import io

import gradio as gr
import librosa
import numpy as np
import soundfile
from inference.infer_tool import Svc
import logging
from logmmse import logmmse

logging.getLogger('numba').setLevel(logging.WARNING)
#
model_name = "logs/32k/uma1.pth"
config_name = "configs/uma1.json"

svc = Svc(model_name, config_name)
sid_map = {
    "米浴":"rice",
    "东海帝皇":"teio",
    "爱慕织姬":"aimya"
}


def vc_fn(sid, input_audio, vc_transform):
    if input_audio is None:
        return "You need to upload an audio", None
    sampling_rate, audio = input_audio
    # print(audio.shape,sampling_rate)
    duration = audio.shape[0] / sampling_rate
    audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
    if len(audio.shape) > 1:
        audio = librosa.to_mono(audio.transpose(1, 0))
    if sampling_rate != 32000:
        audio = librosa.resample(audio, orig_sr=sampling_rate, target_sr=32000)
    audio = logmmse(audio, 32000)
    print(audio.shape)
    out_wav_path = io.BytesIO()
    soundfile.write(out_wav_path, audio, 32000, format="wav")
    out_wav_path.seek(0)
    sid = sid_map[sid]
    out_audio, _out_sr = svc.infer(sid, vc_transform, out_wav_path)

    _audio = out_audio.cpu().numpy()
    return "Success", (32000, _audio)


app = gr.Blocks()
with app:
    with gr.Tabs():
        with gr.TabItem("Basic"):
            gr.Markdown(value="""
                # 前言
                本demo基于[sovits 3.0 32khz版本](https://github.com/innnky/so-vits-svc)训练的，并改写于 `https://huggingface.co/spaces/innnky/nyaru-svc-3.0`，
                https://huggingface.co/spaces/yukie/yukie-sovits3等
                特别感谢innnky佬与yukie佬
                加载赛马娘语音，自用。
                """)
            sid = gr.Dropdown(label="音色", choices=[
                              "东海帝皇", "米浴","爱慕织姬"], value="东海帝皇")
            vc_input3 = gr.Audio(label="上传音频")
            vc_transform = gr.Number(
                label="变调（整数，可以正负，半音数量，升高八度就是12）", value=0)
            vc_submit = gr.Button("转换", variant="primary")
            vc_output1 = gr.Textbox(label="Output Message")
            vc_output2 = gr.Audio(label="Output Audio")
            gr.Markdown(value="""
                ## 注意
                不要使用太长的语音
                """)
        vc_submit.click(vc_fn, [sid, vc_input3, vc_transform], [
                        vc_output1, vc_output2])

    app.launch()