Ukrainian Stanford Question Answering Dataset

[IronTony-Stark]

First thing first, I was not able to setup GoogleDrive as our source for downloading data. I may being dummy, but I just don't see a way to download a file without Auth0. Getting auth token is not that hard, but the real question is where to store it. I may need a deeper investigation on how python packages work and if they can contain secrets. I also tried to store our data in DropBox, but the problem is the same. So I ended up uploading it to GitHub repo :) We can later move it to some of our repos. The only thing is that such repository has to be public. Private would once again require auth token

P.S. And I'll probably need to refactor UaSquadDataset.parse sometime later

[KyloRen1]

LGTM, I thought of creating JSON first and then uploading it somewhere

[Smfun12]

Everything seems fine, unless two things:

1. _check_exists method - I would rather rename it to file_exists, so the line if self._file_exists(): is more readable
2. Why do you call self.parse(self.dataset_path), if parse is static(no need for self)?