How to send utf-8 e-mail?

Question

how to send utf8 e-mail please?

import sys
import smtplib
import email
import re

from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText

def sendmail(firm, fromEmail, to, template, subject, date):
    with open(template, encoding="utf-8") as template_file:
        message = template_file.read()

    message = re.sub(r"{{\s*firm\s*}}", firm, message)
    message = re.sub(r"{{\s*date\s*}}", date, message)
    message = re.sub(r"{{\s*from\s*}}", fromEmail, message)
    message = re.sub(r"{{\s*to\s*}}", to, message)
    message = re.sub(r"{{\s*subject\s*}}", subject, message)

    msg = MIMEMultipart("alternative")
    msg.set_charset("utf-8")
    
    msg["Subject"] = subject
    msg["From"] = fromEmail
    msg["To"] = to

    #Read from template
    html = message[message.find("html:") + len("html:"):message.find("text:")].strip()
    text = message[message.find("text:") + len("text:"):].strip()

    part1 = MIMEText(html, "html")
    part2 = MIMEText(text, "plain")
    
    msg.attach(part1)    
    msg.attach(part2)

    try:
        server = smtplib.SMTP("10.0.0.5")
        server.sendmail(fromEmail, [to], msg.as_string())
        return 0
    except Exception as ex:
        #log error
        #return -1
        #debug
        raise ex
    finally:
        server.quit()

if __name__ == "__main__":
    #debug
    sys.argv.append("Moje")
    sys.argv.append("[email protected]")
    sys.argv.append("[email protected]")
    sys.argv.append("may2011.template")
    sys.argv.append("This is subject")
    sys.argv.append("This is date")

    
    if len(sys.argv) != 7:
        exit(-2)

    firm = sys.argv[1]
    fromEmail = sys.argv[2]
    to = sys.argv[3]
    template = sys.argv[4]
    subject = sys.argv[5]
    date = sys.argv[6]
    
    exit(sendmail(firm, fromEmail, to, template, subject, date))

Output

Traceback (most recent call last):
  File "C:\Documents and Settings\Administrator\Plocha\Newsletter-build-desktop\sendmail.py", line 69, in <module>
    exit(sendmail(firm, fromEmail, to, template, subject, date))   
  File "C:\Documents and Settings\Administrator\Plocha\Newsletter-build-desktop\sendmail.py", line 45, in sendmail
    raise ex
  File "C:\Documents and Settings\Administrator\Plocha\Newsletter-build-desktop\sendmail.py", line 39, in sendmail
    server.sendmail(fromEmail, [to], msg.as_string())
  File "C:\Python32\lib\smtplib.py", line 716, in sendmail
    msg = _fix_eols(msg).encode('ascii')
UnicodeEncodeError: 'ascii' codec can't encode character '\u011b' in position 385: ordinal not in range(128)

Solution 1

You should just add 'utf-8' argument to your MIMEText calls (it assumes 'us-ascii' by default).

For example:

# -*- encoding: utf-8 -*-

from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText

msg = MIMEMultipart("alternative")
msg["Subject"] = u'テストメール'
part1 = MIMEText(u'\u3053\u3093\u306b\u3061\u306f\u3001\u4e16\u754c\uff01\n',
                 "plain", "utf-8")
msg.attach(part1)

print msg.as_string().encode('ascii')

Solution 2

The question asked by Martin Drlík is 7 years and 8 months old... And nowadays, thanks to the developers of Python, encoding problems are solved with version 3 of Python.

Consequently, it is no longer necessary to specify that one must use the utf-8 encoding:

#!/usr/bin/python2
# -*- encoding: utf-8 -*-
...
    part2 = MIMEText(text, "plain", "utf-8")

We will simply write:

#!/usr/bin/python3
...
    part2 = MIMEText(text, "plain")

Ultimate consequence: Martin Drlík's script works perfectly well!

However, it would be better to use the email.parser module, as suggested in email: Examples.

Comments

[orlp]

We are missing line numbers so we can locate the exact line of the error.

[blubb]

The error is probably in the message.as_string(). You need to provide more, or we won't be able to help you.

[hephestos]

Sorry for being this old post and I write a new question underneath but I have to... Is supposed a human to write things like 'u\3035\u3093 ????? you have some japanese word right? how do you come up with these codes? from the tutorials i have read I always see this numbers ( while I understand are codes for letters , I don;t get how the input is done to this code or similar ones ...)

[sasikumar]

These codes are here only for illustration purposes. You can set your python source input encoding to utf-8 and type Unicode constants directly. If you need to get those '\u...' codes for some chars, you can just do something like print repr(u'テストメール')

[hephestos]

First thank you for the quick answer...Check this out ( to understand my issue with this question) I did this test : I put a utf-8 word (greek letters) in a variable and stored in a dictionary 'ΤΕΣΤ'=X. In the "Immediate window" when I do print listOfutfs reports me some u'' codes! ( from mysql I had ΤΕΣΤ retrieved ). Then I say ok give me the value back as listOfutfs['ΤΕΣΤ'], it replies to me KeyError: '\xce\xa6\xce\xa1\xce\x9b\xce\x9a'!!!!!!!!!!!!!!!!

[hephestos]

ok, got it [u'ΤΕΣΤ']! terrible sorry, I didn't know that this whole thing of u'' is for explanatory reasons....How do you concatenate though, unicodes in python then? you can not write u""" ΤΕΣΤ """ also, right ?

[sasikumar]

You actually can. Follow the python Unicode howto: docs.python.org/howto/unicode.html

[hephestos]

Thank you abbot, I just realised how many months lost...cheers

[IonicBurger]

I have tested this on python2.6.4 and it does not work as expected, the way to get past it is to also encode the unicode string as 'utf', it does however seem to work on python 2.7.2. In 2.6.4 you would do as an example, MIMEText( u'\3053'.encode('utf-8'), 'plain', 'utf-8')

[tripleee]

If this still uses the old MIMEText API, it should perhaps be avoided by anyone writing new code.

[Orsiris de Jong]

Well my code is compatible from python 2.7 up to recent 3.10, and tested via continuous integration. So as far as I want to keep retrocompatibility, I'll stick to MIMEText which, except of some shortcommings which I already cover via my code, doesn't have any major flaw AFAIK. Anyway, thanks for your advice.