The lib email parse problem...

[John Machin]

叮叮当当 wrote:
> this is not enough.
>
> when a part is mulitpart/alternative, i must find out which sub part i
> need, not all the subparts. so i must know when the alternative is
> ended.
>

So you'll have to write your own tree-walker. It would seem that
is_multipart(), get_content_type() and get_payload() are the important
methods.

Here's a quickly lashed-up example:

def choose_one(part, html_ok=False):
    last = None
    for subpart in part.get_payload():
        if html_ok or "html" not in subpart.get_content_type():
            last = subpart
    return last

def traverse(part, html_ok=False):
    mp = part.is_multipart()
    ty = part.get_content_type()
    print "multi:%r type:%r file:%r" % (mp, ty,
part.get_filename("<<NoFileName>>"))
    if mp:
        if ty == "multipart/alternative":
            chosen = choose_one(part, html_ok=html_ok)
            traverse(chosen, html_ok=html_ok)
        else:
            for subpart in part.get_payload():
                traverse(subpart, html_ok=html_ok)

import email
pmsg = email.message_from_string(msg_text)
for toggle in (True, False):
    print "--- html_ok is %r ---" % toggle
    traverse(pmsg, html_ok=toggle)

With a suitable message, this produced:

--- html_ok is True ---
multi:True type:'multipart/alternative' file:'<<NoFileName>>'
multi:False type:'text/html' file:'<<NoFileName>>'
--- html_ok is False ---
multi:True type:'multipart/alternative' file:'<<NoFileName>>'
multi:False type:'text/plain' file:'<<NoFileName>>'