hopefuly

src/predict.py

@@ -289,9 +289,20 @@ class Predictor:
         reference_speaker = local_file_path
         target_se, audio_name = se_extractor.get_se(reference_speaker, tone_color_converter, target_dir=openvoice_dir, vad=False)
         src_path = os.path.join(results_dir,f"{gen_id}-tmp.wav")
-        openvoice_output = os.path.join(results_dir,f"{gen_id}-voice-clone-2.wav")
-        base_speaker_tts.tts(passage,src_path,speaker='default',language='English',speed=1.0)
-        return {"status":"Done"}
+        openvoice_output = os.path.join(results_dir,f"{gen_id}-voice-clone-emotions.wav")
+        print("extracting emotions from openai")
+        base_speaker_tts.tts(passage,src_path,speaker='default',language='English',speed=1.0,use_emotions=True)
+        source_se, audio_name = se_extractor.get_se(src_path, tone_color_converter, target_dir=openvoice_dir, vad=False)
+        tone_color_converter.convert(audio_src_path=src_path,src_se=source_se,tgt_se=target_se,output_path=openvoice_output,message='')
+        if process_audio:
+            (new_sr, wav1) = self._fn(openvoice_output,"Midpoint",32,0.5)
+            sf.write(openvoice_output,wav1,new_sr)
+        mp3_final_output_2 = str(openvoice_output).replace('wav','mp3')
+        self.convert_wav_to_mp3(openvoice_output,mp3_final_output_2)
+        self.upload_file_to_s3(mp3_final_output_2,'demovidelyusergenerations',f"{gen_id}-voice-clone-emotions.mp3")
+        shutil.rmtree(os.path.join(output_dir,gen_id))
+        return {"voice_clone_emotions":f"https://demovidelyusergenerations.s3.amazonaws.com/{gen_id}-voice-clone-emotions.mp3",
+                }
         
 
     def predict_with_multi_lang(self,s3_url,passage,process_audio):