Update app.py

app.py

@@ -8,10 +8,13 @@ sess = onnxruntime.InferenceSession('weightsgru.onnx')
 
 
 # Define a function to extract MFCC features from an audio file
+utterance_length = 384
+def extract_mfcc(file_path, utterance_length):
+    # Get raw .wav data and sampling rate from librosa's load function
+    raw_w, sampling_rate = librosa.load(file_path, sr=None)
 
-def extract_mfcc(audio_data, utterance_length):
-    # Obtain MFCC Features from raw audio data
-    mfcc_features = librosa.feature.mfcc(y=audio_data, sr=16000, n_mfcc=40, hop_length=512)
+    # Obtain MFCC Features from raw data
+    mfcc_features = librosa.feature.mfcc(y=raw_w, sr=sampling_rate, n_mfcc=40, hop_length=512)
     
     # Adjust the utterance length
     if mfcc_features.shape[1] > utterance_length:
@@ -26,9 +29,9 @@ def extract_mfcc(audio_data, utterance_length):
 
 
 # Define a function to classify audio using the ONNX model
-def classify_audio(audio_data):
+def classify_audio(audio_file):
     utterance_length = 384
-    mfcc_features = extract_mfcc(audio_data, utterance_length)
+    mfcc_features = extract_mfcc(audio_file, utterance_length)
     features = np.split(mfcc_features, 6)
     inputs = [np.expand_dims(segment, axis=0) for segment in features]
     classes = np.zeros((1, 10))
@@ -46,10 +49,8 @@ def classify_audio(audio_data):
     highest_bird_label = bird_labels[pred_final1]
     second_highest_bird_label = bird_labels[pred_final2]
     return highest_bird_label, second_highest_bird_label
-
-# Create Gradio interface
-inputs = gr.Audio(type="filepath", label="Upload Audio")
-outputs = [gr.Textbox(label="Highest Bird Label"), gr.Textbox(label="Second Highest Bird Label")]
+inputs = gr.Audio(type="filepath")
+outputs = [gr.Textbox(label="highest_bird_label"), gr.Textbox(label="second_highest_bird_label")]
 
 # Create Gradio interface
 interface = gr.Interface(
@@ -57,8 +58,12 @@ interface = gr.Interface(
     inputs=inputs,
     outputs=outputs,
     title="Bird Classification",
-    description="Upload an audio file or speak into the microphone to classify it into bird categories"
+    description="Upload an audio file and classify it into bird categories"
 )
 
 # Launch the interface
-interface.launch()
+interface.launch()    
+# highest_bird, second_highest_bird = classify_audio(audio_file)
+
+# print("Highest Bird Label:", highest_bird)
+# print("Second Highest Bird Label:", second_highest_bird)