Remove entire publications from ACM Digital Library's exported .bib file

detrashify_acm.py

from pybtex.database.input import bibtex
from pybtex.database.output.bibtex import Writer
from pybtex.errors import set_strict_mode
import sys

if __name__ == '__main__':
    if len(sys.argv) != 2:
        print('usage: python3 detrashify_acm.py /path/to/acm.bib')
        exit(1)

    filename = sys.argv[1]
    print(f'Detrashifying {filename}')

    # allow duplicate bibs
    set_strict_mode(False)

    # parse acm.bib
    parser = bibtex.Parser()
    bib_data = parser.parse_file(filename)

    old_len = len(bib_data.entries.items())

    # remove entries with authors == None
    keys_to_remove = []
    excluded = 0
    for cite_key, entry in bib_data.entries.items():
        if 'author' not in entry.persons or len(entry.persons['author']) == 0: # remove
            print(f"Filtering out {entry.fields['title'].split(':')[0]}")
            keys_to_remove.append(cite_key)
            excluded += 1

    for key in keys_to_remove:
        del bib_data.entries[key]

    writer = Writer()
    with open("acm-updated.bib", 'w', encoding='utf-8') as file:
        writer.write_stream(bib_data, file)

    new_len = len(bib_data.entries.items())
    print(f'excluded {excluded} entries without authors')
    print(f'Old len: {old_len}\t New len: {new_len}')
    print(f'Updated file: acm-updated.bib')