Updates examples to newer Icechunk syntax

[norlandrhagen]

PR to address #428.

• Changes the Coiled example to write to Icechunk
• Updates the NOAA append notebook to the newer icechunk syntax.

Blocked by: #421 and the related Zarr-python issue: zarr-developers/zarr-python#2800, so there might be some un-surfaced syntax errors that need addressing after.

@thodson-usgs should we also update your lithops example from the Kerchunk JSON format to Icechunk or would you like to leave it the way it is?

cc @abarciauskas-bgse

[TomNicholas]

This should be unblocked now FYI @norlandrhagen

[norlandrhagen]

The coiled example with icechunk seems to be working, but I might have broken you append example @abarciauskas-bgse.

icechunk: Version: 0.1.2

The store creation and appending all works great until I try to read the store at the end.

IcechunkError: Failed to decode variable 'time':   x error fetching virtual reference
  | 
  | context:
  |    0: icechunk::store::get
  |            with key="time/c/0" byte_range=From(0)
  |              at icechunk/src/store.rs:192
  | 
  |-> error fetching virtual reference
  |-> service error
  |-> unhandled error (PermanentRedirect)
  `-> Error { code: "PermanentRedirect", message: "The bucket you are attempting to access must be addressed using the specified endpoint. Please send all future requests to this endpoint.",
      s3_extended_request_id: "ngBa5W32GDyACUCx8CWH7aPoVJ2ghPOFOiztlHqgRIlp7kn65uAZLDkVQcc4ByM56O5KQTQM0mE=", aws_request_id: "W68YYTHWPYTN8WCM" }
 

Wondering if you've seen anything like this before @TomNicholas or @abarciauskas-bgse before I dig in further.

Also, I think the syntax moved or changed. virtual_ref_config/VirtualRefConfig

[TomNicholas]

I don't recognize that error - any insight @mpiannucci ?

[mpiannucci]

@norlandrhagen Take a look at the latest docs here https://icechunk--720.org.readthedocs.build/en/720/icechunk-python/virtual/

The syntax for virtual refs has changed most likely

[norlandrhagen]

The syntax for virtual refs has changed most likely

Thanks @mpiannucci! That fixed it.

[norlandrhagen]

I think this is G2G!

[TomNicholas]

This looks good, except that in the append notebook, then because you didn't actually run the cell that commits the changes, the reading doesn't yet show the newly appended data.