Provide a base-format only mode? e.g. don't identify specific JSON flavors when JSON is enough

[ross-spencer]

Given there are going to be a lot of different objects out there written using JSON/YAML/TOML and so on, should there be an option to just provide a more concrete identification of JSON without forwarding the data to identification by ruleset?

This might be quicker for some folk, and also it might be all the accuracy they need. It would capture different encodings and identify formats to a high level of reliability. Ruleset identification could be used later on if it was decided there was a benefit.

NB. please comment on naming if base-format doesn't make sense. There are probably better names.