AWS: Investigate why trello-full-backup Python package saves more data

[floehopper]

• This implementation is closely based on our original plain Ruby + Dropbox implementation which only downloads data for each Trello board; it might be worth investigating why the trello-full-backup Python package seems to do more than this

[floehopper]

I think the only significant difference in the data backed up is that the actual card attachments are also downloaded. Another difference is that the data is saved in a more usable form, e.g. separate Markdown files for each card description.

I think our rationale for not saving Trello attachments previously was that we aim to save all important files to Google Drive. I think this is still the case and so I'm happy to move this issue out of the AWS-ify milestone.