Update app.py

mfcc classification

app.py

@@ -1,6 +1,44 @@
 import numpy as np
 import gradio as gr
 
+import pickle
+import librosa
+
+from sklearn.linear_model import LogisticRegression
+
+lr_model = LogisticRegression()
+
+with open("./lr_model_mfcc.pkl", "rb") as f:
+    lr_model = pickle.load(f)
+
+def extract_mfcc_gradio(audio):
+    
+    sample_rate, y = audio
+
+    y = y.astype(np.float32)
+    y /= np.max(np.abs(y))
+
+    if y.ndim == 1:
+        data = y
+    else:
+        data = y[:, 0]
+
+    mfcc = np.mean(librosa.feature.mfcc(y=data, sr=sample_rate).T, axis=0)
+
+    return mfcc
+
+def voice_mfcc_classification(audio):
+
+    mfcc = extract_mfcc_gradio(audio)
+
+    prediction = lr_model.predict([mfcc])
+
+    if prediction[0] == 0:
+        return "engaging"
+    else:
+        return "boring"
+
+
 def voice_classification(audio):
 
     # generate a random number between 0 and 1
@@ -13,4 +51,4 @@ def voice_classification(audio):
     
     return result
 
-gr.Interface(fn=voice_classification, inputs="audio", outputs="text").launch()
+gr.Interface(fn=voice_mfcc)classification, inputs="audio", outputs="text").launch()