separate tasks for fetching/extract seed data? location?

[laurelnaiad]

As of now, dbLoad always tries to run fetch/extract of seed data unless those tasks "think" they've already been run.

1. I think it's confusing that the data are kind of buried in the java area, I liked the seed-data directory in database, it was a relatively obvious place for the data to live.

2. How does it know whether we want the files that are there vs. what's in the remote tgz? I'm uncomfortable letting the dbLoad task make the decision because that's beyond the scope of loading what I've prepeared. I think it would be better to have dbLoad fail if the seed-data directory isn't populated with something that resembles samplestack data.

3. Perhaps an uber task can run all of the tasks (basically what cleanstart.sh does), and the individual tasks won't all "force" themselves on the developer. I don't necessarily think this should apply universaly, but at the least I don't think loading the data should be hardwired to preparing it to be loaded, because that is a moment where people are likely to diverge, especially once we have more than one data set prepared for them.

[laurelnaiad]

leaving this untagged until we've had time to bat it around

[popzip]

Somewhat related to #49

[laurelnaiad]

This shouldn't really be assigned to me. It's a gradle thing...