Create a user interface that runs from command line

[mbjackson-capp]

High priority -- should come before resolving edge cases. Should work something like as follows:

• Ask user "Do you want to use backup .jsons from QBReader, a folder of packets, or a live DB query?"
  • allow for command line flags for each option (and category filter stuff)
• If backup .jsons:
  • check that tossups.json and bonuses.json are in the relevant folder
    • if not: tell user where to find them
  • run backup_to_cards() stuff
  • ask user if they want to filter categories/difficulty/year
• If folder of packets:
  • run through that folder, skipping over invalid packets
  • category filtering isn't really possible here without writing a classifier of some kind and running every clue through it
  • maybe give user a chance to set tags for all cards?
• If live query or backup csv:
  • ask user to filter categories, difficulty, year, etc. first (or pull in from command line)
  • if live query (See #11 ):
    • query the db, using a for loop to iterate over pages if necessary
  • if backup csv:
    • read in the csv at filepath, using pandas selectors to filter down to desired components
• run cleanup() stuff on the .json objects returned

Then:

• ask user if they want to do fuzzy redundancy removal (with warning that it could take hours for large dataframes)
• write to file

[mbjackson-capp]

In progress in interface.py