[Hackathon 7th No.55] Add audiotools to PaddleSpeech

[DrRyanHuang]

PR types

New features

PR changes

APIs

Describe

• 【Hackathon 7th】开源贡献个人挑战赛 Paddle#68244
• RFC: 【Hackathon 7th No.55】在 PaddleSpeech 中实现 audiotools community#1017

在 audio/audiotools/audio_signal.py 中，注释中打✅的均为通过测试的方法，打❌的为未通过 | 未适配

当前进度

• audio_signal
• util
• audiotools.decorators
• audiotools.data
• audiotools.metrics
• unitest
• 移除代码中的✅❌
• 移除文件名字中的✅❌

[paddle-bot]

Thanks for your contribution!

[zxcd]

why there are so many code annotations?

[zxcd]

why this line annotated?

[zxcd]

1. All you test should be added to tests/unit/cli/test_cli.sh to trigger unit test.
2. Also to reduce the repetition of code, it is recommended that you write a separate bash file for voice download in the single test, and after downloading the voice, read it separately from different python files.
3. To ensure accurate alignment, it is recommended to add the current alignment accuracy value in the single measurement annotation, normally fp32 requires 1e-6.
4. Remove symbols and commented out useless code from single test file names.

[zxcd]

开发者你好，感谢你的参与！由于你的黑客松赛题完成度较高，其PR已被锁定，请尽快完善锁定的PR，并确保在2025年1月3日前完成合入。逾期未合入PR将无法获得奖金发放。