Switch from pathy to cloudpathlib

[adrianeboyd]

Description

Switch from pathy to cloudpathlib for remote storage with spacy project push/pull.

• Continue using smart_open for downloads, which are also used for project assets.
• Switch to cloudpathlib for uploads.
• Replace deprecated ignore_ext flag with compression and add corresponding smart_open version requirements.
• Add tests for local filesystem remote storage.
• Add functionality for sorting multiple content hashes by last modified.

Types of change

Enhancement.

Checklist

• I confirm that I have the right to submit this contribution under the project's MIT license.
• I ran the tests, and all new and existing tests passed.
• My changes don't require a change to the documentation, or if they do, I've added all required information.

[polm]

One question about the changes but looks good.

[adrianeboyd]

The test should probably be marked slow in the end.

[polm]

It looks like #11780 introduced a minor conflict in the docs, so I went ahead and resolved that.

Besides that the most recent change looks fine.

[adrianeboyd]

I'll close this for now since the issues with pathy were resolved. We can always come back to it if we need to...

[BLKSerene]

Hi, I noticed that spaCy 3.7 introduced weasel as a dependency which depends on cloudpathlib instead of pathy. Would it be more consistent and easier for the dependency management of spaCy to switch to cloudpathlib as well?

[adrianeboyd]

Yes it would, thanks for the note! Neither of them currently support python 3.12, however.

[adrianeboyd]

Actually, Pathy should have just been removed from spacy, it's no longer used.