Implement smart GitHub crawl

[henricasanova]

Currently, it takes a while to populate the local database based on crawling the GitHub repo. Instead, we should be smart, and detect that when a workflow hasn't changed since last time we retrieved it then we don't need to retrieve it again. This could entail a redesign, where the WfCommon instances are actually "git pulled" from the WfCommons github, as opposed to curl-ed. This would entail quite a few changes, where instead of always re-downloading from the github, one would instead read a local file. And a cront job would be set up to do a git pull every hour for instance.