Remove support for GoogleCloud, which was not working yet actually. WhisperX is way better IMO

app.py

@@ -21,7 +21,7 @@ def main():
     title = '🎙️ Meetings Note Taker 🎙️'
     st.title(title)
     st.write(
-        'Upload an audio file, transcribe it using WhisperX, GoogleCloud or Assembly.AI, and generate meeting notes using your selected model.'
+        'Upload an audio file, transcribe it using WhisperX or AssemblyAI, and generate meeting notes using your selected model.'
     )
 
     openai_api_key = os.environ.get('OPENAI_API_KEY') or st.text_input(
@@ -29,16 +29,11 @@ def main():
     )
 
     engine_type = os.environ.get('TRANSCRIPTION_ENGINE') or st.selectbox(
-        'Select a transcription engine:', ['WhisperX', 'AssemblyAI', 'Google']
+        'Select a transcription engine:', ['WhisperX', 'AssemblyAI']
     )
-    if engine_type in ['AssemblyAI', 'WhisperX']:
-        engine_api_key = os.environ.get(
-            f'{engine_type.upper()}_API_KEY'
-        ) or st.text_input(f'Enter your {engine_type} API key:', type='password')
-    else:
-        engine_api_key = (
-            None  # Google doesn't need an API key but uses a credentials file
-        )
+    engine_api_key = os.environ.get(
+        f'{engine_type.upper()}_API_KEY'
+    ) or st.text_input(f'Enter your {engine_type} API key:', type='password')
     openai_model = os.environ.get('OPENAI_MODEL') or st.selectbox(
         'Select a model:', ['gpt-3.5-turbo-16k', 'gpt-4-0613']
     )

engines.py

@@ -7,8 +7,6 @@ import requests
 import torch
 import whisperx
 
-from google.cloud import speech_v2 as speech
-
 
 class TranscriptEngine(Protocol):
     """Protocol for a transcription engine"""
@@ -58,31 +56,6 @@ class AssemblyAI:
         )
 
 
-class GoogleCloud:
-    def __init__(self, api_key: str):
-        pass  # do not need an API key for Google Cloud
-
-    def transcribe(self, language, audio_file: BytesIO) -> str:
-        client = speech.SpeechClient()
-
-        audio = speech.RecognitionAudio(content=audio_file.read())
-
-        config = speech.RecognitionConfig(
-            encoding=speech.RecognitionConfig.AudioEncoding.ENCODING_UNSPECIFIED,
-            language_code=language,
-            diarization_config=speech.SpeakerDiarizationConfig(
-                enable_speaker_diarization=True,
-            ),
-        )
-
-        operation = client.long_running_recognize(config=config, audio=audio)
-        response = operation.result()
-
-        return ' '.join(
-            result.alternatives[0].transcript for result in response.results
-        )
-
-
 class WhisperX:
     def __init__(self, api_key: str, device: str = 'cuda', compute_type: str = 'int8', batch_size: int = 8):
         self.api_key = api_key  # HuggingFace API key
@@ -129,7 +102,6 @@ class WhisperX:
 def get_engine(engine_type: str, api_key: str | None) -> TranscriptEngine:
     engine_cls = {
         'AssemblyAI': AssemblyAI,
-        'Google': GoogleCloud,
         'WhisperX': WhisperX,
     }[engine_type]
 

requirements.txt

@@ -1,7 +1,6 @@
 requests>=2.31.0
 streamlit>=1.25.0
 python-dotenv>=1.0.0
-google_cloud_speech>=2.21.0
 torch==2.0.0
 torchvision==0.15.1
 torchaudio==2.0.1