Remove OpenAI's Whisper AI and Bug FixingHEADtrout_19.90.0trout/19.90.019.90.0master

 - Removed OpenAI's Whisper AI from agl-service-voiceagent and using
   whisper.cpp for speech-to-text.
 - Fix audio_recorder.
 - Update grpc protoc to include the online-mode status in ServiceStatus
 - Set online_mode flag default to 0
 - Change wake word to "hey automotive"

Bug-AGL: SPEC-5200
Change-Id: I9f1629cdcaef43498bf4cb9fdd950291a415819d
Signed-off-by: Anuj Solanki <anuj603362@gmail.com>

1 files changed, 22 insertions, 22 deletions

diff --git a/agl_service_voiceagent/utils/stt_model.py b/agl_service_voiceagent/utils/stt_model.py
index 7e8ad8b..0a092ea 100644
--- a/agl_service_voiceagent/utils/stt_model.py
+++ b/agl_service_voiceagent/utils/stt_model.py
@@ -21,7 +21,7 @@ import wave
 from agl_service_voiceagent.utils.common import generate_unique_uuid
 
 # import the whisper model
-import whisper
+# import whisper
 # for whisper timeout feature
 from concurrent.futures import ThreadPoolExecutor  
 import subprocess
@@ -93,28 +93,28 @@ class STTModel:
         return result
     
     # Recognize speech using the whisper model
-    def recognize_using_whisper(self,filename,language = None,timeout = 5,fp16=False):
-        """
-        Recognize speech and return the result as a JSON object.
-
-        Args:
-            filename (str): The path to the audio file.
-            timeout (int, optional): The timeout for recognition (default is 5 seconds).
-            fp16 (bool, optional): If True, use 16-bit floating point precision, (default is False) because cuda is not supported.
-            language (str, optional): The language code for recognition (default is None).
-
-        Returns:
-            dict: A JSON object containing recognition results.
-        """
-        def transcribe_with_whisper():
-            return self.whisper_model.transcribe(filename, language = language,fp16=fp16)
+    # def recognize_using_whisper(self,filename,language = None,timeout = 5,fp16=False):
+    #     """
+    #     Recognize speech and return the result as a JSON object.
+
+    #     Args:
+    #         filename (str): The path to the audio file.
+    #         timeout (int, optional): The timeout for recognition (default is 5 seconds).
+    #         fp16 (bool, optional): If True, use 16-bit floating point precision, (default is False) because cuda is not supported.
+    #         language (str, optional): The language code for recognition (default is None).
+
+    #     Returns:
+    #         dict: A JSON object containing recognition results.
+    #     """
+    #     def transcribe_with_whisper():
+    #         return self.whisper_model.transcribe(filename, language = language,fp16=fp16)
         
-        with ThreadPoolExecutor() as executor:
-            future = executor.submit(transcribe_with_whisper)
-            try:
-                return future.result(timeout=timeout)
-            except TimeoutError:
-                return {"error": "Transcription with Whisper exceeded the timeout."}
+    #     with ThreadPoolExecutor() as executor:
+    #         future = executor.submit(transcribe_with_whisper)
+    #         try:
+    #             return future.result(timeout=timeout)
+    #         except TimeoutError:
+    #             return {"error": "Transcription with Whisper exceeded the timeout."}
             
     def recognize_using_whisper_cpp(self,filename):
         command = self.whisper_cpp_path