bib/clean.py

#!/usr/bin/python

import sys
import bibtexparser
import bibtexparser.bparser


def remove_newlines(d):
    for key, value in d.items():
        d[key] = value.replace('\n', ' ')
    return d


class AlignedWriter(bibtexparser.bwriter.BibTexWriter):
    def __init__(self):
        super(AlignedWriter, self).__init__()
        self.order_entries_by = None

    def _entry_to_bibtex(self, entry):
        bibtex = u'@{}{{{}'.format(entry['type'], entry['id'])

        for field in [i for i in sorted(entry) if i not in ['type', 'id']]:
            try:
                bibtex += u",\n  {:<12}= {{{}}}".format(field, entry[field])
            except TypeError:
                raise TypeError("The field %s in entry %s must be a string" % (field, entry['id']))

        bibtex += u"\n}}\n{}".format(self.entry_separator)
        return bibtex.encode('UTF-8')


def main():
    if len(sys.argv) < 2:
        sys.exit("Usage: clean.py file.bib")

    with open(sys.argv[1]) as f:
        parser = bibtexparser.bparser.BibTexParser()
        parser.customization = remove_newlines

        db = bibtexparser.loads(f.read(), parser)

        writer = AlignedWriter()
        print(bibtexparser.dumps(db, writer))


if __name__ == '__main__':
    main()