convertpdfs.py

import os
import subprocess
from multiprocessing import Pool
from tqdm import tqdm

pdfs = sorted([int(f.split('.')[0]) for f in os.listdir('pdfs')])


def process(pdf):
    command = ['pdftotext', f'pdfs/{pdf}.pdf', f'txts/{pdf}.txt']
    subprocess.call(command)


if __name__ == '__main__':
    npoc = 30
    os.makedirs('txts', exist_ok=True)
    with Pool(npoc) as p:
        list(tqdm(p.imap(process, pdfs), total=len(pdfs)))