Explore and view Neuroglancer Precomputed image volumes

[frcroth]

URL of deployed dev instance (used for testing):

• https://precomputed.webknossos.xyz/

Steps to test:

• Explore a dataset
  • https://storage.googleapis.com/neuroglancer-fafb-data/fafb_v14/fafb_v14_clahe/
  • https://storage.googleapis.com/neuroglancer-fafb-data/fafb_v14/fafb_v14_orig/
  • https://storage.googleapis.com/neuroglancer-public-data/kasthuri2011/image_color_corrected/
• View the dataset
• Compare with neuroglancer
  • fafb_v14_clahe
  • fafb_v14_orig
  • kasthuri2011

Issues:

• fixes Read neuroglancer precomputed #6620

Follow-up Issues

• Sharded (Support sharding for remote datasets #6823)
• Compressed data format compressedseg (Support compressed segmentation for Neuroglancer precomputed datasets #6820)
• Handle layers with different voxel sizes/mags (Support adding multiple layers with different voxel sizes separately #6821)
• Meshes (Support loading meshes from remote neuroglancer datasets #6826)

---

• Updated (unreleased) changelog
• Updated documentation if applicable
• Ready for review

[fm3]

This already looks really good! 🎉 I’m looking forward to using it :) I added a few comments, but nothing major. I’ll test this after you had a look at the suggestions. Could you also add a changelog entry and provide some steps to test?

Also, what is Fix janelia_flyem_hemibrain about in the PR description?

Concerning the compressedseg: do you know how this is solved for javascript? If writing java bindings is complicated, maybe we can do this as a separate PR in order not to keep this one open for too long.

Let’s talk about segmentation layer path issues in person

[frcroth]

Compressed segmentation TS implementation can be found here: https://github.com/google/neuroglancer/tree/master/src/neuroglancer/sliceview/compressed_segmentation, the one used in wk-connect is here: https://github.com/seung-lab/compressedseg

[fm3]

@philippotto could you please have a quick look at the front-end changes?

[philippotto]

Front-end looks good 👍 Didn't test anything, though.

[fm3]

🎉 💯