Update app.py

app.py

@@ -5,6 +5,8 @@ from sklearn.metrics.pairwise import cosine_similarity
 import base64
 from pydub import AudioSegment
 from hezar.models import Model
+import librosa
+import soundfile as sf
 
 st.set_page_config(
     page_title="Sing It Forward App",
@@ -66,8 +68,6 @@ st.write('------')
 
 
 
-
-
 def cosine_sim(text1, text2):
     vectorizer = TfidfVectorizer().fit_transform([text1, text2])
     vectors = vectorizer.toarray()
@@ -97,6 +97,45 @@ def take_challenge(music_file, typed_lyrics, key, language):
             st.error('Awful! Try harder next time', icon="🚨")
             st.markdown('<style>div.stAlert { background-color: rgba(241, 36, 36, 0.9); }</style>', unsafe_allow_html=True)
 
+def change_volume(input_file, output_file, volume_factor):
+    sound = AudioSegment.from_mp3(input_file)
+    volume_changed = sound + volume_factor
+    volume_changed.export(output_file, format="mp3")
+
+def change_speed(input_file, output_file, speed_factor):
+    sound, sr = librosa.load(input_file)
+    speed_changed = librosa.effects.time_stretch(sound, rate=speed_factor)
+    sf.write(output_file, speed_changed, sr)
+
+def change_pitch(input_file, output_file, pitch_factor):
+    sound, sr = librosa.load(input_file)
+    pitch_changed = librosa.effects.pitch_shift(sound, sr=sr, n_steps=pitch_factor)
+    sf.write(output_file, pitch_changed, sr)
+
+def low_pass_filter(input_file, output_file, cutoff_freq):
+    sound = AudioSegment.from_mp3(input_file)
+    low_filtered_sound = sound.low_pass_filter(cutoff_freq)
+    low_filtered_sound.export(output_file, format="mp3")
+
+def high_pass_filter(input_file, output_file, cutoff_freq):
+    sound = AudioSegment.from_mp3(input_file)
+    high_filtered_sound = sound.high_pass_filter(cutoff_freq)
+    high_filtered_sound.export(output_file, format="mp3")
+
+def pan_left_right(input_file, output_file, pan_factor):
+    sound = AudioSegment.from_mp3(input_file)
+    pan_sound = sound.pan(pan_factor)
+    pan_sound.export(output_file, format="mp3")
+
+def fade_in_ms(input_file, output_file, fade_factor):
+    sound = AudioSegment.from_mp3(input_file)
+    faded_sound = sound.fade_in(fade_factor)
+    faded_sound.export(output_file, format="mp3")
+
+def fade_out_ms(input_file, output_file, fade_factor):
+    sound = AudioSegment.from_mp3(input_file)
+    faded_sound = sound.fade_out(fade_factor)
+    faded_sound.export(output_file, format="mp3")
 
 
             
@@ -117,25 +156,75 @@ with tab2:
         with open("raw_music.mp3", "wb") as f:
             f.write(uploaded_file.getbuffer())
         st.audio("raw_music.mp3")
-        trimm_check = st.checkbox("Trim?")
+
+        current_input = "raw_music.mp3"
+        output_file = "processed_music.mp3"
+
+        trimm_check = st.checkbox("Trim")
         if trimm_check:   
             st.write("Specify start and end times for trimming:")
-            audio = AudioSegment.from_file("raw_music.mp3")
+            audio = AudioSegment.from_mp3(current_input)
             duration = len(audio) // 1000 
             start_time = st.number_input("Start Time (seconds)", min_value=0, max_value=duration, value=0)
             end_time = st.number_input("End Time (seconds)", min_value=0, max_value=duration, value=duration)
             if start_time < end_time:
                 trimmed_audio = audio[start_time * 1000:end_time * 1000]
-                trimmed_audio.export("trimmed_music.mp3", format="mp3")
-                st.write("Now type what user should sing:")
-                typed_lyrics = st.text_area("Lyrics to be singed:")
-                st.write('------')
-                take_challenge("trimmed_music.mp3", typed_lyrics, "unique_key_1", language)
+                trimmed_audio.export(output_file, format="mp3")
+                current_input = output_file
             else:
                 st.error('Start Time should be smaller than End Time!', icon="❌")
                 st.markdown('<style>div.stAlert { background-color: rgba(241, 36, 36, 0.9); }</style>', unsafe_allow_html=True)
-        else:
-            st.write("Now type what user should sing:")
-            typed_lyrics = st.text_area("Lyrics to be singed:")
-            st.write('------')
-            take_challenge("raw_music.mp3", typed_lyrics, "unique_key_2", language)
+
+        volume_checkbox = st.checkbox("Change Volume")
+        if volume_checkbox:
+            volume_factor = st.slider("Volume Factor (dB)", -30, 30, 0)
+            change_volume(current_input, output_file, volume_factor)
+            current_input = output_file
+
+        speed_checkbox = st.checkbox("Change Speed")
+        if speed_checkbox:
+            speed_factor = st.slider("Speed Factor", 0.25, 2.0, 1.0)
+            change_speed(current_input, output_file, speed_factor)
+            current_input = output_file
+
+        pitch_checkbox = st.checkbox("Change Pitch")
+        if pitch_checkbox:
+            pitch_factor = st.slider("Pitch Shift (fractional steps)", -12, 12, 0)
+            change_pitch(current_input, output_file, pitch_factor)
+            current_input = output_file
+
+        low_pass_checkbox = st.checkbox("Low Pass Filter")
+        if low_pass_checkbox:
+            cutoff_freq = st.slider("Low Pass Filter Cutoff Frequency", min_value=20, max_value=20000, value=2000)
+            low_pass_filter(current_input, output_file, cutoff_freq)
+            current_input = output_file
+
+        high_pass_checkbox = st.checkbox("High Pass Filter")
+        if high_pass_checkbox:
+            cutoff_freq = st.slider("High Pass Filter Cutoff Frequency", min_value=20, max_value=20000, value=2000)
+            high_pass_filter(current_input, output_file, cutoff_freq)
+            current_input = output_file
+
+        pan_checkbox = st.checkbox("Pan Left/Right")
+        if pan_checkbox:
+            pan_factor = st.slider("Pan Factor (-1 for Left, 1 for Right)", -1.0, 1.0, 0.0)
+            pan_left_right(current_input, output_file, pan_factor)
+            current_input = output_file
+
+        fade_in_checkbox = st.checkbox("Fade In")
+        if fade_in_checkbox:
+            fade_in_time = st.slider("Fade In Duration (ms)", min_value=0, max_value=10000, value=1000)
+            fade_in_ms(current_input, output_file, fade_in_time)
+            current_input = output_file
+
+        fade_out_checkbox = st.checkbox("Fade Out")
+        if fade_out_checkbox:
+            fade_out_time = st.slider("Fade Out Duration (ms)", min_value=0, max_value=10000, value=1000)
+            fade_out_ms(current_input, output_file, fade_out_time)
+            current_input = output_file
+
+
+        st.write("Now type what user should sing:")
+        typed_lyrics = st.text_area("Lyrics to be singed:")
+        st.write('------')
+        take_challenge(current_input, typed_lyrics, "unique_key_1", language)