import gradio as gr
import torch
from transformers import AutoProcessor, BarkModel
import scipy


# device = "cuda" if torch.cuda.is_available() else "cpu"
# model = BarkModel.from_pretrained("suno/bark-small", torch_dtype=torch.float16).to(device)
# model.enable_cpu_offload()

device = "cpu"


processor = AutoProcessor.from_pretrained("suno/bark-small")
model = BarkModel.from_pretrained("suno/bark-small").to(device)
num_list = ["1","2","3","4","5","6","7","8","9","10"]
lang_list = ["en","de"]

def run_bark(text, n, lang):
    #history_prompt = []
    semantic_prompt=f"v2/{lang}_speaker_{int(n-1)}"

        #text=["Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe."],
    inputs = processor(text=text,
        voice_preset = semantic_prompt,
        return_tensors="pt",
    )
    
    speech_values = model.generate(**inputs, do_sample=True)
    sampling_rate = model.generation_config.sample_rate

    #sampling_rate = model.config.sample_rate
    #sampling_rate = 24000
    scipy.io.wavfile.write("bark_out.wav", rate=sampling_rate, data=speech_values.cpu().numpy().squeeze())
    return ("bark_out.wav")

with gr.Blocks() as app:
    with gr.Column():
        in_text = gr.Textbox()
        with gr.Row():
            speaker_num = gr.Dropdown(label="Speaker Voice", choices=num_list,value="1")
            speaker_lang = gr.Dropdown(label="Speaker Language", choices=lang_list,value="en")
            
        #speaker_num = gr.Number(value=0)
        go_btn = gr.Button()

    with gr.Column():
        out_audio = gr.Audio()

    go_btn.click(run_bark,[in_text, speaker_num, speaker_lang],out_audio)

app.launch()