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-flutter/resolve/main/"

from dataclasses import dataclass


@dataclass
class APP:
    major: int
    minor: int
    patch: int
    arch: str
    name1: str
    name2: str

    def __init__(self, _s):
        # android
        # sherpa-onnx-1.10.11-x86_64-zh-tts-vits-zh-hf-fanchen-wnj.apk
        # sherpa-onnx-1.10.11-armeabi-v7a-zh-tts-vits-zh-hf-keqing.apk
        # sherpa-onnx-1.10.11-x86_64-zh-tts-vits-zh-hf-fanchen-ZhiHuiLaoZhe_new.apk

        # macos
        # sherpa-onnx-1.10.11-osx-arm64-bn-tts-vits-mimic3-bn-multi_low.app.tar.bz2
        # sherpa-onnx-1.10.11-osx-arm64-zh-tts-vits-zh-hf-theresa.app.tar.bz2

        # Linux
        # sherpa-onnx-1.10.11-linux-x86_64-zh-tts-vits-zh-hf-zenyatta.tar.bz2

        # Windows
        # sherpa-onnx-1.10.11-win-x64-zh-tts-vits-zh-hf-zenyatta.tar.bz2

        s = _s.split("/")[-1]
        split = s.split("-")
        self.major, self.minor, self.patch = list(map(int, split[2].split(".")))
        if "android" in _s:
            self.os = "android"
            self.arch = split[3]
            self.lang = split[4]
            self.name1 = split[-1]

            if "arm" in _s:
                self.arch = split[3] + "-" + split[4]
                self.lang = split[5]
        else:
            self.os = split[3]
            self.arch = split[4]
            self.lang = split[5]
            self.name1 = split[-1]


def sort_by_app(x):
    x = APP(x)
    return (x.major, x.minor, x.patch, x.os, x.arch, x.lang, x.name1)


def get_all_files(d: str, suffix: str) -> List[str]:
    ss = []
    for root, d, files in os.walk(d):
        for f in files:
            if f.endswith(".tar.bz2") or f.endswith(".apk"):
                ss.append(os.path.join(root, f))

    return list(map(lambda x: BASE_URL + x, ss))


def to_file_android(filename: str, files: List[str]):
    content = r"""
<h1> Flutter Android APKs for text-to-speech </h1>

This page lists the <strong>text-to-speech</strong> APKs created with <a href="https://flutter.dev/">Flutter</a>
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., 1.10.11
 <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>

<br/>

<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:
        print("filename", filename)
        print(filename.split("/")[-1].split("."))
        cn_filename = filename.split("/")[-1].split(".")[0] + "-cn.html"
        content += f"""
        For Chinese users, please <a href="./{cn_filename}">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="./{cn_filename}">这个地址</a>
        <br/>
        <br/>
        """

    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            x = x.replace("huggingface.co", "hf-mirror.com")
            x = x.replace("resolve", "blob")
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def to_file_linux(filename: str, files: List[str]):
    content = r"""
<h1> Flutter Linux APPs for text-to-speech </h1>

This page lists the <strong>text-to-speech</strong> Linux APPs created with <a href="https://flutter.dev/">Flutter</a>
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 APP has the following rule:
<ul>
 <li> sherpa-onnx-{version}-linux-{arch}-{lang}-tts-{model}.tar.bz2
</ul>
where
<ul>
 <li> version: It specifies the current version, e.g., 1.10.11
 <li> arch: Only x86_64 is supported for Linux at present.
 <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>

<br/>

<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:
        print("filename", filename)
        print(filename.split("/")[-1].split("."))
        cn_filename = filename.split("/")[-1].split(".")[0] + "-cn.html"
        content += f"""
        For Chinese users, please <a href="./{cn_filename}">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="./{cn_filename}">这个地址</a>
        <br/>
        <br/>
        """

    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            x = x.replace("huggingface.co", "hf-mirror.com")
            x = x.replace("resolve", "blob")
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def to_file_macos_x64(filename: str, files: List[str]):
    content = r"""
<h1> Flutter macoOS x86_64 APPs for text-to-speech </h1>

This page lists the <strong>text-to-speech</strong> macOS x86_64 APPs created with <a href="https://flutter.dev/">Flutter</a>
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 APP has the following rule:
<ul>
 <li> sherpa-onnx-{version}-osx-{arch}-{lang}-tts-{model}.tar.bz2
</ul>
where
<ul>
 <li> version: It specifies the current version, e.g., 1.10.11
 <li> arch: This page is for x86_64. If you need arm64, please see <a href="./tts-macos-arm64.html">./tts-macos-arm64.html</a>
 <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>

<br/>

<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:
        print("filename", filename)
        print(filename.split("/")[-1].split("."))
        cn_filename = filename.split("/")[-1].split(".")[0] + "-cn.html"
        content += f"""
        For Chinese users, please <a href="./{cn_filename}">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="./{cn_filename}">这个地址</a>
        <br/>
        <br/>
        """

    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            x = x.replace("huggingface.co", "hf-mirror.com")
            x = x.replace("resolve", "blob")
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def to_file_macos_arm64(filename: str, files: List[str]):
    content = r"""
<h1> Flutter macoOS arm64 APPs for text-to-speech </h1>

This page lists the <strong>text-to-speech</strong> macOS arm64 APPs created with <a href="https://flutter.dev/">Flutter</a>
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 APP has the following rule:
<ul>
 <li> sherpa-onnx-{version}-osx-{arch}-{lang}-tts-{model}.tar.bz2
</ul>
where
<ul>
 <li> version: It specifies the current version, e.g., 1.10.11
 <li> arch: This page is for arm64. If you need x86_64, please see <a href="./tts-macos-x64.html">./tts-macos-x64.html</a>
 <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>

<br/>

<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:
        print("filename", filename)
        print(filename.split("/")[-1].split("."))
        cn_filename = filename.split("/")[-1].split(".")[0] + "-cn.html"
        content += f"""
        For Chinese users, please <a href="./{cn_filename}">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="./{cn_filename}">这个地址</a>
        <br/>
        <br/>
        """

    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            x = x.replace("huggingface.co", "hf-mirror.com")
            x = x.replace("resolve", "blob")
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def to_file_windows(filename: str, files: List[str]):
    content = r"""
<h1> Flutter Windows APPs for text-to-speech </h1>

This page lists the <strong>text-to-speech</strong> Windows APPs created with <a href="https://flutter.dev/">Flutter</a>
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 APP has the following rule:
<ul>
 <li> sherpa-onnx-{version}-win-{arch}-{lang}-tts-{model}.tar.bz2
</ul>
where
<ul>
 <li> version: It specifies the current version, e.g., 1.10.11
 <li> arch: Only x64 is supported for Windows at present.
 <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>

<br/>

<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:
        print("filename", filename)
        print(filename.split("/")[-1].split("."))
        cn_filename = filename.split("/")[-1].split(".")[0] + "-cn.html"
        content += f"""
        For Chinese users, please <a href="./{cn_filename}">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="./{cn_filename}">这个地址</a>
        <br/>
        <br/>
        """

    with open(filename, "w") as f:
        print(content, file=f)
        for x in files:
            x = x.replace("huggingface.co", "hf-mirror.com")
            x = x.replace("resolve", "blob")
            name = x.rsplit("/", maxsplit=1)[-1]
            print(f'<a href="{x}" />{name}<br/>', file=f)


def get_platform(allFiles: List[str], pattern: str):
    ans = []
    for f in allFiles:
        if pattern in f:
            ans.append(f)
    return ans


def main():
    app = get_all_files("flutter/tts", suffix="*.tar.bz2")

    macos = get_platform(app, "macos")
    macos_x64 = get_platform(macos, "x86_64")
    macos_arm64 = get_platform(macos, "arm64")

    windows = get_platform(app, "windows")
    linux = get_platform(app, "linux")
    android = get_platform(app, "android")

    android = sorted(android, key=sort_by_app, reverse=True)
    to_file_android("./tts-android.html", android)
    to_file_android("./tts-android-cn.html", android)

    linux = sorted(linux, key=sort_by_app, reverse=True)
    to_file_linux("./tts-linux.html", linux)
    to_file_linux("./tts-linux-cn.html", linux)

    macos_x64 = sorted(macos_x64, key=sort_by_app, reverse=True)
    to_file_macos_x64("./tts-macos-x64.html", macos_x64)
    to_file_macos_x64("./tts-macos-x64-cn.html", macos_x64)

    macos_arm64 = sorted(macos_arm64, key=sort_by_app, reverse=True)
    to_file_macos_arm64("./tts-macos-arm64.html", macos_arm64)
    to_file_macos_arm64("./tts-macos-arm64-cn.html", macos_arm64)

    windows = sorted(windows, key=sort_by_app, reverse=True)
    to_file_windows("./tts-win.html", windows)
    to_file_windows("./tts-win-cn.html", windows)


if __name__ == "__main__":
    main()