Fix bibtex parser

[Nagasaki45]

Get the entry that caused pybibs to fail from Ben.

[Nagasaki45]

Here's an example entry:

@ARTICLE{2019arXiv190304691L,
   author = {{Law}, C.~J. and {Margalit}, B. and {Palliyaguru}, N.~T. and 
	{Metzger}, B.~D. and {Sironi}, L. and {Zheng}, Y. and {Berger}, E. and 
	{Margutti}, R. and {Beloborodov}, A. and {Nicholl}, M. and {Eftekhari}, T. and 
	{Vurm}, I. and {Williams}, P.~K.~G.},
    title = "{Radio Time-Domain Signatures of Magnetar Birth}",
  journal = {arXiv e-prints},
archivePrefix = "arXiv",
   eprint = {1903.04691},
 primaryClass = "astro-ph.HE",
 keywords = {Astrophysics - High Energy Astrophysical Phenomena},
     year = 2019,
    month = mar,
   adsurl = {http://adsabs.harvard.edu/abs/2019arXiv190304691L},
  adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

The issue is with the multi-line key / value pairs. pybibs wrongly assume one per line.