streamlit/nx_noise_app.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
import asyncio
from enum import Enum
from io import BytesIO
import os
from pathlib import Path
import shutil
import sys
from typing import List
import uuid

pwd = os.path.abspath(os.path.dirname(__file__))
sys.path.append(os.path.join(pwd, "../"))

import edge_tts
import librosa
import librosa.display
import matplotlib.pyplot as plt
import numpy as np

import streamlit as st
from streamlit_shortcuts import shortcut_button

from project_settings import project_path, temp_dir

# ENTRYPOINT ["streamlit", "run", "streamlit/nx_noise_app.py", "--server.port=8501", "--server.address=0.0.0.0"]
# streamlit run streamlit/nx_noise_app.py --server.port=8501 --server.address=0.0.0.0


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--src_dir",
        default=(project_path / "data/speech/en-PH/2025-01-14/2025-01-14").as_posix(),
        type=str
    )
    parser.add_argument(
        "--tgt_dir",
        default=(project_path / "data/speech/en-PH/2025-01-14/2025-01-14/finished").as_posix(),
        type=str
    )
    args = parser.parse_args()
    return args


class Labels(Enum):
    noise = "noise"
    noisy = "noisy"


async def edge_tts_text_to_speech(text: str, speaker: str = "zh-CN-XiaoxiaoNeural"):
    communicate = edge_tts.Communicate(text, speaker)

    audio_file = temp_dir / f"{uuid.uuid4()}.wav"
    audio_file = audio_file.as_posix()

    await communicate.save(audio_file)
    return audio_file


def generate_spectrogram(filename: str, title: str = "Spectrogram"):
    signal, sample_rate = librosa.load(filename, sr=None)

    mag = np.abs(librosa.stft(signal))
    # mag_db = librosa.amplitude_to_db(mag, ref=np.max)
    mag_db = librosa.amplitude_to_db(mag, ref=20)

    plt.figure(figsize=(10, 4))
    librosa.display.specshow(mag_db, sr=sample_rate)
    plt.title(title)

    buf = BytesIO()
    plt.savefig(buf, format="png", bbox_inches="tight")
    plt.close()
    buf.seek(0)
    return buf


def when_click_annotation_button(filename: Path, label: str, tgt_dir: Path):
    sub_tgt_dir = tgt_dir / label
    sub_tgt_dir.mkdir(parents=True, exist_ok=True)
    shutil.move(filename.as_posix(), sub_tgt_dir)


def main():
    args = get_args()

    src_dir = Path(args.src_dir)
    tgt_dir = Path(args.tgt_dir)

    # @st.cache_data
    # def get_shortcut_audio():
    #     result = {
    #         Labels.noise.value: asyncio.run(edge_tts_text_to_speech("噪音")),
    #         Labels.noisy.value: asyncio.run(edge_tts_text_to_speech("加噪语音")),
    #     }
    #     return result
    #
    # shortcut_audio = get_shortcut_audio()

    # 获取文件列表
    audio_files: List[Path] = [filename for filename in src_dir.glob("**/*.wav")]
    if len(audio_files) == 0:
        st.error("没有未标注的音频了。")
        st.stop()

    audio_file: Path = audio_files[0]

    # session_state
    if "play_audio" not in st.session_state:
        st.session_state.play_audio = False

    # ui
    st.title("🔊 音频文件浏览器")

    column1, column2 = st.columns([4, 1])
    with column1:
        st.audio(audio_file, format=f"{audio_file.suffix}", autoplay=True)

        with st.spinner("生成频谱图中..."):
            spectrogram = generate_spectrogram(audio_file)
        st.image(spectrogram, use_container_width=True)

    with column2:
        shortcut_button(
            label=Labels.noise.value,
            shortcut="1",
            on_click=when_click_annotation_button,
            kwargs={
                "filename": audio_file,
                "label": Labels.noise.value,
                "tgt_dir": tgt_dir,
            },
            type="primary",
        )
        shortcut_button(
            shortcut="2",
            label=Labels.noisy.value,
            on_click=when_click_annotation_button,
            kwargs={
                "filename": audio_file,
                "label": Labels.noisy.value,
                "tgt_dir": tgt_dir,
            },
            type="primary",
        )

    return


if __name__ == "__main__":
    main()