Skip to content

Changelog

All notable changes to the webknossos python library are documented in this file.

Please see the Stability Policy for details about the version schema and compatibility implications.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning MAJOR.MINOR.PATCH. For upgrade instructions, please check the respective Breaking Changes sections.

Unreleased

Commits

Breaking Changes

Added

Changed

Fixed

3.0.1 - 2025-10-30

Commits

Fixed

  • Fixed a bug where some tiff images could not be correctly converted to wkw/zarr, yielding KeyErrors. #1388

3.0.0 - 2025-10-28

Commits

Breaking Changes

  • The endpoint_url in UPath objects is now stored directly in the storage_options dict, instead of being stored in the storage_options["client_kwargs"] dict. #1365
  • Removed Docker image scalableminds/webknossos-cli build. It will not be released further. #1376
  • Removed inference of channels in case numChannels is not set in the datasource-properties.json. numChannels == 1 is assumed and fails upon reading, if otherwise. #1386

Added

  • Dataset.add_layer_as_ref() now accepts RemoteLayer objects as well as Layer objects. #1371
  • RemoteDataset.annotation_id is now exposed, when available. #1380
  • RemoteDataset.open() now accepts an annotation url as well. #1380
  • RemoteAnnotation, now has a get_agglomerate_graph method, to make working with proofreading annotations easier. #1361
  • Added RemoteDataset.used_storage_bytes property. #1386
  • Added read support for N5 and neuroglancer precomputed data formats. #1386

Changed

  • Ported test.sh to test.py, run with uv run test.py. #1379
  • Dataset.add_layer_as_ref(remote_layer.path) does not work anymore. Please use Dataset.add_layer_as_ref(remote_layer) instead. #1371
  • Due to the refactoring, the imports of various classes have changed. Please update your code accordingly. #1371
  • Deprecated a number of methods #1371:
  • Dataset.open_remote(), use RemoteDataset.open instead.
  • RemoteDatset.download_mesh(), use remote_segmentation_layer.download_mesh() instead.
  • Dataset.get_remote_datasets(), use RemoteDataset.list() instead.
  • Dataset.download(), use RemoteDataset.download() instead.
  • mag_view.is_foreign, use layer.is_mag_view_foreign(mag_view) instead.
  • Dataset.trigger_reload_in_datastore(), use RemoteDataset.trigger_reload_in_datastore() instead.
  • Replaced pathlib.Path for upath.UPath internally. pathlib.Path is still supported for user-facing APIs. Adds support for universal_pathlib version 0.3.x. #1374
  • Only use fs-based copy for mags on local file systems and S3. Other protocols, e.g. memory and HTTP, have caveats that break fs-based copying. #1365
  • The add_* methods in Attachments now return the created attachment objects, similar to add_layer and add_mag. #1365
  • The returned dataset ID from dataset exploration is now used. #1378
  • Refactored the architecture, by introducing RemoteLayers, RemoteSegmentationLayers and their abstract base classes. #1371
  • Updated the api version of the webknossos-api to 12. #1371
  • Allowing RemoteDataset to align mags, when down- or upsampling #1382
  • RemoteDatasets that use zarr streaming are no longer read-only. #1383
  • Introduced transfer modes COPY, MOVE_AND_SYMLINK and HTTP for Dataset.upload() #1384

Fixed

  • Fixed test.py to parse the command line arguments correctly. #1383
  • Fixed a bug in the CLI tool export_as_tiff, where filenames would not match the data z position exactly. #1387

2.5.0 - 2025-10-06

Commits

Added

  • Added context manager VolumeLayer.edit for creating and modifying volume annotations. #1340
  • Added overwrite_existing flag to convert, convert-raw and convert-zarr CLI commands. Use with caution. #1372
  • Added downsample, max-mag, interpolation-mode, sampling-mode args to convert, convert-raw and convert-zarr CLI commands, where it was missing, for consistency. #1372
  • Added value rescaling to convert-raw CLI command through source-dtype and rescale-min-max args. #1372
  • Added interpolation_mode and compress kwargs to Dataset.downsample method. #1372
  • Added Team.get_by_id, Team.add_user and Team.delete methods. #1373
  • Added RemoteFolder.get_root, RemoteFolder.get_subfolders, RemoteFolder.get_datasets, RemoteFolder.add_subfolder, RemoteFolder.move_to, RemoteFolder.delete methods and RemoteFolder.allowed_teams, RemoteFolder.name properties. #1373

Changed

  • Team.add now returns the created team object. #1373
  • Moved Team to webknossos.administration.team module. #1373

2.4.12 - 2025-09-17

Commits

Changed

  • Disabled conditional writes for S3 buckets in tensorstore. #1368

Fixed

  • Fixed that some operations that need an authentication token didn’t ask for it but just failed. #1366

2.4.11 - 2025-09-02

Commits

Breaking Changes

  • All internal API client methods (webknossos.client) use keyword-only arguments. #1363
  • The signature of Dataset.trigger_reload_on_datastore changed. dataset_name has been renamed to dataset_name_or_url. Alternatively, a dataset_id can be supplied. #1363

Changed

  • Layer.add_mag_as_copy now automatically uses file-based copy, if possible. Therefore, Dataset.fs_copy_dataset, Dataset.add_fs_copy_layer and Layer.add_fs_copy_mag are deprecated. #1362
  • The API methods now use v10 API version of Webknossos. Datasets are now referenced by ID instead of dataset name or directory name. #1363

2.4.10 - 2025-08-25

Commits

Added

  • Added support for remote datasets and annotations to webknossos export-as-tiff. #1356

2.4.9 - 2025-08-11

Commits

Fixed

  • Fixed a bug where paths of references mags were wrong, if the target dataset's path was a prefix of the source dataset's path. #1346

2.4.8 - 2025-08-06

Commits

Added

  • Added support for reading wkw datasets with the new "mag" key in datasource-properties.json, which replaces "wkwResolutions" for datasets written out by the webknossos server. #1349

2.4.7 - 2025-07-30

Commits

Added

  • Added MagView.rechunk methods to allow for rechunking of datasets. #1342
  • Added the option to configure the codecs of Zarr3 datasets. Supply a Zarr3Config to the compress argument in Layer.add_mag or similar methods. codecs and chunk_key_encoding can be customized. #1343
  • Added method download_mesh to the RemoteDataset and RemoteAnnotation classes to allow download of .stl files. #1307

Changed

  • Enforces that chunk_shape and shard_shape have power-of-two values. This assumptions was used in the code previously, but not explicitly enforced. #1342
  • The sharding_indexed codec is no longer used if chunk_shape == shard_shape. #1343

2.4.6 - 2025-07-28

Commits

2.4.5 - 2025-07-16

Commits

Fixed

  • Fixed an issue where the unit of the scale (voxel_size) was not considered when reading the NML. #1339

2.4.4 - 2025-07-14

Commits

Added

  • The function Dataset.announce_manual_upload now takes an additional optional parameter dataset_url to select the correct datastore in a multi-datastore setup. #1336

Changed

  • Every file that sends logging messages has its own logger now. This can be used to selectively disable loggers. #1335
  • Enabled compression by default for convert-zarr and convert-raw CLI commands. #1338
  • Add more (network) errors ("UNAVAILABLE", "Unknown error") to list of retryable errors. #1341

2.4.3 - 2025-07-07

Commits

Changed

  • Warnings regarding missing pims image readers are only printed once during conversion. Error messages regarding wk-libs version are not displayed in slurm environments. #1266

Fixed

  • Fixed mag and attachment paths under Windows. #1329

2.4.2 - 2025-07-07

Commits

Added

  • Added operations for Project, Task and TaskType classes and objects: #1294
  • Added operations for Project class:
    • Project.create() for creating a new project.
    • my_project.delete() to delete a project object
  • Added my_project.update() to update priority, expected_time and is_blacklisted_from_report values of an existing project object.
  • Added operations for TaskType class:
    • TaskType.get_list() to get all TaskType objects.
    • TaskType.get_by_id() to get a TaskType by id.
    • TaskType.get_by_name() to get a TaskType by name.
    • TaskType.create() to create a new TaskType.
  • Added operations my_task_type.delete() method for TaskType objects.
  • Added Task.get_list()to get all Task objects.
  • Added operations for Task objects:
    • my_task.update() to update a Task object.
    • my_task.delete() to delete a Task object.
  • Added TaskExperience class to store domain and value pair.
  • The libs client now uses versioned datastore routes for improved compatibility across webknossos versions. #1328

Fixed

  • Fixed the tests for the newest WEBKNOSSOS version. #1328

2.4.1 - 2025-07-03

Commits

Changed

  • Limit custom retries (e.g. additional to tensorstore internal retries) to certain network errors ("Too Many Requests", "GatewayTimeout"). #1330

2.4.0 - 2025-06-24

Commits

Breaking Changes

  • Dataset.shallow_copy_dataset does not symlink layers, mags and attachments anymore. Instead mags and attachments are referenced by their path. #1290
  • Deprecated a number of methods #1290:
  • Dataset.add_symlink_layer, use Dataset.add_layer_as_ref instead.
  • Dataset.add_remote_layer, use Dataset.add_layer_as_ref instead (renamed).
  • Dataset.add_copy_layer, use Dataset.add_layer_as_copy instead (renamed).
  • Layer.add_symlink_mag, use Layer.add_mag_as_ref instead.
  • Layer.add_remote_mag, use Layer.add_mag_as_ref instead (renamed).
  • Layer.add_copy_mag, use Layer.add_mag_as_copy instead (renamed).
  • Attachments.add_symlink_attachments, use Attachments.add_attachment_as_ref instead (note the singular form).
  • Attachments.add_attachments, use Attachments.add_attachment_as_ref instead (note the singular form).
  • Attachments.add_copy_attachments, use Attachments.add_attachment_as_copy instead (note the singular form).
  • Dataset.add_layer and Dataset.add_layer_like as well as derived methods do not eagerly create a folder for the layer anymore. Folders will be created, if necessary, when mags or attachments are added. #1290

Changed

  • Changed the default path type to UPath #1326
  • All mags now store their path in the datasource-properties.json file. This eliminates the need for symlinks and removes the need to check multiple possible mag paths (e.g. 1-1-1 or 1). It also unifies local and remote mags. #1290
  • Dataset-local mags and attachments use the ./ prefix for their path in compliance with IETF RFC1808. #1290

2.3.11 - 2025-06-18

Commits

Fixed

  • Fixed excessive RAM consumption during downsampling by chunking the computation. #1325

2.3.10 - 2025-06-17

Commits

Added

  • Added layers_to_ignore argument to Dataset.copy_dataset. #1321

Fixed

  • Make number of retries and backoff factor configurable (mainly for tensorstore reads/writes). See DEFAULT_NUM_RETRIES and DEFAULT_BACKOFF_FACTOR environment variables. #1323

2.3.9 - 2025-06-13

Commits

Added

  • Added a feature to track attached files of a segmentation layer. Previously, these files were only auto-detected by WEBKNOSSOS based on the location in special folders. Attachments can be added, e.g. dataset.get_segmentation_layer("segmentation").attachments.add_mesh('path/to/meshfile.hdf5'). #1312
  • Added a with_attachments parameter to Dataset.copy_dataset to copy attachments. #1312
  • Added a get_segmentation_layer method to Dataset to get a correctly-typed segmentation layer by name. #1312
  • Retry asynchronous tensorstore calls on failure to make operation on object storage datasets accessed via network more robust. #1319

Changed

  • When adding a layer, the used dtype is only valid if it is supported by webknossos. #1316

Fixed

  • Fixed an issue that creates an Error when add_mag() was called just with a chunk_shape. #1315

2.3.8 - 2025-05-28

Commits

Changed

  • Made the folder_id parameter of Dataset.announce_manual_upload optional. #1313

2.3.7 - 2025-05-27

Commits

Changed

  • Pin tifffile version to 2025.5.21 or below. #1314

2.3.6 - 2025-05-20

Commits

Added

  • Add support for in-memory zarr3 datasets (using the Tensorstore in-memory KvStore driver). #1310

Changed

  • Added url parsing in webknossos download CLI command, to get webknossos url without environment variable. #1299

2.3.5 - 2025-05-13

Commits

Added

  • Added --downsample flag to webknossos convert CLI command. #1305

Changed

  • Uses compression by default for webknossos convert CLI command. #1305

Fixed

  • Pinned click<8.2.0 because of incompatibility with typer. #1305

2.3.4 - 2025-05-12

Commits

Fixed

  • Fixed deletion of symlinked mags. #1300

2.3.3 - 2025-05-06

Commits

Fixed

  • Fixed an issue with Dataset.get_remote_datasets() that causes value lookups to fail. #1286
  • Fixed issue with adding remote mags to local dataset due to some UPath checks failing for paths from different file systems. #1298

2.3.2 - 2025-04-25

Commits

Fixed

For S3 mag paths, do not dump endpoint_url protocol ("https") in the mag path when writing to dataset_properties.json. Correctly parse S3 paths for remote mags when reading from dataset_properties.json.#1293

2.3.1 - 2025-04-24

Commits

Added

  • Added argument require_unique_name to Dataset.announce_manual_upload(). Additionally, the method returns a tuple of new_dataset_id and directory_name now. #1283

Changed

  • Upgraded tensorstore to 0.1.74. #1292

2.3.0 - 2025-04-22

Commits

Added

  • Added a version field to the datasource-properties.json file, which is always set to 1, for now. #1288

Changed

  • The Layer.path property now always returns Dataset.path / layer.name, even if all mags are stored remotely. #1288
  • Layer.is_remote_to_dataset is deprecated in favor of Layer.is_foreign and MagView.is_remote_to_dataset is deprecated in favor of MagView.is_foreign. #1288
  • Relative path fields in the mag definition of a datasource-properties.json file are now interpreted relative to the Dataset.path. #1288

2.2.0 - 2025-04-17

Commits

Breaking Changes

  • If buffer_shape is passed to Layer.upsample() it must be a multiple of the shard_shape. #1287

Changed

  • Using x-auth-header to send tokens to datastore. #1270
  • Add clarification that bounding box size is in units of voxels #1289

Fixed

  • Fixed an issue with upsampling views. #1287

2.1.0 - 2025-04-01

Commits

Breaking Changes

  • Dropped support for Python 3.9. #1277

Added

  • Adds the field totalFileSizeInBytes to the dataset upload information. #1264
  • Added support for metadata in segment and trees. #1271
  • Added a warning on import if a newer version of the webknossos package is available from PyPI. #1280

Changed

  • Upgraded tensorstore to 0.1.72. #1277

2.0.7 - 2025-03-26

Commits

Changed

  • Pinned tensorstore to <=0.1.71. #1276

2.0.6 - 2025-03-25

Commits

Fixed

  • Fixed an issue with uploading datasets where the dataset name is already in use. #1275

2.0.5 - 2025-03-25

Commits

Changed

  • Changed the default dimension separator for Zarr 2 arrays from . to /. #1274

2.0.4 - 2025-03-19

Commits

Changed

  • For Zarr3 datasets, the OME-Zarr 0.5 metadata is now written. #1272

Fixed

  • Fixed an issue where it was attempted to overwrite the OME-Zarr metadata for symlinked layers. #1272

2.0.3 - 2025-03-18

Commits

2.0.2 - 2025-03-13

Commits

Added

  • Added tests for NDBoundingBox and VecInt classes. #1268

Changed

  • Removed PimsImagejTiffReader in favor of the unified PimsTiffReader. #1269

Fixed

  • Fixed issues with tiff conversion for tiff files that don't use pages for z and other axes. #1269

2.0.1 - 2025-03-11

Commits

Added

  • Added get_remote_annotations() to AnnotationInfo class to get a list of all remote Annotations of the current user. #1262

Fixed

  • Fixed an issue with RemoteDataset.explore_and_add_remote() where including a remote dataset failed. #1261
  • Fixed an issue with the Zarr array's shape when using Dataset.from_images. #1267

2.0.0 - 2025-03-04

Commits

Breaking Changes

  • Changed writing behavior. There is a new argument allow_resize for MagView.write, which defaults to False. If set to True, the bounding box of the underlying Layer will be resized to fit the to-be-written data. That largely mirrors the previous behavior. However, it is not safe for concurrent operations, so it is disabled by default. It is recommended to set the Layer.bounding_box to the desired size before writing. Additionally, by default, writes need to be aligned with the underlying shard grid to guard against concurrency issues and avoid performance footguns. There is a new argument allow_unaligned, which defaults to False. If set to True, the check for shard alignment is skipped.
  • Deprecated chunks_per_shard arguments in favor of shard_shape, which equals to shard_shape = chunk_shape * chunks_per_shard. The shard shape is more intuitive, because it directly defines the size of shards instead of being a factor of the chunk shape.
  • Deprecated dtype_per_layer argument, because it promotes the use of uncommon dtypes and leads to confusion with the other dtype_per_channel argument. With this change only the use of dtype_per_channel is encouraged.
  • Removed deprecated functions, properties and arguments:
  • Functions:
    • open_annotation, use Annotation.load() instead
    • Dataset.get_color_layer, use Dataset.get_color_layers() instead
    • Dataset.get_segmentation_layer, use Dataset.get_segmentation_layers() instead
    • Dataset.create, use Dataset.__init__ instead
    • Dataset.get_or_create, use Dataset.__init__ with exist_ok=True instead
    • Layer.get_best_mag, use Layer.get_finest_mag instead
    • View.read_bbox, use read with relative_bounding_box or absolute_bounding_box instead
    • View.__enter__ and View.__exit__, context managers are not needed anymore
    • open_nml, use Skeleton.load() instead
    • Group.add_graph, use Group.add_tree instead
    • Group.get_max_graph_id, use Group.get_max_tree_id instead
    • Group.flattened_graphs, use Group.flattened_trees instead
    • Group.get_graph_by_id, use Group.get_tree_by_id instead
    • Skeleton.from_path, use Skeleton.load() instead
    • Skeleton.write, use Skeleton.save() instead
  • Properties:
    • Annotation.username, use Annotation.owner_name instead
    • Annotation.scale, use Annotation.voxel_size instead
    • Annotation.user_id, use Annotation.owner_id instead
    • ArrayInfo.shard_size, use ArrayInfo.shard_shape instead
    • Dataset.scale, use Dataset.voxel_size instead
    • MagView.global_offset, always (0, 0, 0, ...)
    • MagView.size, use mag_view.bounding_box.in_mag(mag_view.mag).bottomright
    • MagViewProperties.resolution, use MagViewProperties.mag instead
    • LayerProperties.resolutions, use LayerProperties.mags instead
    • View.header, use View.info instead
    • View.global_offset, use view.bounding_box.in_mag(view.mag).topleft instead
    • View.size, use view.bounding_box.in_mag(view.mag).size instead
    • Group.graphs, use Group.trees
    • Skeleton.scale, use Skeleton.voxel_size instead
  • Arguments:
    • annotation_type in Annotation.download, not needed anymore
    • annotation_type in Annotation.open_as_remote_dataset, not needed anymore
    • size in BufferedSliceReader.__init__, use relative_bounding_box or absolute_bounding_box instead
    • offset in BufferedSliceReader.__init__, use relative_bounding_box or absolute_bounding_box instead
    • offset in BufferedSliceWriter.__init__, use relative_bounding_box or absolute_bounding_box instead
    • json_update_allowed in BufferedSliceWriter.__init__, not supported anymore
    • offset in BufferedSliceWriter.reset_offset, use relative_offset or absolute_offset instead
    • scale in Dataset.__init__, use voxel_size or voxel_size_with_unit instead
    • dtype in Dataset.add_layer, use dtype_per_channel instead
    • dtype in Dataset.get_or_add_layer, use dtype_per_channel instead
    • chunk_size in Dataset.add_layer_from_images, use chunk_shape instead
    • chunk_size in Dataset.copy_dataset, use chunk_shape instead
    • block_len in Dataset.copy_dataset, use chunk_shape instead
    • file_len in Dataset.copy_dataset, use chunks_per_shard instead
    • args in Dataset.copy_dataset, use executor instead
    • chunk_size in Layer.add_mag, use chunk_shape instead
    • block_len in Layer.add_mag, use chunk_shape instead
    • file_len in Layer.add_mag, use chunks_per_shard instead
    • chunk_size in Layer.get_or_add_mag, use chunk_shape instead
    • block_len in Layer.get_or_add_mag, use chunk_shape instead
    • file_len in Layer.get_or_add_mag, use chunks_per_shard instead
    • args in Layer.downsample, use executor instead
    • args in Layer.downsample_mag, use executor instead
    • args in Layer.redownsample, use executor instead
    • args in Layer.downsample_mag_list, use executor instead
    • args in Layer.downsample_mag_list, use executor instead
    • buffer_edge_len in Layer.upsample, use buffer_shape instead
    • args in Layer.upsample, use executor instead
    • min_mag in Layer.upsample, use finest_mag instead
    • offset in MagView.write, use relative_offset, absolute_offset, relative_bounding_box, or absolute_bounding_box instead
    • json_update_allowed in MagView.write, use allow_resize instead
    • args in MagView.compress, use executor instead
    • offset in View.write, use relative_offset, absolute_offset, relative_bounding_box, or absolute_bounding_box instead
    • json_update_allowed in View.write, not supported anymore
    • offset in View.read, use relative_offset, absolute_offset, relative_bounding_box, or absolute_bounding_box instead
    • offset in View.get_view, use relative_offset, absolute_offset, relative_bounding_box, or absolute_bounding_box instead
    • offset in View.get_buffered_slice_writer, use relative_offset, absolute_offset, relative_bounding_box, or absolute_bounding_box instead
    • offset in View.get_buffered_slice_reader, use relative_bounding_box, or absolute_bounding_box instead
    • size in View.get_buffered_slice_reader, use relative_bounding_box, or absolute_bounding_box instead
    • chunk_size in View.for_each_chunk, use chunk_shape instead
    • source_chunk_size in View.for_zipped_chunks, use source_chunk_shape instead
    • target_chunk_size in View.for_zipped_chunks, use target_chunk_shape instead
    • args in View.content_is_equal, use executor instead
  • Classes:
    • Graph, use Tree instead
  • Changed defaults:
  • exist_ok in Dataset.__init__ is now False
  • compress in Dataset.from_images is now True
  • compress in Dataset.add_layer_from_images is now True
  • DEFAULT_DATA_FORMAT is now Zarr3
  • compress in Layer.add_mag is now True
  • compress in Layer.upsample is now True
  • buffer_size in View.get_buffered_slice_reader is now computed from the shard shape
  • buffer_size in View.get_buffered_slice_writer is now computed from the shard shape
  • Moved from positional argument to keyword-only argument:
  • json_update_allowed in MagView.write
  • organization_id, sharing_token, webknossos_url, bbox, layers, mags, path, exist_ok in Dataset.download
  • layers_to_link, jobs in Dataset.upload
  • dtype_per_layer, dtype_per_channel, num_channels, data_format, bounding_box in Dataset.add_layer
  • dtype_per_layer, dtype_per_channel, num_channels, data_format in Dataset.get_or_add_layer
  • data_format, mag, chunk_shape, chunks_per_shard, shard_shape, compress in Dataset.add_layer_from_images
  • chunk_shape, shard_shape, chunks_per_shard, data_format, compress, executor in Dataset.add_copy_layer
  • organization_id, tags, name, folder_id in Dataset.get_remote_datasets
  • make_relative in Dataset.add_symlink_layer
  • name, make_relative, layers_to_ignore in Dataset.shallow_copy_dataset
  • executor in Dataset.compress
  • sampling_mode, coarsest_mag, executor in Dataset.downsample
  • voxel_size, chunk_shape, shard_shape, chunks_per_shard, data_format, compress, executor, voxel_size_with_unit in Dataset.copy_dataset
  • chunk_shape, shard_shape, chunks_per_shard, compress in Layer.add_mag
  • chunk_shape, shard_shape, chunks_per_shard, compress in Layer.get_or_add_mag
  • extend_layer_bounding_box, chunk_shape, shard_shape, chunks_per_shard, compress, executor in Layer.add_copy_mag
  • make_relative, extend_layer_bounding_box in Layer.add_symlink_mag
  • extend_layer_bounding_box in Layer.add_remote_mag
  • extend_layer_bounding_box in Layer.add_fs_copy_mag
  • move, extend_layer_bounding_box in Layer.add_mag_from_zarrarray
  • from_mag, coarsest_mag, interpolation_mode, compress, sampling_mode, align_with_other_layers, buffer_shape, force_sampling_scheme, allow_overwrite, only_setup_mags, executor in Layer.downsample
  • interpolation_mode, compress, buffer_shape, allow_overwrite, only_setup_mag, executor in Layer.downsample_mag
  • interpolation_mode, compress, buffer_shape, executor in Layer.redownsample
  • interpolation_mode, compress, buffer_shape, allow_overwrite, only_setup_mags, executor in Layer.downsample_mag_list
  • finest_mag, compress, sampling_mode, align_with_other_layers, buffer_shape, executor in Layer.upsample
  • chunk_shape, executor in SegmentationLayer.refresh_largest_segment_id
  • chunk_shape, shard_shape, chunks_per_shard, compression_mode, path in MagView.create
  • target_path, executor in MagView.compress
  • Added arguments:
  • allow_resize in MagView.write with default False
  • allow_unaligned in MagView.write with default False
  • shard_shape in Dataset.from_images
  • shard_shape in Dataset.add_layer_from_images
  • shard_shape in Dataset.copy_dataset
  • Newly deprecated arguments:
  • chunks_per_shard in Dataset.from_images, use shard_shape instead
  • dtype_per_layer in Dataset.add_layer, use dtype_per_channel instead
  • dtype_per_layer in Dataset.get_or_add_layer, use dtype_per_channel instead
  • chunks_per_shard in Dataset.add_layer_from_images, use shard_shape instead
  • chunks_per_shard in Dataset.copy_dataset, use shard_shape instead
  • dtype_per_layer in Layer.__init__, use dtype_per_channel instead
  • chunks_per_shard in Layer.add_mag, use shard_shape instead
  • chunks_per_shard in Layer.get_or_add_mag, use shard_shape instead
  • chunks_per_shard in Layer.add_copy_mag, use shard_shape instead
  • chunks_per_shard in MagView.create, use shard_shape instead
  • Newly deprecated properties:
  • Layer.dtype_per_layer

Added

  • Added the webknossos copy-dataset CLI command. #1259
  • Added Dataset.write_layer method for writing entire layers in one go. #1242

0.16.10 - 2025-02-26

Commits

0.16.9 - 2025-02-24

Commits

Fixed

  • Fixed opening a renamed dataset via an annotation link. #1256

0.16.8 - 2025-02-04

Commits

0.16.7 - 2025-02-03

Commits

Fixed

  • Fixed a misleading error message when Dataset.open() can't find a valid dataset. #1246

0.16.6 - 2025-02-03

Commits

0.16.5 - 2025-01-30

Commits

Added

  • Added support for python 3.13. #1240

Changed

  • When iterating over Tree.nodes(data=True) the nodes data dict was empty. To enable easier access to the data stored in each Node the properties of a Node are included in its data dict now. #1241
  • Dataset.open_remote() adds another method to disambiguate dataset_name and organization_id. #1247

Fixed

  • Fixed a bug with loading metadata lists. #1247

0.16.4 - 2025-01-23

Commits

Added

  • Added list_bounding_boxes() for Zarr-based datasets. #1238

0.16.3 - 2025-01-21

Commits

Breaking Changes

  • RemoteDataset.display_name is deprecated. To change the name of a dataset use the name property instead.
  • Dataset.get_remote_datasets() returns a mapping. The keys of this mapping changed from datasets name to datasets id.
  • Task.create() needs a dataset_id now instead of a dataset_name. Alternatively a RemoteDataset object can be used. The dataset_name is marked as deprecated. As dataset_name is an optional argument now its position has changed, this is important if create() is called only with positional arguments.
  • When uploading an Annotation the organization_id is necessary now. The organization_id might be stored in the Annotation object or it is inferred from the current webknossos_context. #1155

Added

  • Dataset method get_remote_datasets() accepts name and folder_id as arguments now to filter remote datasets.
  • RemoteDataset got an additional property: created.
  • Annotation got an additional property: dataset_id.
  • Dataset.trigger_dataset_import() was added to refresh the datastore to register a newly added dataset. #1236

Changed

  • Updated to WEBKNOSSOS API version 9. This includes support for the new url structure for datasets and the usage of dataset_id. #1231

Fixed

  • Fixed Mag setup for non-public datasets. #1222
  • Fixed an issue when shallow copying datasets with a remote mag. #1224

0.16.2 - 2024-12-18

Commits

Breaking Changes

  • MagView.get_zarr_array now returns a tensorstore array instead of a zarr-python array. #1174

Changed

  • Updated to WEBKNOSSOS API version 8. #1185
  • Using tensorstore for reading and writing zarr 2 and 3 arrays. Removed zarrita and zarr dependency. #1174

0.16.1 - 2024-12-05

Commits

Added

  • Added .nrrd and .nhdr to supported suffixes. #1228
  • Added more docstrings for many public classes and methods. #1225

Changed

  • Removes vcr-py from developer dependencies for testing and adds proxay for recording and replaying API requests. #1198
  • Removed the CZI installation extra from pip install webknossos[all] by default. Users need to manually install it with pip install --extra-index-url https://pypi.scm.io/simple/ webknossos[czi]. #1219
  • Refactored the PimsTiffReader to read the data directly from the tiff file without creating a memmap-able copy first. This greatly reduces the time and storage requirements for converting large tiff files. #1212

Fixed

  • Fixed an issue where adding existing trees to an annotation fails. #1201
  • Fixed unpickling of the SSL_Context to allow for a second or third pickling. #1223
  • Fixed offset error in upsample_cube job #1209

0.16.0 - 2024-11-27

Commits

0.15.11 - 2024-11-26

Commits

Fixed

  • Fixed pickling issue that has been introduced in 0.15.9. #1218

0.15.10 - 2024-11-25

Commits

Fixed

  • Fixed pickling issue that has been introduced in 0.15.9. #1218

0.15.9 - 2024-11-25

Commits

Changed

Removed the CZI installation extra from pip install webknossos[all] by default. Users need to manually install it with pip install --extra-index-url https://pypi.scm.io/simple/ webknossos[czi]. #1219

0.15.8 - 2024-11-15

Commits

Changed

  • Fixed SSL certificate verification for remote datasets by adding CA certificates using certifi. #1211

0.15.7 - 2024-10-25

Commits

Added

  • Added --coarsest-mag argument to the webknossos downsample command. #1203

Fixed

  • Fixed an issue with merging annotations with compressed fallback layers.
  • Fixed an issue where adding a Zarr array with other axes than cxyz leads to an error. #1204

0.15.6 - 2024-10-16

Commits

Added

  • Added add_mag_from_zarrarray to Layer class, to add existing Zarr arrays as a mag of a layer. #1151

Changed

  • Replaced the Python package manager poetry with uv. #1199

0.15.5 - 2024-09-26

Commits

Added

  • Webknossos API functions were added: Team.get_list(), Team.add("new_name"), User.assign_team_roles("teamName", isTeamManager: True) and RemoteDataset.explore_and_add_remote() are available now. #1196

0.15.4 - 2024-09-23

Commits

Added

  • Enable metadata access for remote datasets. #1163

0.15.3 - 2024-09-11

Commits

Breaking Changes

  • Conversion of images with 4 channels creates a dataset with four layers instead of a dataset with one RGB layer. #1192

Changed

  • Updated tifffile dependency to v2024.8.30. #1190

0.15.2 - 2024-09-05

Commits

Changed

  • Updated and clarified documentation for nd_bounding_box intersected_with method

Fixed

  • Fixed an issue with cattrs v24.1.0. #1184

0.15.1 - 2024-08-13

Commits

Changed

  • Updates zarrita to 0.2.7. #1169

0.15.0 - 2024-08-09

Commits

Breaking Changes

  • Python version 3.8 is no longer officially supported. #1068

Added

  • Added example for scaling a skeleton. #1147

Changed

  • Added options --layer-name and --mag for compress command of the CLI. #1141
  • Added options --chunk-shape and --chunks-per-shard for convert command of the CLI. #1150
  • The from_images method of the Dataset supports directories and single files as input_path now. #1152
  • Added support for python version 3.12. #1068
  • The number of pixel limit for JPG conversion is disabled now. #1154
  • Added option --batch-size to the convert command of the CLI. #1158

Fixed

  • Fixed issue with webknossos URL and context URL being considered different when opening a remote dataset due to trailing slashes. #1137
  • Fix an issue where the remote folder was not found when the folder path query includes a trailing slash. #1164

0.14.26 - 2024-07-22

Commits

Fixed

  • Add a converter to the VoxelSize field factor, to ensure it is a tuple.

0.14.25 - 2024-07-18

Commits

Added

  • Added support for new voxel size that stores unit and updated to WEBKNOSSOS API version 7. #1136

0.14.24 - 2024-07-09

Commits

Added

  • Added an implementation of padded_with_margins for NDBoundingBox class. #1120
  • Added a new method add_nx_graphs to skeleton.py which supports to add nx.Graphs to the Skeleton object. #1130

Changed

  • Removed additional logging messages during image conversion. #1124

Fixed

  • Fixed an issue where cube jobs upsampling, downsampling and compress failed when performed on more than 3 dimensions. #1095
  • Fixed an issue where webknossos libs crash when installed with minimal dependencies. #1104

0.14.23 - 2024-06-28

Commits

Breaking Changes

  • Preferring a custom tiff reader over the default PIMS reader to convert tiff files. This change enables the recognition of axis information and the support of tifffiles with more than 3 dimensions. However, it also leads to changed behavior when converting tiff files. Tiffs with axes other than c, x, y, and z, with a shape bigger than 1, are no longer supported for conversion to WKW. Please convert these files to Zarr or Zarr3 Datasets instead. #1043

Added

  • Added a pixel level heuristic for distinguishing color and segmentation layers when importing image data with the from_images or add_layer_from_images method. #1007
  • Added .ims as supported suffix. #1085
  • Added suffixes supported by bioformats for Zeiss CZI, Leica LOF, Zeiss LSM (laser scanning microscope), Zeiss LSM (Laser Scanning Microscope) 510/710, Leica XLEF and Zeiss AxioVision ZVI (Zeiss Vision Image). #1086
  • Added support for setting a default ID mapping for segmentation layers. #1118

Changed

  • Moved functional parts of merge volume annotation CLI to Dataset and Annotation classes. #1055
  • Set a new max value for test_align_with_mag_against_numpy_implementation to avoid failures due to high numbers. #1082
  • Updated dependabot.yml #1087
  • Make lookup for supported suffixes case-insensitive. #1100

Fixed

  • Fixed an issue with downloading annotations through the Command Line Interface. #1083

0.14.22 - 2024-05-13

Commits

Fixed

  • Performing webknossos upload on a windows machine led to loss of directory structure due to backslashes in the relative paths. This was fixed by #1067

0.14.21 - 2024-05-07

Commits

Changed

  • Added layer_name as optional argument to Dataset.from_images method. If the created dataset contains only a single layer, layer_name is used, otherwise the given layer_name is a common prefix for all layers. #1054
  • The context variable of View.get_buffered_slice_writer() is a BufferedSliceWriter now instead of a Generator. Interaction with the SliceWriter does not change, but updating the offset after first initialization is possible now. #1052

0.14.20 - 2024-04-23

Commits

Changed

  • Updated ruff to v0.4.0 #1047
  • Added NIfTI suffix .nii to list of supported bioformats suffixes. #1048

0.14.19 - 2024-04-18

Commits

Changed

  • Removed special CLI command for cubing Nifti files. Use regular conversion command instead. Further, moved Python dependencies for examples and dev dependencies into optional groups which are not installed by default. Install with poetry install --with dev --with examples. #1024

0.14.18 - 2024-04-18

Commits

Fixed

  • Fixed a bug, where using an unaligned topleft value for add_layer_from_images leads to corrupted data. #1036

0.14.17 - 2024-04-10

Commits

Changed

  • The characters @ and $ are allowed within layer names now. #1034

0.14.16 - 2024-04-04

Commits

Added

  • Add CLI tool for offline merging of zip annotations with fallback datasets. #996
  • Added support for converting new file formats such as DICOM, using the bioformats reader. #1014

Changed

  • The rules for naming the layers have been tightened to match the allowed layer names on webknossos. #1016
  • Replaced PyLint linter + black formatter with Ruff for development. #1013
  • The remote operations now use the WEBKNOSSOS API version 6. #1018
  • The conversion of 4D Tiff files to a Zarr3 Dataset is possible. NDBoundingBoxes and VecInt classes are introduced to support working with more than 3 dimensions. #966

0.14.15 - 2024-02-07

Commits

0.14.14 - 2024-01-12

Commits

Added

  • Added a method to the Datasets class that calculates a dataset's bounding box covering all layers. #975

0.14.13 - 2024-01-02

Commits

Fixed

  • Fixed a bug in reading project info from webknossos using the api client for non-admins. #972

0.14.12 - 2023-12-19

Commits

Fixed

  • Fixes that the buffered slice writer could overwrite data when writing less slices than buffer_size at an offset that is not aligned. #973

0.14.11 - 2023-12-06

Commits

0.14.10 - 2023-12-04

Commits

Fixed

  • Fixed a bug in reading project info from webknossos using the api client. #970

0.14.9 - 2023-11-29

Commits

0.14.8 - 2023-11-28

Commits

0.14.7 - 2023-11-17

Commits

0.14.6 - 2023-11-17

Commits

0.14.5 - 2023-11-08

Commits

Changed

  • Performance improvements for reading from and writing to sharded zarr3 datasets, also speeding up the automated tests #963

0.14.4 - 2023-11-07

Commits

Added

  • The DEFAULT_DATA_FORMAT and DEFAULT_CHUNKS_PER_SHARD can now be set with the env variables WK_DEFAULT_DATA_FORMAT and WK_DEFAULT_CHUNKS_PER_SHARD
  • A Vec3Int can now be initialized with a string containing an int or a tuple.

Changed

  • Upgrades mypy to 1.6. #956
  • Refactored the WEBKNOSSOS API client to no longer use generated client code. #948

0.14.3 - 2023-10-19

Commits

Fixed

  • Fixes bug in FSStore creation when using local paths for zarr data format. #955

0.14.2 - 2023-10-18

Commits

0.14.1 - 2023-10-13

Commits Commits

Added

Changed

  • The WK_USE_ZARRITA env variable is no longer required. zarrita is always installed and now the default for Zarr and Zarr3 datasets. #950
  • Updates various dependencies. #943

0.14.0 - 2023-10-11

Commits

Breaking Changes

  • wait_and_ensure_success from webknossos.utils now requires an executor argument. #943

Changed

  • Updates various dependencies. #943

0.13.7 - 2023-10-07

Commits

Fixed

  • Fixed a bug in writing compressed data. #942

0.13.6 - 2023-08-17

Commits

0.13.5 - 2023-08-15

Commits

Added

  • Added task_type property to Task class. #938

Fixed

  • Fixed a bug where parallel access to the properties json leads to an JsonDecodeError in the webknossos CLI #919

0.13.4 - 2023-08-14

Commits

Breaking Changes

  • Task/Project management: open tasks have been renamed to pending. Use Task.status.pending_instance_count instead of Task.status.open_instance_count. #930

Fixed

  • Fixed an infinite loop in the mag calculation during anisotropic downsampling in situations where the target mag cannot possibly be reached while adhering to the anisotropic downsampling scheme. #934

0.13.3 - 2023-08-08

Commits

Added

  • View has a map_chunk method now to run a function on each chunk and collect the results in a list.

Changed

  • As WEBKNOSSOS does not require the largest segment id. It is also not mandatory in the WEBKNOSSOS libs anymore. #917 The method SegmentationLayer.refresh_largest_segment_id was added to lookup the highest value in segmentation data and set largest_segment_id accordingly.
  • The convert command of the cli now has a --category flag, to select the LayerCategoryType.

0.13.2 - 2023-07-26

Commits

Changed

  • The convert command of the cli now has a --category flag, to select the LayerCategoryType.

0.13.1 - 2023-07-17

Commits

Changed

  • The conversion folder structures to layer names does not allow slashes in the layer name anymore. #918

Fixed

  • Fixed a bug where compression in add_layer_from_images uses too much memory #900

0.13.0 - 2023-06-21

Commits

Added

  • Added duration_in_seconds and modified to AnnotationInfo. #914
  • Added zarrita storage backend for arrays. zarrita supports Zarr v2 and v3 including sharding. To activate zarrita, the environment variable WK_USE_ZARRITA must be set. #912
  • Added a Zarr3 data format which supports sharding. #912

Changed

  • Integrated the wkcuber CLI tool into webknossos package. #903
  • To get an overview of all webknossos subcommands type webknossos --help. If the usage of a single subcommand is of interest type webknossos <subcommand> --help
  • These commands were changed:
    • python -m wkcuber, python -m wkcuber.convert_image_stack_to_wkw -> webknossos convert
    • python -m wkcuber.export_wkw_as_tiff -> webknossos export-wkw-as-tiff
    • python -m wkcuber.convert_knossos -> webknossos convert-knossos
    • python -m wkcuber.convert_raw -> webknossos convert-raw
    • python -m wkcuber.downsampling -> webknossos downsample
    • python -m wkcuber.compress -> webknossos compress
    • python -m wkcuber.check_equality -> webknossos check-equality
  • There is one new command:
    • webknossos upload to upload a dataset to a WEBKNOSSOS server
  • These commands have been removed:
    • python -m wkcuber.cubing
    • python -m wkcuber.tile_cubing
    • python -m wkcuber.metadata
    • python -m wkcuber.recubing

Fixed

  • Fixed a bug where upsampling of a layer would fail, if the layer had a bounding box that doesn't align with the from_mag mag. #915

0.12.6 - 2023-06-09

Commits

Changed

  • Upgrades wkw. #911

0.12.5 - 2023-06-01

Commits

Added

  • Added support for Python 3.11. #843

0.12.4 - 2023-05-25

Commits

Added

  • Group.add_tree now also accepts a tree object as a first parameter (instead of only a string). #891
  • Group.remove_tree_by_id was added. #891

Changed

  • Upgrades black, mypy, pylint, pytest. #873

Fixed

  • Fixed poetry build backend for new versions of Poetry. #899
  • Added axis_order fields for Zarr data format. #902

0.12.3 - 2023-02-22

Commits

Added

  • Added support to import ImageJ Hyperstack tiff files via Dataset.from_images and dataset.add_layer_from_images. #877

Changed

  • Dataset.from_images and dataset.add_layer_from_images now automatically convert big endian dtypes to their little endian counterparts by default. #877

Fixed

  • Fixed reading czi files with non-zero axis offsets. #876

0.12.2 - 2023-02-20

Commits Commits

Added

  • Added RemoteFolder for assigning remote datasets to a WEBKNOSSOS folder. #868

0.12.1 - 2023-02-16

Commits

Added

  • Added read_only parameter for annotation.temporary_volume_layer_copy. #866

Changed

  • in Dataset.layer_from_images, layer names generated from multiple input image channels no longer contain equal signs, yielding better url safety. #867

Fixed

  • Fixed a bug where some czi, dm3, dm4 images could not be converted to wkw due to a too-strict check. #865
  • Enforce read_only property of datasets also for down- and upsampling. #866

0.12.0 - 2023-02-10

Commits

Breaking Changes

  • Dropped support for Python 3.7. #833

Added

  • Added support for Python 3.10. #833

0.11.4 - 2023-02-09

Commits

0.11.3 - 2023-02-06

Commits

0.11.2 - 2023-01-18

Commits

Fixed

  • Fixed a bug, where the image order could be randomized when passing a directory path to Dataset.add_layer_from_images. #854

0.11.1 - 2023-01-05

Commits

Added

  • Dataset.from_images and dataset.add_layer_from_images have new features: #842
  • dm3 and dm4 datasets can be read without bioformats now.
  • It's possible to completely disable the bioformats adapter by setting use_bioformats to False.
  • Lists of images can now be handled with other readers, before only images supported by skimage worked in lists.

0.11.0 - 2022-12-09

Commits

Breaking Changes

  • Removed the id attribute of the BoundingBox class, also from the constructor. #836

Fixed

  • Fixed bounding box serialization in NMLs, so that bounding boxes which are uploaded via annotations are now recognized properly by webKnossos. #836
  • Bounding boxes keep their name, color and visibility when transformed via methods, such as bbox.padded_with_margins(). #836

0.10.27 - 2022-12-07

Commits

Added

  • Short links, such as https://webknossos.org/links/93zLg9U9vJ3c_UWp, are now supported for dataset and annotation urls in download and open_remote methods. #837

0.10.26 - 2022-12-05

Commits

Changed

  • Dataset.from_images and dataset.add_layer_from_images now try to import the images via the bioformats after all other options as well. #829

Fixed

  • dataset.add_layer_from_images can now handle paths to folders which only contain a single image. #829

0.10.25 - 2022-11-29

Commits

Breaking Changes

  • Dataset.from_images now adds a layer per timepoint and per channel (if the data doesn't have 1 or 3 channels). #822

Added

  • Added python-native CZI support for Dataset.from_images or dataset.add_layer_from_images, without using bioformats. #822
  • dataset.add_layer_from_images can add a layer per timepoint and per channel when passing allow_multiple_layers=True. #822

0.10.24 - 2022-11-09

Commits

Changed

  • Updated cattrs dependency to 22.2.0. #819

0.10.23 - 2022-11-01

Commits

Breaking Changes

  • ds.add_layer_from_images: Turned some arguments into keyword-only arguments, only affecting positional arguments after the first 8 arguments. #818

Added

  • ds.add_layer_from_images: added topleft and dtype kw-only arguments. #818

0.10.22 - 2022-10-27

Commits

Fixed

  • Fixed a bug where some image sequences could not be read in layer_from_images. #817

0.10.21 - 2022-10-26

Commits

0.10.20 - 2022-10-20

Commits

Changed

  • Make volume locations optional, allowing to parse segment information in future NML-only annotations. #814

Fixed

  • annotation.temporary_volume_layer_copy() works also with empty volume annotations. #814

0.10.19 - 2022-10-18

Commits

Breaking Changes

  • The executor argument can now be passed to the following methods, args is deprecated now for those #805:
  • dataset.copy_dataset(…)
  • layer.upsample(…)
  • layer.downsample(…)
  • layer.downsample_mag(…)
  • layer.downsample_mag_list(…)
  • layer.redownsample(…)
  • mag_view.compress(…)
  • view.content_is_equal(…)

Added

  • Added Dataset.from_images, which converts images to a Dataset, possibly consisting of multiple layers. [#808](https://github.com/scalableminds/webknossos-libs/pull/808
  • Added Annotation.open_as_remote_dataset(…), which is a shorthand for Annotation.download(...).get_remote_annotation_dataset(). #811
  • skeleton.save() now also accepts paths with a .zip suffix. #811
  • Added annotation.get_volume_layer_segments() to interact with information from the Segments tab in annotations. This method returns a dict from segment ids to an object containing optional segment name, color and anchor_position. #812
  • Added convenience methods Dataset.compress and Dataset.downsample for compressing and downsampling all layers and mags in a dataset. #813
  • Added examples for downloading segment masks from webKnossos and cubing & uploading tiff stacks. #813

0.10.18 - 2022-09-29

Commits

Added

  • Annotation.download() now accepts the keyword-only argument skip_volume_data, which can be set to True to omit downloading volume data. #806

0.10.17 - 2022-09-26

Commits

Added

  • Added annotation.get_remote_annotation_dataset() to get a streamed annotation dataset, which also reflects fallback layers or applied mappings. #794
  • Added direct access to an underlying Zarr array with the MagView.get_zarr_array() method. #792

0.10.16 - 2022-09-13

Commits

Added

  • Added direct access to an underlying Zarr array with the MagView.get_zarr_array() method. #792

Changed

  • Upgraded zarr and numcodecs. #798

0.10.14 - 2022-08-30

Commits

Changed

  • dataset.add_copy_layer() and layer.add_copy_mag() now read and write the image data, not copying files. This allows to stream data from remote datasets. To continue using the filesystem copy mechanism, please use dataset.add_fs_copy_layer() or layer.add_fs_copy_mag(). #790

0.10.13 - 2022-08-22

Commits

Breaking Changes

  • Renamed chunk_size arguments to chunk_shape. chunk_size is still available as keyword-only argument, but deprecated. #706

Changed

  • The largest_segment_id is optional now. #786

0.10.12 - 2022-08-15

Commits

Fixed

  • Fixed task.get_project(). #785

0.10.11 - 2022-08-03

Commits

0.10.10 - 2022-07-26

Commits

0.10.9 - 2022-07-22

Commits

0.10.8 - 2022-07-15

Commits

0.10.7 - 2022-07-14

Commits

Breaking Changes

  • The Annotation constructor takes the owner_name argument instead of username. This is only important when using keyword arguments. The username attributes are still available as a proxy for the owner_name attribute, but deprecated. #760
  • user_id on AnnotationInfo objects is deprecated, please use owner_id instead. #760
  • When self-hosting a webKnossos server, please note that a webknossos version >= 22.06.0 is required. #760 & #764

Added

  • Dataset.upload() accepts Layer objects from a RemoteDataset in the layers_to_link argument list. Also, LayerToLink can consume those via LayerToLink.from_remote_layer(). #761
  • Task.create() accepts a RemoteDataset for the dataset_name argument. #761
  • Added annotation.get_remote_base_dataset() returning a RemoteDataset. #761
  • Added Team.get_by_name(). #763
  • Added Dataset.get_remote_datasets(). #763

Changed

  • If a token is requested from the user on the commandline, it is now stored in the current context. Before, it was discarded. #761
  • Annotation.download() does not need the annotation_type anymore, and the type can also be omitted from passed URLs. #764
  • Dataset.add_layer_from_images allows smaller batch sizes for uncompressed writes. #766
  • Dataset.add_layer_from_images uses multiprocessing by default. #766

Fixed

  • Fixed the bounding box inferral for volume annotation layers that were not saved in Mag(1). #765

0.10.6 - 2022-06-27

Commits

0.10.5 - 2022-06-15

Commits

0.10.4 - 2022-06-15

Commits

Breaking Changes

  • Added an align_with_other_layers parameter to Layer.downsample & layer.upsample. When set to true (default), the magnifications of the existing dataset are used as guidance for downsampling/upsampling. Instead of passing a boolean, one can also pass another dataset that should be used as guidance. #730
  • Changed the name of max_mag in Layer.downsample to coarsest_mag. #730

Added

  • Added Dataset.add_layer_from_images() to convert image stacks to wkw or zarr Dataset. This needs pims and possibly more packages, which can be installed using extras, e.g. "webknossos[all]". #741

Changed

  • The Layer.downsample and Layer.upsample function now automatically downsample according to magnifications already existing in the dataset. This behaviour can be turned off by setting the new parameter align_with_other_layers to False. #730
  • View.get_buffered_slice_reader() and View.get_buffered_slice_writer() don't log anything by default now. To get the previous logging, please invoke them with logging=True. #741

0.10.3 - 2022-06-03

Commits

Added

0.10.2 - 2022-05-20

Commits

Changed

  • Added Python 3.9 support to wk-libs #716

Fixed

  • URLs for the webknossos-context (e.g. in the WK_URL env var or via webknossos_context(url=…)) may now contain / in the end and are sanitized. Before, requests would fail if the URL contained a final /. #733

0.10.1 - 2022-05-10

Commits

0.10.0 - 2022-05-09

Commits

Breaking Changes

  • Dataset.upload() now returns a RemoteDataset instead instead of the URL string. You can get the URL via remote_ds.url. #723
  • User.teams now is a tuple instead of a list. #723
  • The deprecated download_dataset function now requires the organization_id argument. #723

Added

  • Added Dataset.open_remote(), which returns an object of the new RemoteDataset. #723 This can
  • give the webknossos URL for the dataset as remote_ds.url,
  • read image data via the webknossos zarr interface, using the inherited Dataset methods, and
  • read and change the following dataset metadata as properties: display_name, description, tags, is_public, sharing_token, allowed_teams.
  • Team instances also contain the organization_id. #723

0.9.24 - 2022-05-03

Commits

Fixed

  • Fixed upsampling with constant z in certain anisotropic cases. #720

0.9.23 - 2022-05-03

Commits

0.9.22 - 2022-05-02

Commits

Breaking Changes

  • Renamed some entities to comply with the webKnossos terminology. #704:
  • The Graph class changed to Tree, also related methods and attributes are renamed now, e.g. add_graph is now add_tree. All previous entities are still available, but deprecated.
  • scale has changed to voxel_size for datasets, skeletons and annotations. Changes in Dataset are backwards-compatible, but scale is deprecated. For Annotation and Skeletons the initializer only supports voxel_size, the scale attribute is deprecated
  • get_best_mag is deprecated, please use get_finest_mag instead
  • In layer.upscale, min_mag is deprecated in favor of finest_mag

Fixed

  • Correctly maintain default_view_configuration property when downloading a dataset. #677

0.9.21 - 2022-04-20

Commits

0.9.20 - 2022-04-20

Commits

Fixed

  • Fixed a bug where the server’s error message during dataset upload was not displayed to the user. #702

0.9.19 - 2022-04-11

Commits

0.9.18 - 2022-04-06

Commits

0.9.17 - 2022-04-05

Commits

0.9.16 - 2022-03-31

Commits

0.9.15 - 2022-03-31

Commits

Added

  • Added cloud storage support for the Dataset classes, by using UPath from universal_pathlib and fsspec. Create remote datasets like this Dataset(UPath("s3://bucket/path/to/dataset", key="...", secret="..."), scale=(11, 11, 24)). Datasets on cloud storage only work with Zarr-based layers. #649

0.9.14 - 2022-03-24

Commits

Breaking Changes

  • Dataset.download: The argument dataset_name was renamed to dataset_name_or_url. #660

Added

  • Dataset.download now also accepts a URL, as well as a sharing_token, which can also be part of the URL. #660

0.9.13 - 2022-03-24

Commits

Breaking Changes

  • Consistently rename organization_name parameters to organization_id (except in deprecated webknossos.client.download_dataset). #639

Changed

  • MagView.compress now skips in-place compression of already compressed mags. #667
  • Replaced uses of pathlib.Path with UPath from universal_pathlib. Since UPath is compatible with pathlib.Path, changes in user code are not necessary. #649

Fixed

  • Fixed compression of downsampled mags for layers with arbitrary and potentially mag-unaligned bounding boxes. #667

0.9.12 - 2022-03-18

Commits

Breaking Changes

  • The annotation class now exposes BoundingBox objects instead of tuples. #646

Added

  • Added groups and graphs property to skeleton.Group to access immediate child groups/graphs. #645
  • The BoundingBox class now supports the following additional properties: id, name, is_visible and color. #646
  • Added support for Zarr arrays in the Dataset classes. Users can set the data_format of layers to zarr to use Zarr for storing data. #627 The current implementation has some limitations, e.g.:
  • Only one type of compression (Blosc+Zstd) is implemented.
  • Sharding is not available in Zarr, yet. Please use chunks_per_shard = (1, 1, 1).
  • Only local filesystem-based arrays are supported. There are changes to the datasource-properties.json for Zarr layers compared to WKW layers:
  • dataFormat needs to be changed to zarr.
  • The list of mags is called mags, instead of wkwResolutions.
  • Each mag is represented by an object with a single attribute mag, e.g. { "mag": [1, 1, 1] }.

Changed

  • Dataset: block_len and file_len attributes are now deprecated, but still available for backwards compatibility. Use chunk_size and chunks_per_shard instead. These new attributes are Vec3Int, so they can be set non-uniformly. However, WKW-backed layers still require uniform chunk_size and chunks_per_shard. #627

Fixed

  • Fixed crash during downsampling and compression of segmentation layers. #657

0.9.11 - 2022-03-16

Commits

0.9.10 - 2022-03-15

Commits

Changed

  • Annotation: Annotation.temporary_volume_layer_copy now uses the NML-provided scale. #644
  • Dataset: Moved the deprecation warning from get_color_layers() to the actually deprecated method get_color_layer(). #635
  • Inconsistent writes to datasets properties (e.g., caused due to multiprocessing) are detected automatically. The warning can be escalated to an exception with warnings.filterwarnings("error", module="webknossos", message=r"\[WARNING\]"). #633
  • Changed the position of a skeleton.Node to use Vec3Int instead of (Float, Float, Float), because webKnossos stores node positions as integers. #645

Fixed

  • Tests: The ./test.sh script works on macOS again and doesn't throw Network Errors anymore. However the introduced fix could lead to slightly different behaviour on macOS tests vs CI tests, when UNIX socket communication is involved. #618

0.9.9 - 2022-03-03

Commits

Breaking Changes

  • Changed the interface and behavior of Annotations and Skeletons, as well as nml-parsing related code. #602
  • webknossos.skeleton.nml is not exposed anymore. The previous functionality may be found in webknossos._nml and webknossos/annotation/_nml_conversion.py if needed, but please not that this is not part of the public API and may change at any time. Please use the respective logic on the Annotation class instead.
  • The name attribute on the Skeleton class changed to dataset_name.
  • The deprecated Skeleton.offset attribute is removed.
  • The following attributes are removed from the Skeleton class, instead they are part of the Annotation class now: time, edit_position, edit_rotation, zoom_level, task_bounding_box, user_bounding_boxes.
  • The following Annotation methods were renamed and their arguments changed slightly:
    • save_volume_annotation ➜ export_volume_layer_to_dataset
    • temporary_volume_annotation_layer_copy ➜ temporary_volume_layer_copy

Added

  • Added new features to the Annotation and Skeleton classes. #602
  • The Skeleton class has new attributes description and organization_id.
  • The Annotation class has new attributes username and annotation_id, as well as the following attributes that are proxies for the skeleton attributes: dataset_name, scale, organization_id, description
  • Annotations can now be initialized from their attributes and an optional skeleton.
  • New methods on Annotation: upload, add_volume_layer, delete_volume_layer
  • Annotation.load() and annotation.save() also works with .nml files.
  • Added MagView.get_views_on_disk() as a shortcut to turning get_bounding_boxes_on_disk into views. #621

Fixed

  • Fixed the download of skeleton-only annotations. #602

0.9.8 - 2022-02-28

Commits

Added

  • Added allow_overwrite parameter to Layer.downsample(). #614

0.9.7 - 2022-02-28

Commits

Added

  • Added only_setup_mag parameter to downsample-related methods in Layer. This parameter allows creating output magnifications before parallelizing downsampling invocations to avoid outdated writes to datasource-properties.json. #610
  • Added Task.create() method to create tasks by providing a dataset name, location, and rotation. #605

0.9.6 - 2022-02-15

Commits

Added

  • Added apply_merger_mode.py example. #592
  • Added support for reading from multiple volume layers in annotations. If an annotation contains multiple volume layers, the layer name has to be provided when reading from a volume layer in an annotation (in Annotation.save_volume_annotation() and Annotation.temporary_volume_annotation_layer_copy()). Also, added the method Annotation.get_volume_layer_names() to see available volume layers. #588

Changed

  • Dataset writes in higher mags do not increase the bounding box if the written data fits in the rounded up box. #595

Fixed

  • Dataset down- & upload: #595
  • Fixed download of higher mags.
  • Bounding box after download is set correctly (was inflated before when downloading higher mags).
  • The returned URL for uploads is corrected, this did not respect new_dataset_name before.

0.9.5 - 2022-02-10

Commits

Fixed

  • Skeleton: Fixed a bug when comparing Graph instances, this fixes failing loads which had the error message Can only compare wk.Graph to another wk.Graph. before. #593

0.9.4 - 2022-02-09

Commits

Added

  • Added AnnotationInfo, Project and Task classes for handling annotation information and annotation project administration. #574

Changed

  • Lifted the restriction that BoundingBox cannot have a negative topleft (introduced in v0.9.0). Also, negative size dimensions are flipped, so that the topleft <= bottomright, e.g. BoundingBox((10, 10, 10), (-5, 5, 5)) -> BoundingBox((5, 10, 10), (5, 5, 5)). #589

0.9.3 - 2022-02-07

Commits

Fixed

  • dataset.upload(layers_to_link=…): Fixed a bug where the upload did not complete if layers_to_link contained layers present in uploading dataset. #584

0.9.2 - 2022-02-03

Commits

Added

  • A custom network request timeout can be set using webknossos_context(…, timeout=300) or export WK_TIMEOUT="300". #577

Changed

  • The default network request timeout changed from ½min to 30 min. #577

0.9.1 - 2022-01-31

Commits

Changed

  • The signatures of (Mag)View.for_each_chunk and (Mag)View.for_zipped_chunks changed:
  • The previous argument work_on_chunk is now called func_per_chunk.
  • The various chunk_size arguments now have to be given in Mag(1). They now have default values.
  • Deprecations in (Mag)View.get_buffered_slice_reader/_writer #564:
  • (Mag)View.get_buffered_slice_reader: using the parameters offset and size is deprecated. Please use the parameter relative_bounding_box or absolute_bounding_box (both in Mag(1)) instead. The old offset behavior was absolute for MagViews and relative for Views.
  • (Mag)View.get_buffered_slice_writer: using the parameter offset is deprecated. Please use the parameter relative_offset or absolute_offset (both in Mag(1)) instead. The old offset behavior was absolute for MagViews and relative for Views.

0.9.0 - 2022-01-19

Commits

Breaking Changes

  • Various changes in View & MagView signatures #553:
  • Breaking Changes:
    • MagView.read: if nothing is supplied and the layer does not start at (0, 0, 0), the default behaviour changes from starting at absolute (0, 0, 0) to the layer's bounding box
    • MagView.write: if no offset is supplied and the layer does not start at (0, 0, 0), the default behaviour changes from starting at absolute (0, 0, 0) to the layer's bounding box
    • (Mag)View.get_view: read_only is a keyword-only argument now
    • MagView.get_bounding_boxes_on_disk() now returns an iterator yielding bounding boxes in Mag(1)
    • BoundingBox cannot have negative topleft or size entries anymore (lifted in v0.9.4).
  • Deprecations The following usages are marked as deprecated with warnings and will be removed in future releases:
    • Using the offset parameter for read/write/get_view in MagView and View is deprecated. There are new counterparts absolute_offset and relative_offset which have to be specified in Mag(1), whereas offset previously was specified in the Mag of the respective View. Also, for read/get_view only using size is deprecated, since it used to refer to the size in the View's Mag. Instead, size should always be used together with absolute_offset or relative_offset. Then it is interpreted in Mag(1).
    • The (Mag)View attributes view.global_offset and view.size are deprecated now, which were in the Mag of the respective View. Please use view.bounding_box instead, which is in Mag(1).
    • read_bbox on the (Mag)View is deprecated as well, please use read with the absolute_bounding_boxor relative_bounding_box parameter instead. You'll have to pass the bounding box in Mag(1) then.

Added

  • Added a check for dataset name availability before attempting to upload. #555

Fixed

  • Fixed the dataset download of private datasets which need a token. #562

0.8.31 - 2022-01-07

Commits

Added

  • Added Annotation.save("file_name") to save an annotation to a file and Annotation.temporary_volume_annotation_layer_copy() to read from the volume layer of an annotation as a WK dataset. #528
  • Added layers_to_link parameter to Dataset.upload() so that layers don't need to be uploaded again if they already exist in another dataset on webKnossos. #544

0.8.30 - 2021-12-27

Commits

Breaking Changes

  • The BoundingBoxNamedTuple was removed. Use BoundingBox instead. #526
  • Some methods of creating, opening and saving have changed. The old methods are still available but deprecated. The documentation gives a good overview. Specifically, the changes are :
  • Dataset.create() → Dataset()
  • Dataset.get_or_create() → Dataset(…, exist_ok=True)
  • Dataset() → Dataset.open()
  • download_dataset() → Dataset.download()
  • open_annotation() → Annotation.load() for local files, Annotation.download() to download from webKnossos
  • open_nml() → Skeleton.load()
  • Skeleton.from_path() → Skeleton.load()
  • Skeleton.write() → Skeleton.save() The deprecated methods will be removed in future releases. #520

Changed

  • The detailed output of e.g. downsampling was replaced with a progress bar. #527
  • Always use the sampling mode CONSTANT_Z when downsampling 2D data. #516
  • Make computation of largestSegmentId more efficient for volume annotations. #531
  • Consistently use resolved instead of absolute path if make_relative is False. #536

0.8.29 - 2021-12-14

Commits

Breaking Changes

  • To download datasets, a recent webknossos server version is necessary (>= 21.12.0). webknossos.org is unaffected. #510

0.8.28 - 2021-12-09

Commits

0.8.27 - 2021-12-09

Commits

v0.8.25 - 2021-12-07

Commits

Added

  • Added support to download datasets from external datastores, which is the case for webknossos.org. #497

Changed

  • Adapt the dataset upload to new webKnossos api. #484
  • get_segmentation_layer() and get_color_layer() were deprecated and should not be used, anymore, as they will fail if no or more than one layer exists for each category. Instead, get_segmentation_layers() and get_color_layers() should be used (if desired in combination with [0] to get the old, error-prone behavior).
  • Renamed the folder webknossos/script-collection to webknossos/script_collection to enable module imports. #505

v0.8.24 - 2021-11-30

Commits

v0.8.23 - 2021-11-29

Commits

Breaking Changes

  • wk.Graph now inherits from networkx.Graph directly. Therefore, the nx_graph attribute is removed. #481
  • The class LayerCategories was removed. COLOR_TYPE and SEGMENTATION_TYPE were renamed to COLOR_CATEGORY and SEGMENTATION_CATEGORY and can now be imported directly. The type of many parameters were changed from str to the literal LayerCategoryType. #454

Added

  • Added redownsample() method to Layer to recompute existing downsampled magnifications. #461
  • Added globalize_floodfill.py script to globalize partially computed flood fill operations. #461

Changed

  • Improved performance for calculations with Vec3Int and BoundingBox. #461

Fixed

  • Resolve path when symlinking layer and make_relative is False (instead of only making it absolute). #492

0.8.22 - 2021-11-01

Commits

Breaking Changes

  • Removed the organization parameter from the webknossos_context function. The organization will automatically be fetched using the token of the user. #470

Fixed

  • Make Views picklable. We now ignore the file handle when we pickle Views. #469

v0.8.19 - 2021-10-21

Commits

Added

  • Added a User class to the client that can be used to get meta-information of users or their logged time. The currently logged in user can be accessed, as well as all managed users. #470

0.8.21 - 2021-10-28

Commits

Changed

  • Downgraded typing-extensions for better dependency compatibility #472

0.8.20 - 2021-10-28

Commits

Breaking Changes

  • BoundingBox.chunk()'s 2nd parameter chunk_border_alignments now does not accept a list with a single int anymore. #452

Fixed

  • Make Views picklable. We now ignore the file handle when we pickle Views. #469

0.8.19 - 2021-10-21

Commits

Breaking Changes

- Views now always open the wkw.Dataset lazily. All explicit calls to View.open() and View.close() must be removed. #448

Added

  • Added a new Annotation class which includes skeletons as well as volume-annotations. #452
  • Added dataset down- and upload as well as annotation download, see the examples learned_segmenter.py and upload_image_data.py. #452

0.8.18 - 2021-10-18

Commits

Added

  • The Dataset class now has a new method: add_shallow_copy. #437

Changed

  • The Vec3Int constructor now asserts that its components are whole numbers also in numpy case. #434
  • Updated scikit-image dependency to 0.18.3. #435
  • BoundingBox.contains now also takes float points in numpy arrays. #450

Fixed

0.8.16 - 2021-09-22

Commits

Breaking Changes

  • Refactored the BufferedSliceWriter and added a BufferedSliceReader. #425
  • BufferedSliceWriter
    • The data no longer gets transposed: previously the format of the slices was [y,x]; now it is [x,y]
    • The interface of the constructor was changed:
    • A View (or MagView) is now required as datasource
    • The parameter dimension can be used to specify the axis along the data is sliced
    • The offset is expected to be in the magnification of the view
    • This class is now supposed to be used within a context manager and the slices are written by sending them to the generator (see documentation of the class).
  • BufferedSliceReader
    • This class was added complementary to the BufferedSliceWriter
  • Added methods to get a BufferedSliceReader/BufferedSliceWriter from a View directly

Added

Changed

Fixed

0.8.15 - 2021-09-22

Commits

Breaking Changes

  • Breaking changes were introduced for geometry classes in #421:

  • BoundingBox

    • is now immutable, use convenience methods, e.g. bb.with_topleft((0,0,0))
    • properties topleft and size are now Vec3Int instead of np.array, they are each immutable as well
    • all to_-conversions return a copy, some were renamed:
    • to_array → to_list
    • as_np → to_np
    • as_wkw → to_wkw_dict
    • from_wkw → from_wkw_dict
    • as_config → to_config_dict
    • as_checkpoint_name → to_checkpoint_name
    • as_tuple6 → to_tuple6
    • as_csv → to_csv
    • as_named_tuple → to_named_tuple
    • as_slices → to_slices
    • copy → (gone, immutable)

  • Mag

    • is now immutable
    • mag.mag is now mag._mag (considered private, use to_list instead if you really need it as list)
    • all to_-conversions return a copy, some were renamed:
    • to_array → to_list
    • scale_by → (gone, immutable)
    • divide_by → (gone, immutable)
    • as_np → to_np

Added

  • An immutable Vec3Int class was introduced that holds three integers and provides a number of convenience methods and accessors. #421

Changed

  • BoundingBox and Mag are now immutable attr classes containing Vec3Int values. See breaking changes above.

Fixed

-

0.8.13 - 2021-09-22

Commits

This is the latest release at the time of creating this changelog.