Auto-build missing OME-NGFF multiscale levels (foundation)#43

Open

davidackerman wants to merge 19 commits into

auto-build-multiscale-pyramid

Commits on Jun 12, 2026

Add pyramid_builder: auto-build missing OME-NGFF multiscale levels
davidackerman
committed
Wire pyramid auto-build into Meshify downsample multires
davidackerman
committed
Symlink-safe pyramid cleanup + change pyramid_alignment_mode default to halo
davidackerman
committed
Parallel pyramid cleanup + rename to delete_intermediate_scales
davidackerman
committed
Chunk-file-level parallel pyramid cleanup
davidackerman
committed
Write pyramid zarr via tensorstore instead of optional zarr package
davidackerman
committed
Parallelize super-chunk pyramid build + flush logs eagerly
davidackerman
committed
Memory-aware sizing for pyramid super-chunks
davidackerman
committed
Memory-aware pyramid build with OOM retry (assembly-style)
davidackerman
committed
Stream LOD writes in pyramid build to reduce peak memory
davidackerman
committed
Move pyramid build to dask cluster + smart processes-per-job sizing
davidackerman
committed
Decouple super-chunk shape from on-disk chunk shape + proactive memory sizing
davidackerman
committed
Fix super-chunk sizing: minimum size, not memory-filling
davidackerman
committed

Commits on Jun 15, 2026

Commits on Jun 16, 2026

Commits on Jun 18, 2026

Set lod_scale_multiplier to voxel_size in mesh info files
davidackerman
committed

Commits on Jun 23, 2026

Default to fixed lod_0_box_size + MIN_USEFUL_FACES cap
davidackerman
committed