ODD PI 25.4 Objective 3: 🗺️Research, develop and document methods for Zarr and VirtualiZarr data services #203
Description
Motivation
The ODD team has demonstrated how we can tile multi-dimensional datasets using a dynamic tiling API based on the proven titiler library. However, this library was developed for use with Cloud-Optimized GeoTIFFs (COGs) and as such may not be the most optimal choice for Zarr. Further, we believe there may be direct browser loading options which would eliminate the need for a server altogether.
Description
Evaluate alternative methods for producing visualizations for Zarr and VirtualiZarr datasets through discussion, research and prototyping.
Acceptance Criteria
- Prototyping and comparison of xpublish, further development of web-optimized Zarr (WOZ)
- Document recommendations on architecture for future Zarr data services and where titiler, xpublish, and other APIs fit in.
- Integration of icechunk reading into titiler.xarray (perhaps prototyping in titiler-multidim)
- Stretch:
- Evaluate options for supporting API-based time series generation for zarr datasets (consider testing the EDR plugin for xpublish, for example, or OGC coverages) and discuss with potential clients, such as the VEDA UI team.
- Deliver a proposal (perhaps using the ADR format) for supporting time series