Wrap up evaluation code

[mnaydan]

• Add missing unit tests
• Add code & unit tests for I/O functionality
• review eval code (RSK)
• discuss /answer questions (LT/RSK)
• update from develop, move span object and poetry detection types into a common location
• add a readme to document evaluation procedure/approach
• make another issue to document the next steps we want to tackle later

[laurejt]

We should likely report some additional page-level result stats such as number of exact matches, partial matches, missed spans, spurious spans

[jerielizabeth]

Refers to #132