Dataset for example live script 000011 is heavy

[aranega]

It seems that the dataset on DANDI have been reorganized a little bit, and the dataset DANDI:000011 is now more than 30Gb. I checked if there was a way to download only sub-parts of a dataset, but I wasn't able to find a way to achieve that using the dandi cli. I'm not sure it's possible. In this situation, does it make sens to keep the live example for 000011?

[rly]

We can use the dandi CLI to download a single NWB file given a known path. See dandi/dandi-cli#1434.

Full help for dandi download

  Accepted resource identifier patterns:
   - DANDI:<dandiset id>[/<version>]
   - https://gui.dandiarchive.org/...
   - https://identifiers.org/DANDI:<dandiset id>[/<version id>] (<version id> cannot be 'draft')
   - https://<server>[/api]/[#/]dandiset/<dandiset id>[/<version>][/files[?location=<path>]]
   - https://*dandiarchive-org.netlify.app/...
   - https://<server>[/api]/dandisets/<dandiset id>[/versions[/<version>]]
   - https://<server>[/api]/assets/<asset id>[/download]
   - https://<server>[/api]/dandisets/<dandiset id>/versions/<version>/assets/<asset id>[/download]
   - https://<server>[/api]/dandisets/<dandiset id>/versions/<version>/assets/?path=<path>
   - https://<server>[/api]/dandisets/<dandiset id>/versions/<version>/assets/?glob=<glob>
   - dandi://<instance name>/<dandiset id>[@<version>][/<path>]
   - https://<server>/...

[aranega]

@rly Indeed, I saw that we can have a specific path when using the server or instance name, but I would have prefered to avoid to have hardcoded somewhere an URL toward a server, or an instance name. If there is no other choice, then I'll push a PR with this infromation hardcoded in the setup.m file.

[rly]

I think it would be OK to hardcode the instance name ("DANDI") for now.

For all sessions of one subject:

dandi download dandi://DANDI/000011@0.220126.1907/sub-255201/

or for one NWB file:

dandi download dandi://DANDI/000011@0.220126.1907/sub-255201/sub-255201_ses-20141124_behavior+ecephys+ogen.nwb