v1.1.0

What's Changed

• Read .zarr folders and plain SpatialData stores in the serverless viewer by @sminot in #5

Full Changelog: v1.0.2...v1.1.0

v1.0.2

Full Changelog: v1.0.1...v1.0.2

v1.0.1

Fix lock_view: deck's camera controller must live on the view, not th…

v1.0.0

Full Changelog: v0.1.10...v1.0.0

v0.1.10

Full Changelog: v0.1.9...v0.1.10

v0.1.9

Full Changelog: v0.1.8...v0.1.9

v0.1.8

Full Changelog: v0.1.7...v0.1.8

v0.1.7

Full Changelog: v0.1.6...v0.1.7

v0.1.6

Full Changelog: v0.1.5...v0.1.6

v0.1.5

Published checkpoint filenames changed

The Nextflow workflow now gives every dataset its own folder, with the same names inside
each one:

results/<dataset path>/
  results-<content hash>.sdata.zarr.zip   # the full checkpoint
  lowres-<content hash>.sdata.zarr.zip    # the image-capped copy
  plots/<NN>_<namespace>.<function>/figure.{svg,pdf}
  results.log

Before this, a dataset's two checkpoints, its log and its plots were siblings named after
the folder it was found in — folderB.sdata.zarr.zip, folderB.sdata.lowres.zarr.zip,
folderB.log, folderB.plots/ — sharing a directory with any neighbouring dataset.

Fixed filenames no longer identify the dataset, so two other things do:

• The content hash is the suffix the app puts on its own saves, so opening a published
  file reports whether its bytes still match the name it was published under.
• The dataset's path is recorded inside each checkpoint (app_state["name"]), so
  reopening either file shows the dataset it holds rather than the word "results".
  backend/cli.py gained --session-name and --lowres-name, and the New Session →
  Load dialog no longer overrides a checkpoint's recorded name with one derived from its
  filename.

What this affects

Anything that consumed the old published paths by name — a script globbing
*.sdata.lowres.zarr.zip, a link into a published tree — needs updating. index.json is
unchanged in shape and still lists every checkpoint with its label, so a served
deployment keeps working without one.

The checkpoint format itself did not change (no viewer/ sidecar or schema change),
so viewer builds either side of this tag read these files: this is a filename change, not
a compatibility break. The layout comes from the workflow revision you run, not from the
viewer bundle — nextflow run … -r v0.1.4 still produces the old layout, and pinning
--viewer_dist to an earlier release only pins the SPA. Cirro runs the workflow straight
from the repository, so runs launched after this landed on main publish the new layout.

Full Changelog: v0.1.4...v0.1.5