* Added python 3.5 to travis * Moved `SentenceTokenizer` to its own file * Added `dept` to `abbrev_types`
0.0.6
------
* [FIX] Sgt, Gov, and No abbreviations accounted for
0.0.5
-----------------
* [FIX] Python2.7 unicode errors * Updated PunktLanguageVars to detect special quotation marks
0.0.4
-----------------
* [FIX] PYPI didn't pick up the new training data in JSON format
0.0.3
-----------------
* [FIX] Tokenizer would think a custom abbreviation, F.B.I., was a sentence break * [FIX] Added sanitizer module to preprocess text before summarizing it * [TESTS] Added ability to pull valid tokenized articles from brevity.detroitnow.io and test that the new tokenizer is still valid