write tests for file data extracting

I've written out a number of sample files to test

todo:

• implement tests that use the file-data-samples/ files
  • figure out what malformed tags should look like
  • add malformed dates to the malformed dates file
• write out the programmatic file data tests
  • generate a huge file, with date, tags embedded
  • generate a file with a huge number of tags
  • malformed utf8 we know what will happen: there will be an io::Error at that point. this is documented on e.g. BufRead::read_line(). i've put in a todo comment that the function could be made to return the current state of the data on error.

this is part of #5