Você pode tentar este script. Eu descobri que precisava de alguns arquivos mbox que eu baixei do Mailman para arquivá-los em um formato mutt reconhecido. Eu acho que é mais pickiest sobre o formato de data. Eu ainda não encontrei uma correção mais fácil. Mas isso funciona para mim.
#!/usr/bin/env python
"""
Usage: ./mailman2mbox.py infile outfile default-to-address
"""
import sys
from time import strftime,strptime,mktime,asctime
from email.utils import parseaddr,formatdate
if len(sys.argv) not in (3,4):
print __doc__
sys.exit()
out = open(sys.argv[2],"w")
listid = None
if len(sys.argv)==4:
listid = sys.argv[3]
date_patterns = ("%b %d %H:%M:%S %Y", "%d %b %H:%M:%S %Y", "%d %b %Y %H:%M:%S", "%d %b %H:%M:%S", "%d %b %y %H:%M:%S", "%d %b %Y %H.%M.%S",'%m/%d/%y %H:%M:%S %p')
class HeaderError(TypeError):
pass
def finish(headers, body):
body.append("\n")
for n,ln in enumerate(headers):
if ln.startswith("Date:"):
break
else:
raise HeaderError("No 'Date:' header:\n" + "".join(headers)+"\n")
if listid is not None:
for ln2 in headers:
if ln2.lower().startswith("list-id:"):
break
else:
headers.append("List-Id: <%s>\n" % (listid,))
date_line = ln[5:].strip()
if date_line.endswith(')'):
date_line = date_line[:date_line.rfind('(')].rstrip()
if date_line[-5] in "+-":
date_line, tz = date_line[:-5].rstrip(), int(date_line[-5:])//100
else:
tz = -5
if date_line[:3] in ("Mon","Tue","Wed","Thu","Fri","Sat","Sun"):
if date_line[3:5] == ', ':
prefix = "%a, "
elif date_line[3] == ',':
prefix = "%a,"
else:
prefix = "%a "
else:
prefix = ""
while True:
for p in date_patterns:
try:
date_struct = strptime(date_line, prefix+p)
except ValueError:
pass
else:
break
else:
if not date_line:
raise ValueError(headers[n])
date_line = date_line[:date_line.rfind(' ')]
continue
break
date_struct = list(date_struct)
try:
headers[n] = 'Date: %s\n' % (formatdate(mktime(date_struct),True))
headers[0] = "%s %s\n" % (headers[0][:-25].rstrip(), asctime(date_struct), )
except ValueError:
raise ValueError(headers[n])
for w in headers, body:
for s in w:
out.write(s)
message = 0
headers, body = None, None
for line in open(sys.argv[1]):
if line.startswith("From "):
message+=1
header = True
if headers is not None:
try:
finish(headers, body)
except HeaderError:
message -= 1
out.write('>')
for w in headers, body:
for s in w:
out.write(s)
headers, body = [], []
line = line.replace(" at ", "@")
elif line == '\n':
header = False
elif header and line.startswith('From:'):
line = line.replace(" at ","@")
(headers if header else body).append(line)
try:
finish(headers, body)
except HeaderError:
out.write('>')
for w in headers, body:
for s in w:
out.write(s)
out.close()