Ask HN: Text comprehension algorithms/libraries?

Can anyone recommend a starting point towards writing lightweight software that can recognize if an article discusses one specific thing? It doesn't have to be full human level comprehension of the entire text, but something better than a keyword search too.

For examples, software that can understand if an article states whether a) A company is repurchasing its own shares b) When Apple or Samsung is releasing its next phone c) Trump or Hillary is leading the polls.

I know there is the NLTK library for Python, but am wondering if that is overkill for what I'd like to do.

I'm not a software engineer and have no formal CS training, so I'm hoping for pointers to accomplish this goal in the simplest way.

That said, if there is no simple way, I'd be open to learning more difficult algorithms and/or machine learning. Really just trying to be time-efficient, rather than avoid the hard stuff.

Thanks!