OSGeo
zarr with lerc compression
Feature description
Some time ago, we started to use large zarr datasets with zstd+lerc compression. We create and read these now using python (with rioxarray, where we can manually insert a custom lerc compressor). It would be a tremendous advantage to have direct gdal access (read + write) to these datasets. Would it be possible to add (lossy, with custom precision) lerc compression to gdal's zarr capabilities?
I can provide access to a simple example of such a dataset (on gcs, prefereably) if necessary.
Thanks, Vincent (Satelligence).
Additional context
No response
We create and read these now using python (with rioxarray, where we can manually insert a custom lerc compressor).
I'd prefer this to go first into imagecodecs first (preferably, or at least some public and advertized extension of it) so there is interoperability with the Python Zarr ecosystem and not be a GDAL-specific solution
Example would be definitely needed
Does this mean that something is going on? https://github.com/cgohlke/imagecodecs/blob/master/imagecodecs/_lerc.pyx Link found from https://zarr.dev/codecs-registry/
