dbm_save.py

import collections, fileinput, os, dbm
word_pos = collections.defaultdict(list)
for line in fileinput.input():
    pos = f'{fileinput.filename()}{os.pathsep}{fileinput.filelineno()}'
    for word in line.split():
        word_pos[word].append(pos)
sep2 = os.pathsep * 2
with dbm.open('indexfile','n') as dbm_out:
    for word in word_pos:
        dbm_out[word.encode('utf-8')] = sep2.join(word_pos[word]).encode('utf-8')