Text this: Pointwise Prediction and Sequence-Based Reranking for Adaptable Part-of-Speech Tagging