Sunset kerchunk and support all of virtualizarr parsers with a new top level API

[betolink]

…solidated API to virtualize() and removed consolidate_metadata()

Description

Addresses #1271 by consolidating consolidate_metadata, open_virtual_mfdatasets and open_virtual_dataset into a single virtualize() API. The user can control which parser earthaccess needs to use but there is a hierarchy: dmrpp first, then HDF then the rest.

We may need another couple methods in case our users decide to use VirtualiZarr directly but don't want to deal with configuring Obstore/Icechunk. Not required for this PR IMO.

---

"Ready for review" checklist

• Open PR as draft
• Please review our Pull Request Guide
• Mark "ready for review" after following instructions in the guide

Merge checklist

• PR title is descriptive
• PR body contains links to related and resolved issues (e.g. closes #1)
• If needed, CHANGELOG.md updated
• If needed, docs and/or README.md updated
• If needed, unit tests added
• All checks passing (comment pre-commit.ci autofix if pre-commit is failing)
• At least one approval

---

📚 Documentation preview 📚: https://earthaccess--1272.org.readthedocs.build/en/1272/

[github-actions]

👈 Launch a binder notebook on this branch for commit 7054bb1

👈 Launch a binder notebook on this branch for commit eab2365

👈 Launch a binder notebook on this branch for commit 67fd706

👈 Launch a binder notebook on this branch for commit ee15996

👈 Launch a binder notebook on this branch for commit 896ac4e

👈 Launch a binder notebook on this branch for commit fd2d152

👈 Launch a binder notebook on this branch for commit bf178d3

👈 Launch a binder notebook on this branch for commit 4eac74b

👈 Launch a binder notebook on this branch for commit 768d81e

👈 Launch a binder notebook on this branch for commit 8dc248a

👈 Launch a binder notebook on this branch for commit c74eaca

👈 Launch a binder notebook on this branch for commit 9a09ae5

👈 Launch a binder notebook on this branch for commit e33e37f

[chuckwondo]

@betolink, this looks great. I have only a few minor suggestions, but I'm approving, as nothing is a show-stopper from my perspective.

[chuckwondo]

The list of granules could be empty, so I suggest handling this case with a pertinent ValueError, otherwise the user will get a less helpful IndexError here.

[betolink]

Done, not sure if the Obstore registry instantiation is the best way to place this but for now... if the granules are not a empty list and they have data inks, in theory they could be virtualized. We can later use a custom "flag" if there is one in CMR about QA e.g. "vritualizable"

[betolink]

@chuckwondo since I updated the code with the granule validation for Obstore, I think I need your approval again =)

[danielfromearth]

Awesome work, @betolink! This is such a good step in making the virtualization process more streamlined :)