clean.py

#!/usr/bin/env python

import sys

import bibtexparser
from bibtexparser.bparser import BibTexParser


FIELDS_TO_KEEP = {
    '*': ['title', 'author', 'year', 'pages', 'ENTRYTYPE', 'ID'],
    'article': ['volume', 'number', 'journaltitle', 'shortjournal']
}


def main(bibtex_file):
    with open(bibtex_file) as f:
        bib = bibtexparser.load(f, BibTexParser(ignore_nonstandard_types=False))
    for entry in bib.entries:
        _add_year(entry)
        _filter_fields(entry)
    print(bibtexparser.dumps(bib))


def _add_year(entry):
    if 'year' not in entry and 'date' in entry:
        entry['year'] = entry['date'][:4] 


def _filter_fields(entry):
    fields_to_keep = FIELDS_TO_KEEP['*'] + FIELDS_TO_KEEP.get(entry['ENTRYTYPE'], [])
    for field in list(entry.keys()):
        if field not in fields_to_keep:
            del entry[field]


if __name__ == '__main__':
    main(sys.argv[1])