Interactive result comparison

[bauersimon]

• see list of solved tasks at a glance (similar to a JUnit html report)
  • including coverage count
  • or encountered errors
  • and any additional metrics we (will) collect (processing time, character count, ...)
  • with possibility to click and see the actual result (i.e. what the model generated)
• diff/compare two model results
  • on a list basis (i.e. which tasks was which model able to solve and where are differences in coverage: character count, ...)
  • on a task basis (i.e. diffing the model results 1:1)
  • maybe also the ability to switch between different results from different runs (i.e. compare plain.go result from model A's run 2 against plain.go result from model B's result 4)