Update app.py

app.py

@@ -49,6 +49,7 @@ with open('dict.pickle', 'rb') as handle:
 import streamlit as st
 import tempfile
 import librosa
+import numpy as np
 
 from streamlit_webrtc import (
     AudioProcessorBase,
@@ -68,11 +69,12 @@ class MicrophoneRecorder(AudioProcessorBase):
 
     def get_audio(self):
         y = np.concatenate(self.frames)
-        return y
+        return y.tobytes()
 
     def save_to_file(self, file_path):
         y = self.get_audio()
-        wf.write(file_path, self.sample_rate, y)
+        with open(file_path, "wb") as f:
+            f.write(y)
 
     def on_receive(self, data):
         if data["type"] == "media":
@@ -98,21 +100,20 @@ def record_audio():
         ),
     )
     if webrtc_ctx.audio_receiver and webrtc_ctx.state.playing:
-        audio = recorder.get_audio()
-        return audio
+        recorder.save_to_file('recording.wav')
+        st.write("Audio saved!")
+        return True
     else:
         st.write("WebRTC state is not playing")
-        return None
+        return False
 
 st.set_page_config(page_title='Tune Recognition', page_icon=':musical_note:')
 st.title('Chord Song Recognition')
 st.markdown('Click the button below to start recording your tune.')
-audio = record_audio()
-if audio is not None:
-    st.write("Audio saved!")
-    
-    rec_file = 'recording.wav'
-    song, sr = librosa.load(rec_file)
+audio_recorded = record_audio()
+
+if audio_recorded:
+    song, sr = librosa.load('recording.wav')
     to_match = np.copy(song[0:220500])
 
     # Create spectrogram image of the song to match
@@ -125,7 +126,7 @@ if audio is not None:
     # Get the embedding of the song to match
     to_match_emb = embedding_model.predict(to_match_img)
 
-   # Calculate the distances between the song to match and the songs in the database
+    # Calculate the distances between the song to match and the songs in the database
     songsdistdict = {}
     for key, values in songspecdict.items():
         dist_array = []