Simplified storage creation with `from_url` helper method.

[jbusecke]

Problem

I find myself repeatedly writing code related to url parsing to set up icechunk storage configurations:

Example (taken from here):

url = "s3://nasa-eodc-public/icechunk/MUR-JPL-L4-GLOB-v4.1-virtual-v2-p2"
url_parsed = urlparse(url)

storage = ic.s3_storage(
    bucket = url_parsed.netloc,
    prefix = url_parsed.path.lstrip('/'),
    from_env=True,
)

Not too terrible, but I would love if this could be simplified, an particularly if the sanitizing (i.e. the stripping of leading slashes above) could happen internally in icechunk.

Proposed Solution

An ideal solution to me would look something like this:

url = "s3://nasa-eodc-public/icechunk/MUR-JPL-L4-GLOB-v4.1-virtual-v2-p2"
storage = icechunk.Storage.from_url(url, **auth_kwargs)

I guess this would principally do two things:

1. Auto detect the storage type from the beginning of the url
2. parse the bucket and prefix from the url

I personally care more about 2. but it would be great to have both.

Is this generally within scope? If yes, id be happy to help out with this.

[ianhi]

This is a great point @jbusecke. So good in fact that there's some prior art here, that ultimately coalesced in ZEP8

Prior issue discussing: #576
rendered: https://zeps--48.org.readthedocs.build/en/48/draft/ZEP0008.html

Which was ultimately included in ZEP8: zarr-developers/zeps#48

ZEP8 in Icechunk: #1161
ZEP8 in Zarr-python: zarr-developers/zarr-python#3567

Does that look like what you are hoping for?

[jbusecke]

Ah yes, thats a great point @ianhi. I knew about zep8, but had not put this together in the moment.
I do think that yes, in particular .create_readonly_session_from_path could satisfy both of my points above.

[paraseba]

I have this for Icechunk 2 #1438