Add create_imaris_crops rule for high-resolution atlas region extraction

[Copilot]

Implements a new Snakemake rule that creates Imaris datasets from SPIM data based on atlas region bounding boxes, using zarrnii's get_region_bounding_box(), crop(), and to_imaris() functions. The level wildcard defaults to 0 for high-resolution output.

Implementation

• New rule: create_imaris_crops in patches.smk following the pattern of create_patches rules

  • Takes SPIM zarr, deformed atlas segmentation, and label TSV as inputs
  • Outputs BIDS-compliant Imaris dataset directories with configurable level wildcard

• Processing script: create_imaris_crops.py

  • Extracts bounding boxes per atlas region using ZarrNiiAtlas.get_region_bounding_box()
  • Crops image data via ZarrNii.crop()
  • Saves as Imaris format via ZarrNii.to_imaris()

• Configuration: New CLI options --crop_labels and --crop_atlas_segs (defaults to roi22)

• Target rule: all_imaris_crops generates level 0 crops by default

Skips generating a region if any of the cropped shape is greater than 5000.

Usage

# Generate high-res Imaris crops for specific labels
spimquant /path/to/bids /path/to/output participant \
  --crop_labels Hippocampus Cortex \
  --crop_atlas_segs roi22

Output structure follows BIDS naming:

sub-*/micr/sub-*_seg-roi22_from-ABAv3_level-0_desc-crop_SPIM.imaris/
  └── sub-{subject}_seg-roi22_label-{region}_SPIM.ims

Original prompt

similar to the create_patches rule, make a rule that will create imaris datasets (using zarrnii's to_imaris() function based on bounding boxes from zarrnii atlas get_region_bounding_box, followed by crop(). Level should be a wildcard for this rule but should be 0 by default, since the purpose is to generate high-res crops of imaris data.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.