Sentiment-aware environment

[settinghead]

Goal: incorporate CoreNLP into an environment that we can use.

What you probably need: CoreNLP Server, and the python code that connects to and calls the server (see sentiment_env.py as an example).

[drannielee]

Hey @settinghead the title and the comment doesn't match?
Why do you need to call CoreNLP? Stanford NLP doesn't have the best sentiment analyzer; should just use an off-the-shelf pre-trained model.

[settinghead]

@drannielee I'm not sure. @alicerueda showed me how to assess sentiment with CoreNLP Server (the code is also in the repo). At the end of the day we just need an environment that can hold a state and give rewards based on actions. So any library we use is fine with me.