Study period detection

[dogversioning]

This PR makes the following changes:

• Adds a new lambda for processing study periods from library exports
• Creates new SNS topic for differentiating upload tasks
• Light reworking of shared functions to accomodate new processing types
• Some namespace diffentiation between extant metadata and new metadata type

[dogversioning]

explaining the business semantics here: a study period encapsulates all exports associated with a study, so there's only one date range element per study topic. This gets us around 'not every count table will have a date column in it'.

[mikix]

Got some classic Mike nitpicks but looks good!

[mikix]

Binaries. ☹️ It's fine if this makes the most sense. But often in these cases, I try to keep the git-file as a human readable and reviewable one, and then transform it in a setUp method to the binary version.

[dogversioning]

so, this is a copy/paste of the library output, which is what we're expectinging and I think the best test case for "will this actually work". but - i can add a copy of the csv output for the same and use that during the validation step, which will also help to self document.

[dogversioning]

Ok, added and used for validation.