Soundevent

soundevent version 0.4.0 Release Notes:

What's New:

Introducing loading and saving functionalities for sound event objects in version 0.4.0. We now provide six functions, namely "load_dataset", "save_dataset", "load_annotation_projects", "save_annotation_project", "load_model_run", and "save_model_run". These functions enable storing objects in the AOEF (Acoustic Objects Exchange Format), defined using Pydantic models. Refer to `soundevent.io.formats.aoef` for the schema. While unit tests have been implemented to ensure expected behavior, further developments are underway in this area.

Documentation Enhancements:

We have revamped the "Examples" section, renaming it as the "User Guide" and incorporating an improved introduction. Additionally, a new page has been added to the user guide, specifically addressing the utilization of the new loading and saving functions.

What's Coming Next:

In the upcoming release, we have plans to introduce functions for managing sound event objects, with a focus on computing distances between them. These functionalities will lay the foundation for model predictions evaluation functions. Furthermore, we will expand the documentation to include a dedicated page showcasing the available geometries for sound event localization within recordings.

Release Notes:

Key Updates:

* Introducing a gallery of examples in the documentation to demonstrate the usage of soundevent package.
* Added audio loading functions for seamless integration of audio files. Check out the "loading audio" example in the documentation gallery.
* Enhanced capabilities with new functions for computing spectrograms. Explore the "computing spectrograms" example in the documentation gallery.

Upcoming Release Preview:

* The next release will feature serialization and de-serialization functions. This addition will enable importing and exporting soundevent objects in a specific format, facilitating data sharing and preservation.

We are excited to announce the release of `soundevent` v0.1.1, a comprehensive package designed for Bioacoustic research. This initial version focuses on laying the foundation for efficient sound event analysis and management.

Key features of this release include:

* Data Class Definitions: All data class definitions within the `soundevent.data` module are now available. These classes provide a standardized and intuitive representation of various Bioacoustic objects, enabling consistent data handling and analysis.

* Extensive Documentation: We have provided detailed documentation to guide users in understanding the purpose and usage of each defined data class. This documentation aims to explain the rationale behind the design decisions, empowering researchers to leverage the package effectively.

* CI/CD Pipeline and Development Environment: We have established a robust CI/CD pipeline and set up the development environment for smooth integration and collaboration. This ensures the stability and reliability of the package throughout the development process.

Looking ahead, our focus for upcoming releases will be on enhancing the functionality of `soundevent`. Planned additions include:

* Data Loading and Writing: Functions to load and write data, facilitating seamless integration with various data formats and streamlining data management.

* Audio Loading: Capability to load recordings and clips as `numpy` arrays, optimizing memory usage and enabling faster processing of audio data.

* Visualization Tools: Plotting functions to visualize sound events, allowing researchers to gain valuable insights into the acoustic content of recordings.

* Prediction Evaluation: Matching functions to evaluate predictions, enabling researchers to assess the performance of sound event detection models and refine their analysis pipelines.