Add backfill study function

[clementzach]

This PR adds a helper function that allows a user to easily download data from a whole study.

One question: Is there a reason that BACKFILL_WINDOW is set to be 5? It seems like that slows down the process of downloading data a good amount.

In my experience, we should be able to have a default value of BACKFILL_WINDOW that is at least 100 or so.

[biblicabeebli]

@tokeetim would you like me to do some review here to speed the process along?

[tokeefe]

@biblicabeebli Yes, please! Thank you!

[biblicabeebli]

Crap I just finished the modernizing/cleanup/python-3.8-fixes refactor I was working on and then looked at this. I totally forgot this pull existed.

@clementzach the refactor for the sync file was just to move it to a file called sync (should maintain import compatibility) from the /sync/__init__.py. I could probably pull in your code additions in manually but that then it wouldn't be attributed to you. Would you like me to do that (because it's been..... a while....) or would you like to handle the merge/rebase?

I'm still new to this codebase, I will start using it, I don't know why a longer window would speed up/slow down requests. One change I did make internally that may affect speed was to up an internal chunking value. Otherwise, I intend to try and get into this code a little more for speed and, eventually, a v2 download endpoint that pulls actually-compressed data.

(it was once compressed! ... but there was a server-destroying memory not-exactly-a-leak-but-kinda-a-leak and I had to turn it off...)