Mailing List Archive

On 2023-03-19, Stefan Ram <ram@zedat.fu-berlin.de> wrote:
> Peng Yu <pengyu.ut@gmail.com> writes:
>>But when I try the following code, get_body() is not found. How to get
>>get_body() to work?
>
> Did you know that this post of mine here was posted to
> Usenet with a Python script I wrote?
>
> That Python script has a function to show the body of
> a post before posting. The post is contained in a file,
> so it reads the post from that file.
>
> I copy it here, maybe it can help some people to see
> how I do this.
>
> # Python 3.5
>
> import email
>
> ...
>
> def showbody( file ): # lightly edited for posting on 2023-03-19
> output = ''
> msg = email.message_from_binary_file\
> ( file, policy=email.policy.default )

I wouldn't generally be pedantic about code style, but that's giving me
painful convulsions. Backslashes for line continuations are generally
considered a bad idea (as they mean that any whitespace after the
backslash, which is often invisible, becomes significant). And not
indenting the continuation line(s) is pretty shocking. Writing it as
below is objectively better:

msg = email.message_from_binary_file(
file, policy=email.policy.default )

(Also, I too find it annoying to have to avoid, but calling a local
variable 'file' is somewhat suspect since it shadows the builtin.)
--
https://mail.python.org/mailman/listinfo/python-list

On 3/18/2023 10:49 PM, Peng Yu wrote:
> Hi,
>
> https://docs.python.org/3/library/email.parser.html
>
> It says "For MIME messages, the root object will return True from its
> is_multipart() method, and the subparts can be accessed via the
> payload manipulation methods, such as get_body(), iter_parts(), and
> walk()."
>
> But when I try the following code, get_body() is not found. How to get
> get_body() to work?
>
> $ python3 -c 'import email, sys; msg =
> email.message_from_string(sys.stdin.read()); print(msg.get_body())'
> <<< some_text
> Traceback (most recent call last):
> File "<string>", line 1, in <module>
> AttributeError: 'Message' object has no attribute 'get_body'
>
A Message object does not have a get_body method, but an EmailMessage
object does.

In the Python 3.10 docs, Just before the part you quoted, there is this
sentence:

"You can pass the parser a bytes, string or file object, and the parser
will return to you the root EmailMessage instance of the object
structure".

So if you want to use get_body(), you should be feeding the parser your
message string, rather than using email.message_from_string(), which
returns a Message, not an EmailMessage.

With a Message object, you could use get_payload(), which will give you
a list of Messages for each MIME part of the document. When a part is
not a multipart, it will give you a string, which sounds like what you
want to end up with.

(see https://docs.python.org/3.10/library/email.compat32-message.html)
--
https://mail.python.org/mailman/listinfo/python-list

On 2023-03-19, Stefan Ram <ram@zedat.fu-berlin.de> wrote:
> Jon Ribbens <jon+usenet@unequivocal.eu> writes:
>>(Also, I too find it annoying to have to avoid, but calling a local
>>variable 'file' is somewhat suspect since it shadows the builtin.)
>
> Thanks for your remarks, but I'm not aware
> of such a predefined name "file"!

Ah, apparently it got removed in Python 3, which is a bit odd as the
last I heard it was added in Python 2.2 in order to achieve consistency
with other types.
--
https://mail.python.org/mailman/listinfo/python-list

On 20/03/23 7:07 am, Jon Ribbens wrote:
> Ah, apparently it got removed in Python 3, which is a bit odd as the
> last I heard it was added in Python 2.2 in order to achieve consistency
> with other types.

As far as I remember, the file type came into existence
with type/class unification, and "open" became an alias
for the file type, so you could use open() and file()
interchangeably.

With the Unicode revolution in Python 3, file handling got
a lot more complicated. Rather than a single file type,
there are now a bunch of classes that handle low-level I/O,
encoding/decoding, etc, and open() is a function again
that builds the appropriate combination of underlying
objects.

--
Greg



--
https://mail.python.org/mailman/listinfo/python-list

On 2023-03-19, Greg Ewing <greg.ewing@canterbury.ac.nz> wrote:
> On 20/03/23 7:07 am, Jon Ribbens wrote:
>> Ah, apparently it got removed in Python 3, which is a bit odd as the
>> last I heard it was added in Python 2.2 in order to achieve consistency
>> with other types.
>
> As far as I remember, the file type came into existence
> with type/class unification, and "open" became an alias
> for the file type, so you could use open() and file()
> interchangeably.
>
> With the Unicode revolution in Python 3, file handling got
> a lot more complicated. Rather than a single file type,
> there are now a bunch of classes that handle low-level I/O,
> encoding/decoding, etc, and open() is a function again
> that builds the appropriate combination of underlying
> objects.

This is true, however there does exist a base class which, according to
the documentation, underlies all of the different IO classes - IOBase -
so it might have been neater to make 'file' be an alias for that.
--
https://mail.python.org/mailman/listinfo/python-list