search

seung-lab / cloud-volume

Read and write Neuroglancer datasets programmatically.
https://twitter.com/thundercloudvol
BSD 3-Clause "New" or "Revised" License
132 stars 47 forks source link

what kinds of sharded operation is supported now? #558

Open xiuliren opened 2 years ago

xiuliren commented 2 years ago

Hi Will,

I would like to upload some synapse detection results to cloud storage. Is it supported in cloud-volume?

I noticed that there is documentation to upload sharded image chunks on the wiki page. what else is supported to upload? It seems that skeleton and synapse point, line segments are not supported yet.

william-silversmith commented 2 years ago

Hi Jingpeng! Currently images, meshes, and skeletons support mesh uploads from Igneous. Images use a simpler method for determining where each chunk goes (a block of adjacent chunks all go in the same shard), so easy user uploads are supported.

Segmentation images are supported as you mentioned. Point annotations are not yet supported (at all) in CloudVolume. I'm expecting to get to them after I finish submitting the Igneous paper. I tried skimming the Neuroglancer spec for them just now, but I think these objects work sufficiently differently from the others that it will take some study for me to understand them. It will probably be a few months before they are fully supported.