modusa is a Python library providing utility tools for research.
It offers purpose-built tools especially for the audio domain but not limited to. It combines audio loading, annotation handling, visualization, and analysis in a unified, easy-to-use API. modusa simplifies common workflows so you can focus on experimentation and insight, not boilerplate code.
-
Flexible Audio Loader Load audio in multiple formats (
WAV,MP3,FLAC,M4A,AAC,OPUS,AIFF) — even directly from YouTube links. -
Unified Annotation Interface Work with
.txt(Audacity labels) and.ctm(ASR/FA outputs) annotations seamlessly.TextGridsupport coming soon. -
Modular Plotter Create time-aligned visualizations combining waveforms, annotations, and spectrograms with minimal code.
Supports multi-tier figures, dark mode, legends, tier IDs, and grouped color patterns. -
Interactive Audio Player Play audio with visible annotation labels directly inside notebooks.
-
Built-in Audio Recorder Capture and instantly analyze microphone input from within Jupyter.
-
Analytical Tools Includes quick plotting utilities like distribution (hill) plots for comparing numerical features.
modusa is under active development. You can install the latest version via:
pip install modusa
pytest tests/
modusa is in early alpha. Expect rapid iteration, breaking changes, and big ideas.
If you like the direction, consider ⭐ starring the repo and opening issues or ideas.
To push doc changes
ghp-import -n -p -f docs/build/html
To create a dist
pdm build
To upload on pypi
twine upload dist/*
modusa is developed and maintained by meluron,
MIT License. See LICENSE for details.
Pull requests, ideas, and discussions are welcome!
No matter which domain you are in, if you work with any signal, we'd love your input.