Update main.py

main.py

@@ -1,15 +1,17 @@
 from google.colab import drive
 drive.mount('/content/drive')
 
-"""Install Dependencies
+"""Install Dependencies"""
 
 pip install transformers librosa torch soundfile numba numpy TTS datasets gradio protobuf==3.20.3
 
-"""Emotion Detection (Using Text Dataset)""""
+"""Emotion Detection (Using Text Dataset)
+
 """
+
 !pip install --upgrade numpy tensorflow transformers TTS
 
-!pip freeze > requirements.txt"""
+!pip freeze > requirements.txt
 
 from transformers import pipeline
 
@@ -51,6 +53,27 @@ def generate_emotional_speech(text, emotion):
     "shame": {"pitch": 0.8, "speed": 0.85},      # Quiet, subdued tone
 
 }
+    
+import librosa
+import soundfile as sf
+
+def adjust_pitch(audio_path, pitch_factor):
+    # Load audio
+    y, sr = librosa.load(audio_path)
+    # Adjust pitch
+    y_shifted = librosa.effects.pitch_shift(y, sr, n_steps=pitch_factor)
+    # Save adjusted audio
+    sf.write(audio_path, y_shifted, sr)
+
+def adjust_speed(audio_path, speed_factor):
+    # Load the audio file
+    y, sr = librosa.load(audio_path)
+
+    # Adjust the speed (this alters the duration of the audio)
+    y_speeded = librosa.effects.time_stretch(y, speed_factor)
+
+    # Save the adjusted audio
+    sf.write(audio_path, y_speeded, sr)
 
     # Retrieve pitch and speed based on detected emotion
     settings = emotion_settings.get(emotion, {"pitch": 1.0, "speed": 1.0})
@@ -75,7 +98,6 @@ emotion = "happy"
 output_audio = generate_emotional_speech("Welcome to the smart library!", emotion)
 print(f"Generated Speech Saved At: {output_audio}")
 
-
 """Integrating the Workflow"""
 
 from IPython.display import Audio, display
@@ -250,28 +272,6 @@ save_path = "/content/drive/My Drive/fine_tuned_tacotron2.pth"
 torch.save(model.state_dict(), save_path)
 
 
-import librosa
-import soundfile as sf
-
-def adjust_pitch(audio_path, pitch_factor):
-    # Load audio
-    y, sr = librosa.load(audio_path)
-    # Adjust pitch
-    y_shifted = librosa.effects.pitch_shift(y, sr, n_steps=pitch_factor)
-    # Save adjusted audio
-    sf.write(audio_path, y_shifted, sr)
-
-def adjust_speed(audio_path, speed_factor):
-    # Load the audio file
-    y, sr = librosa.load(audio_path)
-
-    # Adjust the speed (this alters the duration of the audio)
-    y_speeded = librosa.effects.time_stretch(y, speed_factor)
-
-    # Save the adjusted audio
-    sf.write(audio_path, y_speeded, sr)
-
-
 """Set up the Gradio interface"""
 
 import gradio as gr
@@ -318,7 +318,8 @@ def emotion_aware_tts_pipeline(input_text=None, file_input=None):
 
             # Generate audio
             audio_path = "output.wav"
-            tts_model.tts_to_file(text=input_text, file_path=audio_path)
+            tts_model.tts_to_file(text=input_text, file_path=audio_path, speed=speed, pitch=pitch)
+
 
 
             return f"Detected Emotion: {emotion} (Confidence: {confidence:.2f})", audio_path
@@ -345,4 +346,4 @@ iface = gr.Interface(
 )
 
 # Launch Gradio interface
-iface.launch()
+iface.launch()