Chris Havlin - Introducing yt xarray | SciPy 2023

Introducing yt xarray, an xarray accessor object that streamlines data transfer to yt for geoscience applications, providing a simple and efficient way to integrate yt with xarray and enabling 2D/3D visualization and analysis.

Key takeaways
  • yt xarray is an xarray accessor object that streamlines data transfer from xarray to yt.
  • yt xarray is designed to make it easy to work with yt’s 2D/3D visualization and analysis tool, especially for geoscience-oriented applications.
  • The main goal of yt xarray is to provide a simple and efficient way to integrate yt with xarray, without having to spend time figuring out how to load data into yt.
  • The transition from xarray to yt is relatively straightforward, as long as the data matches the shape of the yt grid.
  • yt xarray supports various data formats, including HDF and netCDF.
  • yt xarray is still under development, with several bugs and caveats to be resolved.
  • The yt xarray API is designed to simplify the process of loading data into yt and making it available for use in yt operations.
  • Some of the benefits of using yt xarray include avoiding memory issues, supporting chunked arrays, and providing dynamic references to xarray fields.
  • yt xarray has several limitations, including limitations on time-series data and the need to pre-compute some functions.
  • The presentation included several examples of using yt xarray, including a 3D volume rendering example.
  • yt xarray is still being developed, with plans to add more features and improve performance.
  • The speaker is hoping to expose more of the yt API within yt xarray to make it easier for users to use yt with their xarray data.