This mostly moves logic out of rake tasks and into dedicated classes, which makes it easier to test. It also separates the harvesting portion from the indexing portion; this might make it easier to have the geoblacklight harvester inherit some of this behavior later. The denylist is one of the things the harvester class includes, which closes #119. I think the logic mentioned in #112 was already in place, so this closes #112 too.
It also adds logic to filter indexing based on your desired schema version (set SCHEMA_VERSION to Aardvark to index only aardvark records; the default is 1.0), which fixes #130.
I added a dependency on ruby-git to do the git pull/clone instead of just calling out to system(), which made things easier to test and might also allow for things to run on windows.
This mostly moves logic out of rake tasks and into dedicated classes, which makes it easier to test. It also separates the harvesting portion from the indexing portion; this might make it easier to have the geoblacklight harvester inherit some of this behavior later. The denylist is one of the things the harvester class includes, which closes #119. I think the logic mentioned in #112 was already in place, so this closes #112 too.
It also adds logic to filter indexing based on your desired schema version (set
SCHEMA_VERSIONtoAardvarkto index only aardvark records; the default is 1.0), which fixes #130.I added a dependency on ruby-git to do the git pull/clone instead of just calling out to
system(), which made things easier to test and might also allow for things to run on windows.