Fractal-tasks-core

* Fix bug in `cellpose_segmentation` upon using masked loading and setting `channel2` (\639). Thanks [FranziskaMoos-FMI](https://github.com/FranziskaMoos-FMI) and [enricotagliavini](https://github.com/enricotagliavini).
* Improve handling of potential race condition in "Apply Registration to image" task (\638).

* Breaking changes in tasks:
* Make `NapariWorkflowsOutput.label_name` attribute required, and use it to fill the `region["path"]` table attribute (\613).
* Breaking changes in core library:
* ⚠️ Refactor the whole package structure, leading to breaking changes for most imports (\613); more details at [this page](https://fractal-analytics-platform.github.io/fractal-tasks-core/version_updates/v0_14_0/).
* In `prepare_label_group` helper function:
* Make `label_attrs` function argument required (\613).
* Validate `label_attrs` with `NgffImageMeta` model (\613).
* Override multiscale name in `label_attrs` with `label_name` (\613).
* In `write_table` helper function:
* Drop `logger` function argument (\613).
* Add `table_name` function argument, taking priority over `table_attrs` (\613).
* Raise an error if no table type is not provided (\613).
* Raise an error if table attributes do not comply with table specs (\613).
* Other internal changes:
* Comply with table specs V1, by writing all required Zarr attributes (\613).
* Remove `has_args_schema` obsolete property from manifest (\603).
* Handle `GroupNotFoundError` in `load_NgffImageMeta` and `load_NgffWellMeta` (\622).
* Bug fixes:
* Fix table selection in calculate registration image-based (\615).
* Documentation
* Clarify table specs V1 (\613).
* Testing:
* Use more recent Zenodo datasets, created with `fractal-tasks-core>=0.12` (\623).
* Use poetry 1.7.1 in GitHub actions (\620).
* Align with new Zenodo API (\601).
* Update `test_valid_manifest` (\606).
* Use [pooch](https://www.fatiando.org/pooch) to download test files (\#610).
* Documentation:
* Add list of tasks (\625).
* Dependencies:
* Remove Pillow `<10.1.0` constraint (\626).

* Always use `write_table` in tasks, rather than AnnData `write_elem` (\581).
* Remove assumptions on ROI-table columns from `get_ROI_table_with_translation` helper function of `calculate_registration_image_based` task (\591).
* Testing:
* Cache Zenodo data, within GitHub actions (\585).
* Documentation:
* Define V1 of table specs (\582).
* Add mathjax support (\582).
* Add cross-reference inventories to external APIs (\582).

* Tasks:
* New task and helper functions:
* Introduce `import_ome_zarr` task (\557, \579).
* Introduce `get_single_image_ROI` and `get_image_grid_ROIs` (\557).
* Introduce `detect_ome_ngff_type` (\557).
* Introduce `update_omero_channels` (\579).
* Make `maximum_intensity_projection` independent from ROI tables (\557).
* Make Cellpose task work when `input_ROI_table` is empty (\566).
* Fix bug of missing attributes in ROI-table Zarr group (\573).
* Dependencies:
* Restrict `Pillow` version to `<10.1` (\571).
* Support AnnData `0.10` (\574).
* Testing:
* Align with new Zenodo API (\568).
* Use ubuntu-22 for GitHub CI (\576).

* Relax `check_valid_ROI_indices` to support search-first scenario (\555).
* Do not install `docs` dependencies in GitHub CI (\551).

* Make `Channel.window` attribute optional in `lib_ngff.py` (\548).
* Automate procedure for publishing package to PyPI (\545).