node-summary

Sentence Tokenizing has been improved with sentence-tokenizer.

Minor changes here and there, but since this seems pretty solid, I'm setting it as a first release.

Latest changes include:

• Using lodash-node over underscore for speed. I'm more hoping to remove this dependency in the future as we don't need all of the overhead.
• I've prevented the tests from failing by setting the mocha and should.js versions to a static value.
• UTF-8 support.

Enjoy!