app.py

#!/usr/bin/env python3

import os
import re
import time
import uuid
from datetime import datetime
from pathlib import Path

import gradio as gr
import sherpa_onnx
import soundfile as sf

from model import get_pretrained_model, get_speaker_map, language_to_models


def MyPrint(s):
    now = datetime.now()
    date_time = now.strftime("%Y-%m-%d %H:%M:%S.%f")
    print(f"{date_time}: {s}")


def get_num_speakers(repo_id: str) -> int:
    m = re.search(r"\|(\d+)\s*speaker", repo_id)
    if m:
        return int(m.group(1))
    return 1


def get_speaker_choices(repo_id: str) -> list:
    n = get_num_speakers(repo_id)
    speaker_map = get_speaker_map(repo_id)
    if speaker_map:
        return [f"{v}" for _, v in sorted(speaker_map.items())]
    return [f"说话人 {i}" for i in range(n)]


def extract_sid_from_label(repo_id: str, label: str) -> int:
    speaker_map = get_speaker_map(repo_id)
    if speaker_map:
        for sid, name in speaker_map.items():
            if name == label:
                return sid
    m = re.search(r"说话人\s*(\d+)", label)
    if m:
        return int(m.group(1))
    return 0


title = "# 文字转语音 (TTS)"

css = """
.result {display:flex;flex-direction:column}
.result_item {padding:15px;margin-bottom:8px;border-radius:15px;width:100%}
.result_item_success {background-color:mediumaquamarine;color:white;align-self:start}
.result_item_error {background-color:#ff7070;color:white;align-self:start}
"""


def update_model_dropdown(language: str):
    if language in language_to_models:
        choices = language_to_models[language]
        speaker_choices = get_speaker_choices(choices[0])
        return gr.Dropdown(
            choices=choices,
            value=choices[0],
            interactive=True,
        ), gr.Dropdown(
            choices=speaker_choices,
            value=speaker_choices[0],
            visible=len(speaker_choices) > 1,
            interactive=True,
        )
    raise ValueError(f"不支持的语言: {language}")


def update_speaker_dropdown(repo_id: str):
    speaker_choices = get_speaker_choices(repo_id)
    return gr.Dropdown(
        choices=speaker_choices,
        value=speaker_choices[0],
        visible=len(speaker_choices) > 1,
        interactive=True,
    )


def build_html_output(s: str, style: str = "result_item_success"):
    return f"""
<div class='result'>
  <div class='result_item {style}'>
    {s}
  </div>
</div>
"""


def process(language: str, repo_id: str, text: str, speaker: str, speed: float):
    max_len = 4000
    sid = extract_sid_from_label(repo_id, speaker)
    MyPrint(f"输入文本长度 {len(text)}: {text[:max_len]}. 说话人: {speaker}(id={sid}), 语速: {speed}")

    if len(text) > max_len:
        MyPrint(f"文本过长！{len(text)}")
        info = "为保证响应速度，请使用短文本进行测试。如需处理长文本，请在本地运行。"
        return None, build_html_output(info)

    n = get_num_speakers(repo_id)
    if n > 1 and sid >= n:
        sid = n - 1

    tts = get_pretrained_model(repo_id, speed)

    start = time.time()
    audio = tts.generate(text, sid=sid)
    end = time.time()

    if len(audio.samples) == 0:
        raise ValueError("语音生成出错，请查看上方错误信息。")

    duration = len(audio.samples) / audio.sample_rate
    elapsed_seconds = end - start
    rtf = elapsed_seconds / duration

    info = f"""
音频时长: {duration:.3f} 秒<br/>
处理时间: {elapsed_seconds:.3f} 秒<br/>
实时率(RTF): {rtf:.3f}<br/>
说话人: {speaker}
"""

    MyPrint(info)
    MyPrint(f"\nrepo_id: {repo_id}\ntext: {text}\nsid: {sid}\nspeed: {speed}")

    filename = str(uuid.uuid4())
    filename = f"{filename}.wav"
    sf.write(
        filename,
        audio.samples,
        samplerate=audio.sample_rate,
        subtype="PCM_16",
    )
    return filename, build_html_output(info)


demo = gr.Blocks(css=css)

with demo:
    gr.Markdown(title)
    language_choices = list(language_to_models.keys())
    language_radio = gr.Radio(
        label="语言",
        choices=language_choices,
        value=language_choices[0],
    )
    model_dropdown = gr.Dropdown(
        choices=language_to_models[language_choices[0]],
        label="选择模型",
        value=language_to_models[language_choices[0]][0],
    )
    first_model = language_to_models[language_choices[0]][0]
    first_speakers = get_speaker_choices(first_model)
    speaker_dropdown = gr.Dropdown(
        choices=first_speakers,
        value=first_speakers[0],
        label="选择说话人",
        visible=len(first_speakers) > 1,
        interactive=True,
        allow_custom_value=True,
    )
    language_radio.change(
        update_model_dropdown,
        inputs=language_radio,
        outputs=[model_dropdown, speaker_dropdown],
    )
    model_dropdown.change(
        update_speaker_dropdown,
        inputs=model_dropdown,
        outputs=speaker_dropdown,
    )
    with gr.Tabs():
        with gr.TabItem("输入文本"):
            input_text = gr.Textbox(
                label="输入文本",
                info="请输入要转换为语音的文字",
                lines=3,
                value="大家好，这是一个文字转语音的测试。",
                placeholder="请输入要转换为语音的文字",
            )
            input_speed = gr.Slider(
                minimum=0.1,
                maximum=10,
                value=1,
                step=0.1,
                label="语速（越大越快，越小越慢）",
            )
            input_button = gr.Button("生成语音")
            output_audio = gr.Audio(label="生成的语音")
            output_info = gr.HTML(label="信息")

    input_button.click(
        process,
        inputs=[
            language_radio,
            model_dropdown,
            input_text,
            speaker_dropdown,
            input_speed,
        ],
        outputs=[
            output_audio,
            output_info,
        ],
    )


def download_espeak_ng_data():
    os.system(
        """
cd /tmp
wget -qq https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/espeak-ng-data.tar.bz2
tar xf espeak-ng-data.tar.bz2
"""
    )
    if not Path("/tmp/dict").is_dir():
        os.system(
            "cd /tmp; curl -SL -O https://github.com/csukuangfj/cppjieba/releases/download/sherpa-onnx-2024-04-19/dict.tar.bz2; tar xvf dict.tar.bz2"
        )
    os.system("ls -lh /tmp/dict")


if __name__ == "__main__":
    download_espeak_ng_data()
    demo.launch()