ome zarr notebook

[thewtex]

Add OME-Zarr image registration notebook

A simple example that registers one timepoint in an IDR dataset to another.

We use the IDR dataset as it is a familiar reference sample OME-Zarr
dataset in the community. This also demonstrates fetching the data
dynamically from the cloud.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[thewtex]

@mstaring @N-Dekker @stefanklein please take a look 🙏

[N-Dekker]

Very interesting, Matt! I'm still trying to get it to work here locally but I got an error, on import ngff_zarr as nz, below here. Do you have a clue what I might have done wrong?

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
Cell In[4], line 4
      2 import itk
      3 from itkwidgets import view
----> 4 import ngff_zarr as nz

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\ngff_zarr\__init__.py:6
      1 # SPDX-FileCopyrightText: 2022-present Matt McCormick <matt.mccormick@kitware.com>
      2 #
      3 # SPDX-License-Identifier: MIT
      5 from .__about__ import __version__
----> 6 from .cli_input_to_ngff_image import cli_input_to_ngff_image
      7 from .config import config
      8 from .detect_cli_io_backend import ConversionBackend, detect_cli_io_backend

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\ngff_zarr\cli_input_to_ngff_image.py:3
      1 import sys
----> 3 import zarr
      4 from dask.array.image import imread as daimread
      5 from rich import print

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\__init__.py:2
      1 from zarr._version import version as __version__
----> 2 from zarr.api.synchronous import (
      3     array,
      4     consolidate_metadata,
      5     copy,
      6     copy_all,
      7     copy_store,
      8     create,
      9     create_array,
     10     create_group,
     11     empty,
     12     empty_like,
     13     full,
     14     full_like,
     15     group,
     16     load,
     17     ones,
     18     ones_like,
     19     open,
     20     open_array,
     21     open_consolidated,
     22     open_group,
     23     open_like,
     24     save,
     25     save_array,
     26     save_group,
     27     tree,
     28     zeros,
     29     zeros_like,
     30 )
     31 from zarr.core.array import Array, AsyncArray
     32 from zarr.core.config import config

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\api\synchronous.py:7
      3 from typing import TYPE_CHECKING, Any, Literal
      5 from typing_extensions import deprecated
----> 7 import zarr.api.asynchronous as async_api
      8 import zarr.core.array
      9 from zarr._compat import _deprecate_positional_args

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\api\asynchronous.py:12
      9 import numpy.typing as npt
     10 from typing_extensions import deprecated
---> 12 from zarr.core.array import Array, AsyncArray, create_array, get_array_metadata
     13 from zarr.core.array_spec import ArrayConfig, ArrayConfigLike
     14 from zarr.core.buffer import NDArrayLike

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\core\__init__.py:9
      6 from __future__ import annotations
      8 from zarr.core.buffer import Buffer, NDBuffer  # noqa: F401
----> 9 from zarr.core.codec_pipeline import BatchedCodecPipeline

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\core\codec_pipeline.py:20
     18 from zarr.core.config import config
     19 from zarr.core.indexing import SelectorTuple, is_scalar, is_total_slice
---> 20 from zarr.core.metadata.v2 import _default_fill_value
     21 from zarr.registry import register_pipeline
     23 if TYPE_CHECKING:

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\core\metadata\__init__.py:3
      1 from typing import TypeAlias, TypeVar
----> 3 from .v2 import ArrayV2Metadata, ArrayV2MetadataDict
      4 from .v3 import ArrayV3Metadata, ArrayV3MetadataDict
      6 ArrayMetadata: TypeAlias = ArrayV2Metadata | ArrayV3Metadata

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\zarr\core\metadata\v2.py:9
      6 from functools import cached_property
      7 from typing import TYPE_CHECKING, TypedDict, cast
----> 9 import numcodecs.abc
     11 from zarr.abc.metadata import Metadata
     13 if TYPE_CHECKING:

File ~\.conda\envs\ITKElxPoints\Lib\site-packages\numcodecs\__init__.py:116
    112 from numcodecs.json import JSON
    114 register_codec(JSON)
--> 116 from numcodecs import vlen as vlen
    117 from numcodecs.vlen import VLenArray, VLenBytes, VLenUTF8
    119 register_codec(VLenUTF8)

File numcodecs\\vlen.pyx:1, in init numcodecs.vlen()

ImportError: numpy.core.multiarray failed to import (auto-generated because you didn't call 'numpy.import_array()' after cimporting numpy; use '<void>numpy._import_array' to disable if you are certain you don't need it).

[thewtex]

@N-Dekker thanks for testing!

What versions of ngff_zarr, numpy, numcodecs, zarr do you have?

[N-Dekker]

What versions of ngff_zarr, numpy, numcodecs, zarr do you have?

I just did a brand new install on a new pc. Now it says:

ImportError                               Traceback (most recent call last)
Cell In[1], line 5
      3 import ngff_zarr as nz
      4 from rich import print
----> 5 from zarr.storage import FSStore, LRUStoreCache
      6 import zarr
      7 import copy

ImportError: cannot import name 'FSStore' from 'zarr.storage' (C:\Users\...\.conda\envs\ITKElx\Lib\site-packages\zarr\storage\__init__.py)

I also printed the versions, as follows:

import sys
print(sys.version)
print("itk", itk.__version__)
import itkwidgets
print("itkwidgets", itkwidgets.__version__)
print("ngff_zarr", nz.__version__)
import numpy as np
print("numpy", np.__version__)
import numcodecs
print("numcodecs", numcodecs.__version__)
import zarr
print("zarr", zarr.__version__)

Which says:

3.11.11 | packaged by Anaconda, Inc. | (main, Dec 11 2024, 16:34:19) [MSC v.1929 64 bit (AMD64)]
itk 5.4.0
itkwidgets 0.32.6
ngff_zarr 0.12.1
numpy 2.2.2
numcodecs 0.15.0
zarr 3.0.1

Having pip-installed the "WindowsWheel311.zip" from the ITKElastix CI, from the last commit at the main, at https://github.com/InsightSoftwareConsortium/ITKElastix/actions/runs/12258642912

Do you have a clue?

[thewtex]

@N-Dekker thanks for the information!

The issue is zarr 3, which was recently released. I'll have to make updates, but for now you could try with zarr 2.

[N-Dekker]

Thanks Matt. I created a new environment, using zarr version 2:

3.11.11 | packaged by Anaconda, Inc. | (main, Dec 11 2024, 16:34:19) [MSC v.1929 64 bit (AMD64)]
itk 5.4.0
itkwidgets 0.32.6
ngff_zarr 0.12.1
numpy 1.26.4
numcodecs 0.14.1
zarr 2.13.3

Successfully moved one cell further, but unfortunately, now it says:

---------------------------------------------------------------------------
GroupNotFoundError                        Traceback (most recent call last)
Cell In[2], line 7
      5 # Least recently used cache -- avoid re-fetching timepoints
      6 lru_store = LRUStoreCache(fs_store, max_size=1e9)
----> 7 ome_zarr = nz.from_ngff_zarr(lru_store, version='0.3')
      9 # 3D time series
     10 print(ome_zarr)

File ~\.conda\envs\ITKElx-zarr2\Lib\site-packages\ngff_zarr\from_ngff_zarr.py:65, in from_ngff_zarr(store, validate, version)
     59 if version and zarr_version_major >= 3:
     60     format_kwargs = (
     61         {"zarr_format": 2}
     62         if packaging.version.parse(version) < packaging.version.parse("0.5")
     63         else {"zarr_format": 3}
     64     )
---> 65 root = zarr.open_group(store, mode="r", **format_kwargs)
     66 root_attrs = root.attrs.asdict()
     68 if not version:

File ~\.conda\envs\ITKElx-zarr2\Lib\site-packages\zarr\hierarchy.py:1374, in open_group(store, mode, cache_attrs, synchronizer, path, chunk_store, storage_options, zarr_version, meta_array)
   1372         if contains_array(store, path=path):
   1373             raise ContainsArrayError(path)
-> 1374         raise GroupNotFoundError(path)
   1376 elif mode == 'w':
   1377     init_group(store, overwrite=True, path=path, chunk_store=chunk_store)

GroupNotFoundError: group not found at path ''

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

N-Dekker commented on 2025-01-27T13:43:42Z
----------------------------------------------------------------

Line #7.    ome_zarr = nz.from_ngff_zarr(lru_store, version='0.3')

from_ngff_zarr returns a Multiscales object, so I would rather name the variable "multiscales" (instead of "ome_zar"). Just like in https://ngff-zarr.readthedocs.io/en/latest/python.html#read-an-ome-zarr

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

N-Dekker commented on 2025-01-27T13:43:42Z
----------------------------------------------------------------

Line #5.    vmin = ome_zarr.metadata.omero.channels[0].window.start

For my understanding, is the index [0] in channels[0] related to the index [0] in images[0]?

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

N-Dekker commented on 2025-01-27T13:43:43Z
----------------------------------------------------------------

Line #2.    view(fixed_image, vmin=vmin, vmax=vmax)

"vmin=vmin" and "vmax=vmax" look non-informative tautologies to me. I would like it better if it would say:

view(fixed_image, vmin=window.start, vmax=window.end)

Assuming window = multiscales.metadata.omero.channels[0].window

[N-Dekker]

My review-notebook-app comments are only minor, stylish. No show-stopper.

I just wish I could make it run here locally!