Improve how the KnownPubsDB gets updated

[tnabtaf]

Right now, the Known Pubs DB (a TSV spreadsheet) is updated in two ways:

1. By running PubSpork. This updates it with information from the latest alerts, and from the relevant pubs library itself.
2. By opening it in a spreadsheet program and manually updating the state and annotation columns. These columns are set when a publication is irrelevant, or is relevant, but we just can't curate it yet.

I'd really like to improve that manual updating step. I'm just not sure how.

In my ideal world, the generated HTML report (that drives the curation process) would also include forms/pulldowns/whatever to flag a paper as irrelevant and say why. We wouldn't need to open and manually update the spreadsheet.

This might requiring moving the DB to, say, SQLite. I like having the spreadsheet, but I could adapt to SQLite too.

Suggestions (and fixes!) are encouraged.