Closed Niuhao1998 closed 1 year ago
Niuhao - for the BossDB implementation at BossDB.org, we've mostly migrated to using precomputed format for our backend spatial data storage (https://github.com/google/neuroglancer/tree/master/src/neuroglancer/datasource/precomputed).
The BossDB system still supports SPDB, and we're happy to set up a time to walk through deployment with you (please send an email to info@bossdb.org. If you have specific questions about SPDB or BossDB, we can use this forum to discuss (please let us know).
I have seen this system use morton index curve to index the cude. so how to manage a big cube,for example, 102400 102400 7000.Upload and cutout the cube?
Hi Niuhao,
If you'd like to manage your data in BossDB, you can upload it to our public BossDB service (https://api.bossdb.io) using the ingest-client
(https://github.com/jhuapl-boss/ingest-client). This tool will read in your data in supported formats (TIFF, Zarr, PNG, etc) and chunk up the cube into small objects. Our backend service here (SPDB) will assign a morton index to the cuboids.
Once the data is in BossDB, you can access it with intern
(https://github.com/jhuapl-boss/intern), our python client to interact with the BossDB API. You can also access the data directly through the RESTful API: https://docs.theboss.io/reference/introduction
Want to store your data locally in a BossDB-like system? We also offer Bossphorus (https://github.com/aplbrain/bossphorus) to do exactly that!
Let me know if I can help further.
Thanks!
From: dxenes1 @.> Sent: Tuesday, November 22, 2022 10:45 PM To: jhuapl-boss/spdb @.> Cc: Wester, Brock A. @.>; Comment @.> Subject: [EXT] Re: [jhuapl-boss/spdb] how to depoly this system (Issue #34)
APL external email warning: Verify sender @.*** before clicking links or attachments
Hi Niuhao,
If you'd like to manage your data in BossDB, you can upload it to our public BossDB service (https://api.bossdb.io) using the ingest-client (https://github.com/jhuapl-boss/ingest-client). This tool will read in your data in supported formats (TIFF, Zarr, PNG, etc) and chunk up the cube into small objects. Our backend service here (SPDB) will assign a morton index to the cuboids.
Once the data is in BossDB, you can access it with intern (https://github.com/jhuapl-boss/intern), our python client to interact with the BossDB API. You can also access the data directly through the RESTful API: https://docs.theboss.io/reference/introduction
Want to store your data locally in a BossDB-like system? We also offer Bossphorus (https://github.com/aplbrain/bossphorus) to do exactly that!
Let me know if I can help further.
— Reply to this email directly, view it on GitHubhttps://github.com/jhuapl-boss/spdb/issues/34#issuecomment-1324512051, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADNHFBQZUEBJVD23X3JUEH3WJWHJ5ANCNFSM6AAAAAASGOEJTY. You are receiving this because you commented.Message ID: @.***>
Thanks for your reply!I have a question about the parameter in kvio.py. the function of generate_cached_cuboid_keys(). what is the time samples? why we need this parameter to index cubus?
We have the ability to store time-series data with this system, so cuboid objects have a time sample index value in their key. In cases where there is just spatial data (no time-series dimension) the time sample just defaults to [0,1] as seen here:
In the case where 4D data is being uploaded, we assume the first dimension is time. https://github.com/jhuapl-boss/spdb/blob/44d41e2b7a7b961e55746e1a5527d5419a74c2ce/spdb/spatialdb/spatialdb.py#L755
Best, Danny
Thank you for your reply!
I have another question about the boss. Why we use collection,experiment,channel to organize the storage path?
HI Niuhao, The idea behind collection/experiment/channel is you setup an "experiment" to be on a single animal ,like a mouse. Then you could have several channels associated with experiment, like an electron microscopy (EM) channel, an annotation channel, a channel marking specific regions of interest. Neuroglancer can be used to overlay multiple channels. The collection would be a set of experiments you are performing that should be logically be created together. Often times there is only a single experiment under a collection.
Thank you for your reply! I have a question about the boss,what is the channel-layer map? And why need this map?
how to use spdb as part of BossDB and depoly it?