Since the publication of the last paper describing the methods and capabilities of yt, it has been dramatically transformed; while many of the underlying algorithms for processing grid-based data may remain similar or identical, it has been expanded considerably in scope to include data of many different forms.
Furthermore, each of these classes of data -- most notably octree, smoothed particle hydrodynamics, and unstructured meshes -- requires substantial care to ensure that the way that class of data is represented is a high-fidelity reflection of the underlying methods.
In this paper, we have presented our approach to making yt a functional, scientist-driven library to access, process, and visualize data.
This includes the selection of spatial (and non-spatial) regions, converting between spaces for representing data in index coordinates and geometric coordinates, the multi-dimensional reduction of data along spatial and non-spatial dimensions, production of publication-quality plots, volume rendering (both hardware and software) and the approaches we have taken toward developing a community of individuals using and developing yt.
We are grateful for the community of individuals who have participated in yt's development; the authorship of this paper reflects a large, but not complete, fraction of those who have contributed changesets, although invitations have been extended to everyone we were able to reach.
The landscape of astrophysical computation is different now than it was when yt was created, and even since the publication of the first yt method paper.
Code development has become more open, conducted on platforms such as Github, and investment in community around software is largely recognized as necessary, rather than supplemental.
The usage of Python as a library, at least for high-level APIs, has become quite widespread, albeit not quite ubiquitous.
While the community of individuals who participate in yt usage and development is not as large as some in the astrophysical community, or in the broader pydata ecosystem, it is thriving.
The future development of yt will be focused on solidifying our existing community and growing to support the modern needs of scientists from with different approaches to and conceptions of data analysis and visualization.