grep-like functionality

[giftig]

Being able to search through an S3 file while in the s3-browser cli would be beneficial in some cases. To stick with the unix-like interface it'd be nice to make this a grep command. This could either be stripped-down functionality like simply supporting fgrep [literal text] [key] / grep [regex] [key] or could simply use grep itself to filter the contents (where available)

Either way it should perform a binary check first by using utils.print_object like the cat command currently does.

[giftig]

A simpler solution here might just be to open the file in the user's text editor of choice as a quick option, since otherwise this would open the door to wanting other tools like jq for json, etc.