Rnlp

This update patches an error caused by where the `sentenceID` was incremented. Previously it was incremented before iterating through the words, so the `wordInSentence` predicate would occasionally refer to sentences that did not exist.

This release does more behind-the-scenes than in front, changes include:

**Bugfixes**

* Changed from using `.strip(_punctuation)` to a version of `translate`. strip would remove leading and trailing punctuation marks, which would occasionally lead to internal errors during BoostSRL's parsing step.
* Fixed a bug where the `lateSentenceInBlock` predicate was not generated properly during makeIdentifiers.

**Minor changes**

* `rnlp.makeIdentifiers` is now imported by default when `rnlp` is imported.
* Added unit tests (now under `rnlp/tests/rnlptests/`) and integrated with Travis-CI and CodeCov.io
* Aligned the modules with the pep8 style guide.

Release `0.2.0` had a set bug caused by overlapping versions and a "cannot find README.md" error.

rnlp` may now be used fairly interchangeably as either a package or a commandline program. A `tqdm` progress bar, parameterized options, and a few other quality-of-life features were added.

Official first release for `rnlp`.