You might find some parallels to the field of citation field extraction - where citation/bibliography strings are parsed into author, affiliation, conference name, page/volume of a journal the article appeared in, etc. It is a well-studied problem in NLP and CRF-based taggers have known to yield the best results.