dvpio.read.image.read_czi

Contents

dvpio.read.image.read_czi#

dvpio.read.image.read_czi(path, chunk_size=(10000, 10000), channels=None, scene=None, timepoint=0, z_stack=0, **kwargs)#

Read .czi to Image2DModel

Uses the CZI API to read .czi Carl Zeiss image format to spatialdata image format.

Parameters:

  • path (str) – Path to file

  • chunk_size (tuple[int, int] (default: (10000, 10000))) – Size of the individual regions that are read into memory during the process.

  • channels (int | list[int] | None (default: None)) – Defaults to None which automatically selects all available channels. Passing the numeric index of a single or multiple channels subsets the data to the specified channels.

  • scene (int | None (default: None)) – Index of the scene to read. If None (default), all scenes will be considered. If specified, only subblocks of the specified scene contribute to the parsed image.

  • timepoint (int (default: 0)) – If timeseries, select the given index (defaults to 0 [first])

  • z_stack (int (default: 0)) – If z_stack, selects the given z-plane (defaults to 0 [first])

  • kwargs (Mapping[str, Any]) – Keyword arguments passed to spatialdata.models.Image2DModel.parse()

Return type:

Image2DModel

Returns:

spatialdata.models.Image2DModel

Example

We can read czi images with a very simple API

from dvpio.read.image import read_czi

czi_path = ...
read_czi(czi_path)
# > <xarray.DataArray 'image' (c: 2, y: 1440, x: 21718)> Size: 125MB

Note that you can also select subsets of the data that you would like to read. Currently, the function supports reading specific channel indices, scenes (regions of interest), timepoint indices, or z-stack indices. This might significantly reduce the storage demands of your data

czi_path_multi_scene = ...
# Only read the first scene
read_czi(czi_path_multi_scene, scene=0)
# > <xarray.DataArray 'image' (c: 2, y: 1416, x: 1960)> Size: 11MB

You can pass additional keyword arguments to spatialdata.models.Image2DModel.parse(). For example, to generate a pyramidal image for overall faster data access, pass the scale_factors argument

# Create a pyramidal data representation
read_czi(czi_path_multi_scene, scale_factors=[2, 2, 2])
# > <xarray.DataTree>
#   Group: /
#   |-- Group: /scale0
#   |-- Group: /scale1
#   `-- Group: /scale2
Contents