Upgrade to 2.0

[dryan]

This is a near complete code refactor to bring the module back into conformance with the canonical ruby gem.

• All regular expressions updated to match ruby gem
• Additional methods added to validation module
• Rewrite of Extractor to match ruby gem
• Changing how HitHighlighter.hit_highlight works to match ruby gem (see note below)
• Refactor tests.py to use https://github.com/twitter/twitter-text-conformance
• Addresses issues auto_link bug with ending parenthesis #13, Url regex #14, extract_urls_with_indices returns url with comma #15 and latin characters in hashtags breaks the entities extraction #16

The ruby gem changed how HitHighlighter.hit_highlight works. The query parameter from v1.x must now be passed as a named argument and not a positional argument. The positional argument is now an array of arrays of indices indicating where to place the highlight tags, ie [[0,4]]. These indices are based on the text with all HTML tags stripped. Additionally, the option for customizing what HTML tag to use as the highlight has changed to tag.

Tests for Validation are skipped on narrow builds of python (which includes the system and homebrew-installed versions on OS X), though the validation methods will work on these systems.

Parsing extended Kanji characters on narrow build systems works in all tests, but my fail in some cases. Usage of this module for work with extended unicode characters is only recommended on wide builds of python.