Add encoding support to `files.count_lines()`.

This seemed to be necessary when working with a file with 'latin_1'
encoding. Hopefully this is a good change and doesn't break anything..?