This PR does a few things:

• decode now returns either SourceMapIndex or SectionedSourceMapIndex depending on whether the input has a 'sections' key.

• SectionedSourceMapIndex implements the same interface as SourceMapIndex allowing for the external interaction to be transparent to which type of index is returned

• lookup now returns both a token and the SourceMapIndex the token came from. This is relevant for working with the index around the token especially with SectionedSourceMapIndex as the lookup from SourceMapIndex just returns self.

• Additional API methods were created to make it easier to work with the underlying index without having to reach into the internals:

  • columns_for_line
  • total_number_of_lines
  • files
  • sources_content_map
  • raw_sources