getbibs

#!/usr/bin/env python3

import argparse
import sys

"""
Extract all bibtex references from a plaintext document
"""

if __name__ == "__main__":
    # parse input arguments
    parser = argparse.ArgumentParser(description="Extract all bibtex-formatted references from a plaintext document.")
    parser.add_argument("-i", "--infile", type=str, default=None, help="File from which to extract references, default stdin")
    parser.add_argument("-o", "--output", type=str, default=None, help="Output file to save references to, default stdout")
    parser.add_argument("-r", "--remove", action="store_true", default=False, help="Instead of producing references file, produce file with references removed")
    args=parser.parse_args()

    reference_lines = []
    text_lines = []
    # Open text file for processing
    with (open(args.infile, 'r') if args.infile else sys.stdin) as infile:
        line = infile.readline()
        while line:
            # If a line doesn't start with '@', ignore it
            if not line.lstrip().startswith('@'):
                text_lines.append(line)
                line = infile.readline()
                continue
            # Otherwise, start recording a new reference
            this_ref = []
            running_total = line.count('{') - line.count('}')
            # While the reference is open, and EOF not reached...
            while running_total>0 and line:
                this_ref.append(line)
                line = infile.readline()
                running_total += (line.count('{') - line.count('}'))
            # Either reference is closed, or EOF reached
            reference_lines += ( this_ref + ['}\n\n']) 
            line = infile.readline()

    output_lines = text_lines if args.remove else reference_lines
    if args.output is None:
        print(''.join(output_lines))
    else:
        with open(args.output, 'w') as outfile:
            outfile.write(''.join(output_lines))