generate-tts.py

#!/usr/bin/env python3
import os
import re
from pathlib import Path
from typing import List

BASE_URL = "https://huggingface.co/csukuangfj/sherpa-onnx-apk/resolve/main/"

from dataclasses import dataclass


@dataclass
class APK:
    major: int
    minor: int
    patch: int
    arch: str
    lang: str
    src: str  # piper, coqui

    def __init__(self, s):
        s = str(s)
        if "tts-" in s:
            s = s[len("tts-") + 1 :]
        else:
            s = s[len("tts") :]
        split = s.split("-")
        self.major, self.minor, self.patch = list(map(int, split[2].split(".")))
        self.arch = split[3]
        self.lang = split[4]
        self.src = split[7]
        if "arm" in s:
            self.arch += "-" + split[4]
            self.lang = split[5]
            self.src = split[8]

        if "armeabi" in self.arch:
            self.arch = "y" + self.arch

        if "arm64" in self.arch:
            self.arch = "z" + self.arch


def sort_by_apk(x):
    x = APK(x)
    return (x.major, x.minor, x.patch, x.arch, -ord(x.src[0]), -ord(x.lang[0]))


def generate_url(files: List[str]) -> List[str]:
    ans = []
    base = BASE_URL
    for f in files:
        ans.append(base + str(f))
    return ans


def get_all_files(d: List[str], suffix: str) -> List[str]:
    files = []
    for k in d:
        files += list(Path(k).glob(suffix))

    ans = sorted(files, key=sort_by_apk, reverse=True)
    return list(map(lambda x: BASE_URL + str(x), ans))


def to_file(filename: str, files: List[str]):
    content = r"""
<h1> APKs for text-to-speech </h1>
This page lists the <strong>text-to-speech</strong> APKs for <a href="http://github.com/k2-fsa/sherpa-onnx">sherpa-onnx</a>,
one of the deployment frameworks of <a href="https://github.com/k2-fsa">the Next-gen Kaldi project</a>.
<br/>
The name of an APK has the following rule:
<ul>
 <li> sherpa-onnx-{version}-{arch}-{lang}-tts-{model}.apk
</ul>
where
<ul>
 <li> version: It specifies the current version, e.g., 1.8.7
 <li> arch: The architecture targeted by this APK, e.g., arm64-v8a, armeabi-v7a, x86_64, x86
 <li> lang: The language supported by this APK, e.g., en for English, zh for Chinese, fr for French, de for German, es for Spanish
 <li> model: The name of the model used in the APK, e.g., vits-ljs, vits-piper-de_DE-thorsten-low, vits-piper-de_DE-thorsten-medium
</ul>

<span style="color:red;">Note:</span> For text-to-speech engine APKs, please see
<a href="https://k2-fsa.github.io/sherpa/onnx/tts/apk-engine.html">https://k2-fsa.github.io/sherpa/onnx/tts/apk-engine.html</a>
<br/><br/>

<span style="color:red;">Note:</span> Models from
<a href="https://github.com/rhasspy/piper">piper</a> have their names prefixed
with <strong>vits-piper-</strong>. For instance, for the model
<strong>vits-piper-en_US-lessac-medium.apk</strong>, its original name
in <a href="https://github.com/rhasspy/piper">piper</a> is
<strong>en_US-lessac-medium.apk</strong>, which is available at
<a href="https://huggingface.co/rhasspy/piper-voices/blob/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx">
https://huggingface.co/rhasspy/piper-voices/blob/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx
</a><br/><br/>

<span style="color:red;">Note:</span> Models from
<a href="https://github.com/MycroftAI/mimic3-voices">MycroftAI/mimic3-voices</a> have their names prefixed
with <strong>mimic3-</strong>.
<br/><br/>

<span style="color:red;">Note:</span> Models from
<a href="https://github.com/coqui-ai/TTS">coqui-ai/TTS</a> have their names prefixed
with <strong>coqui-</strong>.
<br/><br/>

You can find many more models that have not been converted to <strong>sherpa-onnx</strong>
at
<a href="https://huggingface.co/rhasspy/piper-voices">https://huggingface.co/rhasspy/piper-voices</a>

<br/>
<br/>

<strong>Note about the license</strong> The code of Next-gen Kaldi is using
<a href="https://www.apache.org/licenses/LICENSE-2.0">Apache-2.0 license</a>. However,
we support models from different frameworks. Please check the license of your selected model.


<br/>
<br/>
<div/>
    """
    if "-cn" not in filename:
        content += """
        For Chinese users, please <a href="./apk-cn.html">visit this address</a>,
        which replaces <a href="huggingface.co">huggingface.co</a> with <a href="hf-mirror.com">hf-mirror.com</a>
        <br/>
        <br/>
        中国用户, 请访问<a href="./apk-cn.html">这个地址</a>
        <br/>
        <br/>
        """
    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def main():
    apk = get_all_files(["tts", "tts-2"], suffix="*.apk")
    to_file("./apk.html", apk)

    # for Chinese users
    apk2 = []
    for a in apk:
        a = a.replace("huggingface.co", "hf-mirror.com")
        a = a.replace("resolve", "blob")
        apk2.append(a)

    to_file("./apk-cn.html", apk2)


if __name__ == "__main__":
    main()