = If specified, speed up recognition by avoiding auto-detection. Default to system locale.
 voice.config.whisperstt.maxSeconds.label = Max Transcription Seconds
 voice.config.whisperstt.maxSeconds.description = Seconds to force transcription before silence detection.
 voice.config.whisperstt.maxSilenceSeconds.label = Max Silence Seconds
 voice.config.whisperstt.maxSilenceSeconds.description = Seconds of silence to trigger transcription.
 voice.config.whisperstt.minSeconds.label = Min Transcription Seconds
 voice.config.whisperstt.minSeconds.description = Min transcription seconds passed to whisper.
-voice.config.whisperstt.modelName.label = Model Name
-voice.config.whisperstt.modelName.description = Model name without extension.
+voice.config.whisperstt.mode.label = Local Mode Or API
+voice.config.whisperstt.mode.description = Use the local model or the OpenAI compatible API.
+voice.config.whisperstt.mode.option.LOCAL = Local
+voice.config.whisperstt.mode.option.API = OpenAI API
+voice.config.whisperstt.modelName.label = Local Model Name
+voice.config.whisperstt.modelName.description = Model name without extension. Local mode only.
 voice.config.whisperstt.openvinoDevice.label = OpenVINO Device
 voice.config.whisperstt.openvinoDevice.description = Initialize OpenVINO encoder. (built-in binaries do not support OpenVINO, this has no effect)
 voice.config.whisperstt.preloadModel.label = Preload Model