Add argument to preserve unicode characters in json output.

[wayneworkman]

Here's a snippit from the Anthropology article before this code change using the --json argument.

Their New Latin ' derived from the combining forms of the Greek words \"\u00e1nthr\u014dpos\" (, \"human\") and \"l\u00f3gos\" (, \"study\").

Here it is after these code changes, using --json and --preserve-unicode

Their New Latin ' derived from the combining forms of the Greek words \"ánthrōpos\" (, \"human\") and \"lógos\" (, \"study\").

For text computing activities, it's nice to have the option to preserve these unicode characters in their true form, rather than ASCII representations.

[wayneworkman]

Thinking more on this pull request, when you load the JSON output text using json.loads() in python, the unicode is represented correctly in the loaded JSON. Given that, I think this pull request is only valid for a few scenarios:

• When you're not loading the JSON output with a JSON tool; i.e. parsing it manually (not a great idea).
• When you want to view the output manually as a sanity check to see what the source text looks like.
• When you're using the JSON output for personal reading purposes.