Use partial clone strategy

[thatbudakguy]

Cloning a large number of repositories with their full history is an expensive operation, both in terms of network and disk/inode usage. In most cases we don't need the full history of changes, especially since the user can find them on github, so we can use a partial strategy to save on resources.

In addition to the classic --depth=1 ("shallow clone") strategy there's now apparently something called a "blobless clone" that saves on resources but doesn't sacrifice as much as removing the history entirely. For more info, see this article on partial clone strategies.

Related to #155 which might support this more easily.

[thatbudakguy]

Apparently we already do a --depth=1 clone:

GeoCombine/lib/geo_combine/harvester.rb

Line 104 in 8048b75

Git.clone(repo_url, nil, path: ogm_path, depth: 1)