Perform unicode migration for reviewboard/webapi
This change represents an audit of all of our string literals to make sure that

bytestrings are annotated with a leading 'b', and then adds

"from __future__ import unicode_literals" to the beginning of every non-empty

file.
This also includes a number of changes to use six.text_type for python 2/3

compatibility, using 'with' for file opens, as well as some clean-ups to use

raw strings when testing escaping.

Ran unit tests.