- Drop support for Python 3.8.
- Document parameters and return value of
parse_url
.
- Write metadata using delayed that depends on writing array(s)
- Document return values of sample data
- Simplify Python tutorial
- pin zarr at < 3
- Update version of pyupgrade
- Correctly specify maximum compatible fsspec version. (#338)
- Add tests on Python 3.12. (#338)
- Write OMERO metadata. (#261)
- Fixed chunking when a scalar value for chunks is given. Previously
passing
storage_options={"chunks": <int>}
only set the chunk size of the final dimension. Now the chunk size of all dimensions is set to this value, which is identical behaviour tozarr-python
. (#365)
- Fix reading HCS file on AWS S3 (#322)
- Update docs to include write example for HCS dataset (#317)
- Minor improvements to format docs (#316)
- Add dask to intersphinx mapping (#315)
- Fix compute on plate grid (#299)
- Check Well exists before trying to load data (#296)
- Avoid hard-coded field-of-view path "0". Thanks to aeisenbarth (#300)
- Update python.rst to have min, max for window. Thanks to Sean Martin (#309)
- Exclude fsspec 2023.9.0 fixes FileNotFoundError (#307)
- Set minimum fsspec version. Thanks to Evan Lyall (#295)
- Add conda install instructions to README. Thanks to David Stansby (#313)
- Cody tidy fixes. Thanks to Dimitri Papadopoulos Orfanos (#310, #312)
- Add
ome_zarr view image.zarr
to server and view in browser (#285) - Allow
main
with empty args: thanks to Dominik Kutra (#279)
- Optimize zarr multiscale writing by permitting delayed execution: thanks to Camilo Laiton (#257)
- Use default compressor by default; thanks to Juan Nunez-Iglesias (#253)
- Fix bug in dask resizing of edge chunks (#244)
- Fix dimension separator used in dask array to_zarr; thanks to Luca Marconato (#243)
- Add
compute
option to image writing methods; thanks to Camilo Laiton (#257) - Allow errors to be raised when reading Multiscales data (#266)
- Add a pre-commit configuration for codespell; thanks to Yaroslav Halchenko (#272)
- Fix pyramid dtype for scaler.local_mean (#233)
- Add support for writing
dask
arrays. Thanks to Daniel Toloudis, with code contributions from Andreas Eisenbarth (#192) - Fix a bug in using
storage_options
withwrite_images
, thanks to Marc Claesen (#221) - Use multiscales name for node metadata (#214)
- Disable auto_mkdir when path begins with s3, thanks to @colobas (#212)
- Read multi-resolution pyramids in
ome_zarr.reader.Well
, thanks to @tcompa (#209)
- Add
fmt
option towrite_label_metadata
and store version (#206)
- Relax version detection (#189)
- Fix format warning duplication (#190)
- Fix plate pyramid loading (#195)
- Unify docstrings in
ome_zarr.writer
module (#185)
- Add API for writing data according to the
image-label
spec:write_labels
,write_multiscale_labels
&write_label_metadata
(#178, #184) - Remove
byte_order
fromwrite_image
signature (#184) - Deprecate
chunks
argument in favor ofstorage_options
(#184)
- Add support for OME-NGFF 0.4 specification (#124, #159, #162)
- Add support for passing storage options (compression, chunks..) to the writer API, thanks to @satra (#161)
- Add API for writing plate & well metadata (#153, #157)
- Add API for writing multiscales metadata (#149, #165)
- Add support for writing and scaling 2D, 3D and 4D data (#114)
- Add dispatch GitHub workflow (#118)