Skip to content

An open-source package providing standardized tools for sound event analysis and data management.

License

Notifications You must be signed in to change notification settings

mbsantiago/soundevent

Repository files navigation

soundevent

License: MIT PyPI version tests docs Python 3.9 + Static Badge codecov

Warning This package is under active development, use with caution.

soundevent is an open-source Python package that aims to support the computational biocoustic community in developing well-tested, coherent, and standardized software tools for bioacoustic analysis. The main goal of the package is to provide a flexible yet consistent definition of what a sound event is in a computational sense, along with a set of tools to easily work with this definition. The package comprises three key components:

Main features

1. Data Schemas for Bioacoustic Analysis

The soundevent package introduces several data schemas designed to conceptualize and standardize recurring objects in bioacoustic analysis. These data schemas establish relationships between various concepts and define the attributes each object possesses. They provide flexibility to cover a broad spectrum of use cases in bioacoustic analysis while incorporating data validation mechanisms to ensure stored information is both valid and meaningful. Notably, the package defines schemas related to sound events, including user annotations and model predictions.

2. Serialization, Storage, and Reading Functions

To promote standardized data formats for storing annotated datasets and other information about sounds in recordings, the soundevent package provides several functions for serialization, storage, and reading of the different data classes offered. These functions enable easy sharing of information about common objects in bioacoustic research. By employing a consistent data format, researchers can exchange data more efficiently and collaborate seamlessly.

3. Handling Functions for Sound Events

The soundevent package also includes a variety of functions that facilitate the handling of sound event objects. These functions serve multiple purposes, such as matching sound events for model prediction evaluation, transforming sound events, and managing metadata and labels. By offering a comprehensive set of handling functions, the package aims to streamline the analysis workflow for bioacoustic researchers, providing them with powerful tools to manipulate and extract insights from their data.

Installation

You can install soundevent using pip:

pip install soundevent

Documentation

For detailed information on how to use the package, please refer to the documentation.

Example Usage

To see practical examples of how to use soundevent, you can explore the collection of examples provided in the documentation's gallery.

Contributing

We welcome contributions from the community to make soundevent even better. If you would like to contribute, please refer to the contribution guidelines.

License

soundevent is released under the MIT License.