restore option to compress raster#944
Conversation
|
perhaps still some refactoring to clean up the code a bit, but first want opinions on the exposed parameter. |
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #944 +/- ##
==========================================
- Coverage 92.06% 91.96% -0.11%
==========================================
Files 51 51
Lines 7792 7841 +49
==========================================
+ Hits 7174 7211 +37
- Misses 618 630 +12
🚀 New features to boost your workflow:
|
|
Thanks @melonora, I picked this up and merged the conflicts arising after the Zarr v3 support. The tests currently fail mainly due to the disconnect between the new format options and dask, but it should be straightforward to make the PR green again after https://github.com/ome/ome-zarr-py/pull/510/changes is merged. |
|
Sharding support has been added to ome-zarr-py ome/ome-zarr-py#534 and is being added to spatialdata via #1106. |
Use "compressors" (plural) with native zarr v3 codec objects (ZstdCodec, BloscCodec) in storage_options, which ome-zarr-py >= 0.16.0 + dask >= 2026.3.0 correctly forwards to zarr_array_kwargs. Bump minimum deps accordingly and fix test assertions for lz4 (.clevel) vs zstd (.level) and the s0 array path. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
@melonora I merged the latest main and bumped the min Can you please do a review of the changes? If you also think it's ok, feel free to merge (before the sharding PR). |
2 participants
closes #37
This PR adds the capability back of using either
lz4orzstdcompression at various compression levels. It is exposed through acompressorargument inwriteandwrite_elementthat takes as value adictwith key being the compression name and the value being the compression level. By default if compression is specified bytes will be reordered for more efficient compressionSHUFFLEinblosc.