reconstruct_pyx.py

from collections import defaultdict
import re
import sys

all = defaultdict(dict)

lineno = None
for line in sys.stdin:
    m = re.match(r' */\* "(.*\.pyx)":(\d+)', line)
    if m:
        file = m.group(1)
        lineno = int(m.group(2)) - 2
        print "found", m.groups()
    elif '*/' in line:
        lineno = None
    elif lineno is not None:
        all[file][lineno] = line[2:].rstrip()
        lineno += 1

for file, lines in all.items():
    print '=' * 20, file, '=' * 20
    for lineno, line in sorted(lines.items()):
        print lineno, ':', line