linumpy.geometry.resampling#

Mosaic grid resampling utilities.

Consolidated from linum_resample_mosaic_grid.py.

Functions#

resolution_is_mm(source_res)

Heuristic: source resolution in mm if all components < 1, otherwise µm.

resample_mosaic_grid(vol, source_res, target_res_um[, ...])

Resample a mosaic grid volume to a target isotropic resolution.

Module Contents#

linumpy.geometry.resampling.resolution_is_mm(source_res)[source]#

Heuristic: source resolution in mm if all components < 1, otherwise µm.

Used across the pipeline to accept either unit in OME-Zarr metadata or CLI arguments without breaking legacy data. Pixel sizes below 1 µm would imply sub-nanometre voxels, so the heuristic is safe for all realistic acquisitions.

Parameters:

source_res (tuple | list)

Return type:

bool

linumpy.geometry.resampling.resample_mosaic_grid(vol, source_res, target_res_um, n_levels=5, out_path=None)[source]#

Resample a mosaic grid volume to a target isotropic resolution.

Processes tiles individually to avoid loading the entire mosaic into memory. Uses anti-aliasing and 1st-order interpolation.

Parameters:

  • vol (dask array or zarr array) – Mosaic grid volume with chunk structure (each chunk = one tile). Shape: (Z, nx*tile_h, ny*tile_w)

  • source_res (tuple) – Source resolution (res_z, res_y, res_x) in whatever unit.

  • target_res_um (float) – Target isotropic resolution in microns.

  • n_levels (int) – Number of pyramid levels in output.

  • out_path (str or None) – If provided, save the result to this OME-Zarr path.

Returns:

Resampled array if out_path is None, else None (writes to file).

Return type:

np.ndarray or None