import os
import torch
from openvoice import se_extractor
from openvoice.api import ToneColorConverter

ckpt_converter = '/kaggle/working/converter'
device = "cuda:0" if torch.cuda.is_available() else "cpu"
output_dir = 'outputs_v2'

tone_color_converter = ToneColorConverter(f'{ckpt_converter}/config.json', device=device)
tone_color_converter.load_ckpt(f'{ckpt_converter}/checkpoint.pth')

os.makedirs(output_dir, exist_ok=True)

reference_speaker = '/kaggle/working/industrialRevolution.wav' # This is the voice you want to clone
target_se, audio_name = se_extractor.get_se(reference_speaker, tone_color_converter, vad=False)

source_se = torch.load(f'/kaggle/working/model.pth', map_location=device)

import time
#!wget https://huggingface.co/datasets/leafspark/openai-voices/resolve/main/juniper/RPReplay_Final1716272820.mp4.wav
out = '/kaggle/working/output.wav'
cloned_voice = "RPReplay_Final1716272820.mp4.wav"
#target_se, audio_name = se_extractor.get_se(cloned_voice, tone_color_converter, vad=False)
wav = tts.tts(text="no, yes, nah, yeah, damn!", speaker_wav=cloned_voice, language="en", emotion="happy")
scipy.io.wavfile.write(out, 24000, np.array(wav))
time1 = time.time()
tone_color_converter.convert(
            audio_src_path=out, 
            src_se=source_se, 
            tgt_se=target_se, 
            output_path="/kaggle/working/tmp.wav")
print(time.time() - time1)
from IPython.display import Audio
display(Audio("/kaggle/working/tmp.wav"))
display(Audio(wav, rate=24000))