Update app.py to use firebase storage

app.py

@@ -5,16 +5,24 @@ from transformers import Wav2Vec2Processor, AutoModelForCTC
 import zipfile
 import os
 import firebase_admin
-from firebase_admin import credentials, firestore
-from datetime import datetime
+from firebase_admin import credentials, firestore, storage
+from datetime import datetime, timedelta
 import json
 import tempfile
+import uuid
 
-# Initialize Firebase
+''' LOCAL INITIALIZATION - ONLY USE ON YOUR OWN DEVICE 
+os.chdir(os.path.dirname(os.path.abspath(__file__)))
+cred = credentials.Certificate("serviceAccountKey.json")
+'''
+# Deployed Initialization
 firebase_config = json.loads(os.environ.get('firebase_creds'))
 cred = credentials.Certificate(firebase_config)
-firebase_admin.initialize_app(cred)
+firebase_admin.initialize_app(cred, {
+    "storageBucket": "amis-asr-corrections-dem-8cf3d.firebasestorage.app"
+})
 db = firestore.client()
+bucket = storage.bucket()
 
 # Load the ASR model and processor
 MODEL_NAME = "eleferrand/xlsr53_Amis"
@@ -37,23 +45,37 @@ def transcribe(audio_file):
 def transcribe_both(audio_file):
     start_time = datetime.now()
     transcription = transcribe(audio_file)
-    processing_time = (datetime.now() - start_time).total_seconds()
-    return transcription, transcription, processing_time
+    return transcription, transcription
 
 def store_correction(original_transcription, corrected_transcription, audio_file, age, native_speaker):
     try:
         audio_metadata = {}
+        audio_file_url = None
+        
+        # If an audio file is provided, upload it to Firebase Storage
         if audio_file and os.path.exists(audio_file):
             audio, sr = librosa.load(audio_file, sr=16000)
             duration = librosa.get_duration(y=audio, sr=sr)
             file_size = os.path.getsize(audio_file)
             audio_metadata = {'duration': duration, 'file_size': file_size}
+            
+            # Generate a unique identifier for the audio file
+            unique_id = str(uuid.uuid4())
+            destination_path = f"audio/{unique_id}.mp3"
+            
+            # Create a blob and upload the file
+            blob = bucket.blob(destination_path)
+            blob.upload_from_filename(audio_file)
+            
+            # Generate a signed download URL valid for 1 hour (adjust expiration as needed)
+            audio_file_url = blob.generate_signed_url(expiration=timedelta(hours=1))
         
         combined_data = {
             'original_text': original_transcription,
             'corrected_text': corrected_transcription,
             'timestamp': datetime.now().isoformat(),
             'audio_metadata': audio_metadata,
+            'audio_file_url': audio_file_url,
             'model_name': MODEL_NAME,
             'user_info': {
                 'native_amis_speaker': native_speaker,
@@ -117,7 +139,6 @@ def toggle_language(switch):
             "Download ZIP File"
         )
 
-# Interface
 # Interface
 with gr.Blocks() as demo:
     lang_switch = gr.Checkbox(label="切換到繁體中文 (Switch to Traditional Chinese)")
@@ -125,26 +146,22 @@ with gr.Blocks() as demo:
     title = gr.Markdown("Amis ASR Transcription & Correction System")
     step1 = gr.Markdown("Step 1: Audio Upload & Transcription")
     
-    # Audio input and playback (Original section)
     with gr.Row():
         audio_input = gr.Audio(sources=["upload", "microphone"], type="filepath", label="Audio Input")
     
     step2 = gr.Markdown("Step 2: Review & Edit Transcription")
-    # Transcribe button below the audio input (Added this section to place the button below the playback)
-    with gr.Row():  # Added this Row to position the button below the audio input
+    with gr.Row():
         transcribe_button = gr.Button("Transcribe Audio")
     
     original_text = gr.Textbox(label="Original Transcription", interactive=False, lines=5)
     corrected_text = gr.Textbox(label="Corrected Transcription", interactive=True, lines=5)
 
     step3 = gr.Markdown("Step 3: User Information")
-    
     with gr.Row():
         age_input = gr.Slider(minimum=0, maximum=100, step=1, label="Age", value=25)
         native_speaker_input = gr.Checkbox(label="Native Amis Speaker?", value=True)
 
     step4 = gr.Markdown("Step 4: Save & Download")
-    
     with gr.Row():
         save_button = gr.Button("Save Correction")
         save_status = gr.Textbox(label="Save Status", interactive=False)
@@ -153,7 +170,6 @@ with gr.Blocks() as demo:
         download_button = gr.Button("Download ZIP File")
         download_output = gr.File()
 
-    # Toggle language dynamically
     lang_switch.change(
         toggle_language, 
         inputs=lang_switch,