* refactor tokenizer more

* #16 refactor around type predictor

* #16 refactor code generation part

* #16 refactor benchmark and related code with better error handling

* #16 refactor benchmark as unit test

* #24 add testcase for code generation

* #16 unit tests for client & bug fix and refactor directory structure

* #16 remove bad typo

* #16 rename core -> src

* #16 Move TokenManager from util to format_prediction

* #16 Move tester and its unit test

* #16 Rename test

* #16 rename run.py -> atc_env and make it configurable and add test

* #16 change module names to more python-ish(?) ones

* #16 Applied AutoPep Whether it is good or not is doubtful though.

* #16 unsupport 3.4 and refactor atc_env -> envgen

* #16 add lang argument and fix test in order to use abs path always regardless where it is called from

* #16 update readme