Tutorial Request: S3-only usage

[CodyCBakerPhD]

One of several issue tickets @satra requested to be opened for tracking purposes

@yarikoptic came up with this request originally - to have a notebook demonstrating how to use the S3 bucket directly to perform various operations usually delegated to the DANDI API

Would include:

• fetching blob download IDs/URLs from asset metadata files
• navigating Dandisets directly from the metadata files
• downloading a Dandiset version using something like s5cmd instead of the DANDI CLI

Happy to take this one myself as I've done all these operations before internally

[yarikoptic]

I think it might even a good excercise for some claude code or another coding agent to take and existing notebook and perform extension to

explore content and structure of s3://dandiarchive bucket and in particular dandisets/ blobs/ and zarr/ prefixes, prepare an overview and take notebook X which performs IO operations on DANDI archive via dandi-cli and dandi REST API, and convert to directly access dandiset manifest json from S3 bucket which gives access to the listing of assets and corresponding blobs and zarrs, and then loading nwb's directly from s3.