feat: adds Resource.hash in datapackage.json

[dangotbanned]

Related

• https://datapackage.org/standard/data-resource/#hash

Description

This PR adds a hash for each dataset, piggybacking directly from the GitHub trees API.

More info

vega-datasets/scripts/build_datapackage.py

Lines 520 to 547 in 062ee8a

	"""
	Use `Get a tree`_ to retrieve a hash for each dataset.
	Parameters
	----------
	ref
	The SHA1 value or ref (`branch`_ or `tag`_) name of the tree.
	api_version
	The `GitHub REST API version`_.
	Returns
	-------
	Mapping from `Resource.path`_ to `Resource.hash`_.
	.. _Get a tree:
	https://docs.github.com/en/rest/git/trees?apiVersion=2022-11-28#get-a-tree
	.. _branch:
	https://github.com/vega/vega-datasets/branches
	.. _tag:
	https://github.com/vega/vega-datasets/tags
	.. _GitHub REST API version:
	https://docs.github.com/en/rest/about-the-rest-api/api-versions?apiVersion=2022-11-28
	.. _Resource.path:
	https://datapackage.org/standard/data-resource/#path-or-data
	.. _Resource.hash:
	https://datapackage.org/standard/data-resource/#hash
	"""

Mostly just upstreaming existing code from https://github.com/vega/altair/blob/fdffed0a15be3967c6b9513787fd40feb59c9cdc/tools/datasets/github.py#L145-L159 with some tweaks.

I think this is a broadly useful inclusion to datapackage.json, since we have 74 datasets that have very few revisions.
I've been using this in vega/altair#3631 for caching datasets across versions - and found that as of vega-datasets@v2.11.0 there have only been 115 unique hash values.

Moving this logic here will greatly simplify (vega/altair#3631), as the hash is the last bit of metadata I'm currently not able to get from datapackage.json.
I've been planning out some revisions to get that PR over the line, this will let me remove most of (https://github.com/vega/altair/tree/fdffed0a15be3967c6b9513787fd40feb59c9cdc/tools/datasets) since I no longer need to collect any metadata from multiple endpoints

[dangotbanned]

@dsmedia ah you saw this early 😉

I've still got some docs to do, and then filling out the description (this PR simplifies altair.datasets significantly)