tiff-as-virtual-zarr.js

Read OME-TIFF data via the Zarr store interface, with "virtualized" OME-NGFF metadata. In other words, given an OME-TIFF file, we want to conceptually/virtually map the image tiles and pyramid resolutions to Zarr chunks and OME-Zarr multiscales. We translate OME-XML metadata to OME-NGFF metadata.

For now, we only support the single-file OME-TIFF use case, and we translate metadata to OME-NGFF v0.5.

The implementation was originally based on Viv's TiffPixelSource (696f1d52) but is now based on Vizarr's OmeTiffStore.

Development

pnpm install

Run unit tests:

pnpm run test

Publish:

pnpm publish --filter "tiff-as-virtual-zarr"

Virtual tiff-to-zarr mapping

/zarr.json (Zarr v3) attrs for root metadata

This should list the dimensions, multiscales, and channels according to the OME-NGFF metadata schema.

Image array /{scale}/zarr.json

This should list the shape and dtype of the array and other properties such as compression parameters.

Tile chunk /{scale}/{tile_index}

This key identifies a particular tile in the multidimensional array representing a particular scale of a multiscale image.

Related work

• https://github.com/hms-dbmi/vizarr/pull/254
• https://github.com/cgohlke/tifffile/blob/976465b328f7724041b230917f65ca29a3e2294b/tifffile/zarr.py#L235
• https://github.com/zarr-developers/VirtualiZarr
• https://github.com/virtual-zarr/virtual-tiff and #32
• https://github.com/pymmcore-plus/ome-writers
  • https://github.com/tlambert03/ome-writers/blob/6f14f841700a0586f57e9e518123baa4756a497b/src/ome_writers/_backends/_live_tiff_store.py
• https://github.com/keller-mark/hdf5-as-virtual-zarr.js

Citation

Cite the Viv paper if you use this code:

Trevor Manz, Ilan Gold, Nathan Heath Patterson, Chuck McCallum, Mark S Keller, Bruce W Herr II, Katy Börner, Jeffrey M Spraggins, Nils Gehlenborg, "Viv: multiscale visualization of high-resolution multiplexed bioimaging data on the web." Nature Methods (2022), doi:10.31219/10.1038/s41592-022-01482-7