The central class in the email package is the Message class, imported from the email.message module. It is the base class for the email object model. Message provides the core functionality for setting and querying header fields, and for accessing message bodies.
Conceptually, a Message object consists of headers and payloads. Headers are RFC 2822 style field names and values where the field name and value are separated by a colon. The colon is not part of either the field name or the field value.
Headers are stored and returned in case-preserving form but are matched case-insensitively. There may also be a single envelope header, also known as the Unix-From header or the From_ header. The payload is either a string in the case of simple message objects or a list of Message objects for MIME container documents (e.g. multipart/* and message/rfc822).
Message objects provide a mapping style interface for accessing the message headers, and an explicit interface for accessing both the headers and the payload. It provides convenience methods for generating a flat text representation of the message object tree, for accessing commonly used header parameters, and for recursively walking over the object tree.
Here are the methods of the Message class:
If policy is specified (it must be an instance of a policy class) use the rules it specifies to update and serialize the representation of the message. If policy is not set, use the compat32 policy, which maintains backward compatibility with the Python 3.2 version of the email package. For more information see the policy documentation.
Changed in version 3.3: The policy keyword argument was added.
Return the entire message flattened as a string. When optional unixfrom is true, the envelope header is included in the returned string. unixfrom defaults to False. For backward compabitility reasons, maxheaderlen defaults to 0, so if you want a different value you must override it explicitly (the value specified for max_line_length in the policy will be ignored by this method). The policy argument may be used to override the default policy obtained from the message instance. This can be used to control some of the formatting produced by the method, since the specified policy will be passed to the Generator.
Flattening the message may trigger changes to the Message if defaults need to be filled in to complete the transformation to a string (for example, MIME boundaries may be generated or modified).
Note that this method is provided as a convenience and may not always format the message the way you want. For example, by default it does not do the mangling of lines that begin with From that is required by the unix mbox format. For more flexibility, instantiate a Generator instance and use its flatten() method directly. For example:
from io import StringIO
from email.generator import Generator
fp = StringIO()
g = Generator(fp, mangle_from_=True, maxheaderlen=60)
g.flatten(msg)
text = fp.getvalue()
If the message object contains binary data that is not encoded according to RFC standards, the non-compliant data will be replaced by unicode “unknown character” code points. (See also as_bytes() and BytesGenerator.)
Changed in version 3.4: the policy keyword argument was added.
Equivalent to as_string(). Allows str(msg) to produce a string containing the formatted message.
Return the entire message flattened as a bytes object. When optional unixfrom is true, the envelope header is included in the returned string. unixfrom defaults to False. The policy argument may be used to override the default policy obtained from the message instance. This can be used to control some of the formatting produced by the method, since the specified policy will be passed to the BytesGenerator.
Flattening the message may trigger changes to the Message if defaults need to be filled in to complete the transformation to a string (for example, MIME boundaries may be generated or modified).
Note that this method is provided as a convenience and may not always format the message the way you want. For example, by default it does not do the mangling of lines that begin with From that is required by the unix mbox format. For more flexibility, instantiate a BytesGenerator instance and use its flatten() method directly. For example:
from io import BytesIO
from email.generator import BytesGenerator
fp = BytesIO()
g = BytesGenerator(fp, mangle_from_=True, maxheaderlen=60)
g.flatten(msg)
text = fp.getvalue()
New in version 3.4.
Equivalent to as_bytes(). Allows bytes(msg) to produce a bytes object containing the formatted message.
New in version 3.4.
Return True if the message’s payload is a list of sub-Message objects, otherwise return False. When is_multipart() returns False, the payload should be a string object. (Note that is_multipart() returning True does not necessarily mean that “msg.get_content_maintype() == ‘multipart’” will return the True. For example, is_multipart will return True when the Message is of type message/rfc822.)
Set the message’s envelope header to unixfrom, which should be a string.
Return the message’s envelope header. Defaults to None if the envelope header was never set.
Add the given payload to the current payload, which must be None or a list of Message objects before the call. After the call, the payload will always be a list of Message objects. If you want to set the payload to a scalar object (e.g. a string), use set_payload() instead.
Return the current payload, which will be a list of Message objects when is_multipart() is True, or a string when is_multipart() is False. If the payload is a list and you mutate the list object, you modify the message’s payload in place.
With optional argument i, get_payload() will return the i-th element of the payload, counting from zero, if is_multipart() is True. An IndexError will be raised if i is less than 0 or greater than or equal to the number of items in the payload. If the payload is a string (i.e. is_multipart() is False) and i is given, a TypeError is raised.
Optional decode is a flag indicating whether the payload should be decoded or not, according to the Content-Transfer-Encoding header. When True and the message is not a multipart, the payload will be decoded if this header’s value is quoted-printable or base64. If some other encoding is used, or Content-Transfer-Encoding header is missing, the payload is returned as-is (undecoded). In all cases the returned value is binary data. If the message is a multipart and the decode flag is True, then None is returned. If the payload is base64 and it was not perfectly formed (missing padding, characters outside the base64 alphabet), then an appropriate defect will be added to the message’s defect property (InvalidBase64PaddingDefect or InvalidBase64CharactersDefect, respectively).
When decode is False (the default) the body is returned as a string without decoding the Content-Transfer-Encoding. However, for a Content-Transfer-Encoding of 8bit, an attempt is made to decode the original bytes using the charset specified by the Content-Type header, using the replace error handler. If no charset is specified, or if the charset given is not recognized by the email package, the body is decoded using the default ASCII charset.
Set the entire message object’s payload to payload. It is the client’s responsibility to ensure the payload invariants. Optional charset sets the message’s default character set; see set_charset() for details.
Set the character set of the payload to charset, which can either be a Charset instance (see email.charset), a string naming a character set, or None. If it is a string, it will be converted to a Charset instance. If charset is None, the charset parameter will be removed from the Content-Type header (the message will not be otherwise modified). Anything else will generate a TypeError.
If there is no existing MIME-Version header one will be added. If there is no existing Content-Type header, one will be added with a value of text/plain. Whether the Content-Type header already exists or not, its charset parameter will be set to charset.output_charset. If charset.input_charset and charset.output_charset differ, the payload will be re-encoded to the output_charset. If there is no existing Content-Transfer-Encoding header, then the payload will be transfer-encoded, if needed, using the specified Charset, and a header with the appropriate value will be added. If a Content-Transfer-Encoding header already exists, the payload is assumed to already be correctly encoded using that Content-Transfer-Encoding and is not modified.
The following methods implement a mapping-like interface for accessing the message’s RFC 2822 headers. Note that there are some semantic differences between these methods and a normal mapping (i.e. dictionary) interface. For example, in a dictionary there are no duplicate keys, but here there may be duplicate message headers. Also, in dictionaries there is no guaranteed order to the keys returned by keys(), but in a Message object, headers are always returned in the order they appeared in the original message, or were added to the message later. Any header deleted and then re-added are always appended to the end of the header list.
These semantic differences are intentional and are biased toward maximal convenience.
Note that in all cases, any envelope header present in the message is not included in the mapping interface.
In a model generated from bytes, any header values that (in contravention of the RFCs) contain non-ASCII bytes will, when retrieved through this interface, be represented as Header objects with a charset of unknown-8bit.
Return the total number of headers, including duplicates.
Return true if the message object has a field named name. Matching is done case-insensitively and name should not include the trailing colon. Used for the in operator, e.g.:
if 'message-id' in myMessage:
print('Message-ID:', myMessage['message-id'])
Return the value of the named header field. name should not include the colon field separator. If the header is missing, None is returned; a KeyError is never raised.
Note that if the named field appears more than once in the message’s headers, exactly which of those field values will be returned is undefined. Use the get_all() method to get the values of all the extant named headers.
Add a header to the message with field name name and value val. The field is appended to the end of the message’s existing fields.
Note that this does not overwrite or delete any existing header with the same name. If you want to ensure that the new header is the only one present in the message with field name name, delete the field first, e.g.:
del msg['subject']
msg['subject'] = 'Python roolz!'
Delete all occurrences of the field with name name from the message’s headers. No exception is raised if the named field isn’t present in the headers.
Return a list of all the message’s header field names.
Return a list of all the message’s field values.
Return a list of 2-tuples containing all the message’s field headers and values.
Return the value of the named header field. This is identical to __getitem__() except that optional failobj is returned if the named header is missing (defaults to None).
Here are some additional useful methods:
Return a list of all the values for the field named name. If there are no such named headers in the message, failobj is returned (defaults to None).
Extended header setting. This method is similar to __setitem__() except that additional header parameters can be provided as keyword arguments. _name is the header field to add and _value is the primary value for the header.
For each item in the keyword argument dictionary _params, the key is taken as the parameter name, with underscores converted to dashes (since dashes are illegal in Python identifiers). Normally, the parameter will be added as key="value" unless the value is None, in which case only the key will be added. If the value contains non-ASCII characters, it can be specified as a three tuple in the format (CHARSET, LANGUAGE, VALUE), where CHARSET is a string naming the charset to be used to encode the value, LANGUAGE can usually be set to None or the empty string (see RFC 2231 for other possibilities), and VALUE is the string value containing non-ASCII code points. If a three tuple is not passed and the value contains non-ASCII characters, it is automatically encoded in RFC 2231 format using a CHARSET of utf-8 and a LANGUAGE of None.
Here’s an example:
msg.add_header('Content-Disposition', 'attachment', filename='bud.gif')
This will add a header that looks like
Content-Disposition: attachment; filename="bud.gif"
An example with non-ASCII characters:
msg.add_header('Content-Disposition', 'attachment',
filename=('iso-8859-1', '', 'Fußballer.ppt'))
Which produces
Content-Disposition: attachment; filename*="iso-8859-1''Fu%DFballer.ppt"
Replace a header. Replace the first header found in the message that matches _name, retaining header order and field name case. If no matching header was found, a KeyError is raised.
Return the message’s content type. The returned string is coerced to lower case of the form maintype/subtype. If there was no Content-Type header in the message the default type as given by get_default_type() will be returned. Since according to RFC 2045, messages always have a default type, get_content_type() will always return a value.
RFC 2045 defines a message’s default type to be text/plain unless it appears inside a multipart/digest container, in which case it would be message/rfc822. If the Content-Type header has an invalid type specification, RFC 2045 mandates that the default type be text/plain.
Return the message’s main content type. This is the maintype part of the string returned by get_content_type().
Return the message’s sub-content type. This is the subtype part of the string returned by get_content_type().
Return the default content type. Most messages have a default content type of text/plain, except for messages that are subparts of multipart/digest containers. Such subparts have a default content type of message/rfc822.
Set the default content type. ctype should either be text/plain or message/rfc822, although this is not enforced. The default content type is not stored in the Content-Type header.
Return the message’s Content-Type parameters, as a list. The elements of the returned list are 2-tuples of key/value pairs, as split on the '=' sign. The left hand side of the '=' is the key, while the right hand side is the value. If there is no '=' sign in the parameter the value is the empty string, otherwise the value is as described in get_param() and is unquoted if optional unquote is True (the default).
Optional failobj is the object to return if there is no Content-Type header. Optional header is the header to search instead of Content-Type.
Return the value of the Content-Type header’s parameter param as a string. If the message has no Content-Type header or if there is no such parameter, then failobj is returned (defaults to None).
Optional header if given, specifies the message header to use instead of Content-Type.
Parameter keys are always compared case insensitively. The return value can either be a string, or a 3-tuple if the parameter was RFC 2231 encoded. When it’s a 3-tuple, the elements of the value are of the form (CHARSET, LANGUAGE, VALUE). Note that both CHARSET and LANGUAGE can be None, in which case you should consider VALUE to be encoded in the us-ascii charset. You can usually ignore LANGUAGE.
If your application doesn’t care whether the parameter was encoded as in RFC 2231, you can collapse the parameter value by calling email.utils.collapse_rfc2231_value(), passing in the return value from get_param(). This will return a suitably decoded Unicode string when the value is a tuple, or the original string unquoted if it isn’t. For example:
rawparam = msg.get_param('foo')
param = email.utils.collapse_rfc2231_value(rawparam)
In any case, the parameter value (either the returned string, or the VALUE item in the 3-tuple) is always unquoted, unless unquote is set to False.
Set a parameter in the Content-Type header. If the parameter already exists in the header, its value will be replaced with value. If the Content-Type header as not yet been defined for this message, it will be set to text/plain and the new parameter value will be appended as per RFC 2045.
Optional header specifies an alternative header to Content-Type, and all parameters will be quoted as necessary unless optional requote is False (the default is True).
If optional charset is specified, the parameter will be encoded according to RFC 2231. Optional language specifies the RFC 2231 language, defaulting to the empty string. Both charset and language should be strings.
If replace is False (the default) the header is moved to the end of the list of headers. If replace is True, the header will be updated in place.
Changed in version 3.4: replace keyword was added.
Remove the given parameter completely from the Content-Type header. The header will be re-written in place without the parameter or its value. All values will be quoted as necessary unless requote is False (the default is True). Optional header specifies an alternative to Content-Type.
Set the main type and subtype for the Content-Type header. type must be a string in the form maintype/subtype, otherwise a ValueError is raised.
This method replaces the Content-Type header, keeping all the parameters in place. If requote is False, this leaves the existing header’s quoting as is, otherwise the parameters will be quoted (the default).
An alternative header can be specified in the header argument. When the Content-Type header is set a MIME-Version header is also added.
Return the value of the filename parameter of the Content-Disposition header of the message. If the header does not have a filename parameter, this method falls back to looking for the name parameter on the Content-Type header. If neither is found, or the header is missing, then failobj is returned. The returned string will always be unquoted as per email.utils.unquote().
Return the value of the boundary parameter of the Content-Type header of the message, or failobj if either the header is missing, or has no boundary parameter. The returned string will always be unquoted as per email.utils.unquote().
Set the boundary parameter of the Content-Type header to boundary. set_boundary() will always quote boundary if necessary. A HeaderParseError is raised if the message object has no Content-Type header.
Note that using this method is subtly different than deleting the old Content-Type header and adding a new one with the new boundary via add_header(), because set_boundary() preserves the order of the Content-Type header in the list of headers. However, it does not preserve any continuation lines which may have been present in the original Content-Type header.
Return the charset parameter of the Content-Type header, coerced to lower case. If there is no Content-Type header, or if that header has no charset parameter, failobj is returned.
Note that this method differs from get_charset() which returns the Charset instance for the default encoding of the message body.
Return a list containing the character set names in the message. If the message is a multipart, then the list will contain one element for each subpart in the payload, otherwise, it will be a list of length 1.
Each item in the list will be a string which is the value of the charset parameter in the Content-Type header for the represented subpart. However, if the subpart has no Content-Type header, no charset parameter, or is not of the text main MIME type, then that item in the returned list will be failobj.
The walk() method is an all-purpose generator which can be used to iterate over all the parts and subparts of a message object tree, in depth-first traversal order. You will typically use walk() as the iterator in a for loop; each iteration returns the next subpart.
Here’s an example that prints the MIME type of every part of a multipart message structure:
>>> for part in msg.walk():
... print(part.get_content_type())
multipart/report
text/plain
message/delivery-status
text/plain
text/plain
message/rfc822
text/plain
walk iterates over the subparts of any part where is_multipart() returns True, even though msg.get_content_maintype() == 'multipart' may return False. We can see this in our example by making use of the _structure debug helper function:
>>> for part in msg.walk():
... print(part.get_content_maintype() == 'multipart'),
... part.is_multipart())
True True
False False
False True
False False
False False
False True
False False
>>> _structure(msg)
multipart/report
text/plain
message/delivery-status
text/plain
text/plain
message/rfc822
text/plain
Here the message parts are not multiparts, but they do contain subparts. is_multipart() returns True and walk descends into the subparts.
Message objects can also optionally contain two instance attributes, which can be used when generating the plain text of a MIME message.
The format of a MIME document allows for some text between the blank line following the headers, and the first multipart boundary string. Normally, this text is never visible in a MIME-aware mail reader because it falls outside the standard MIME armor. However, when viewing the raw text of the message, or when viewing the message in a non-MIME aware reader, this text can become visible.
The preamble attribute contains this leading extra-armor text for MIME documents. When the Parser discovers some text after the headers but before the first boundary string, it assigns this text to the message’s preamble attribute. When the Generator is writing out the plain text representation of a MIME message, and it finds the message has a preamble attribute, it will write this text in the area between the headers and the first boundary. See email.parser and email.generator for details.
Note that if the message object has no preamble, the preamble attribute will be None.
The epilogue attribute acts the same way as the preamble attribute, except that it contains text that appears between the last boundary and the end of the message.
You do not need to set the epilogue to the empty string in order for the Generator to print a newline at the end of the file.
The defects attribute contains a list of all the problems found when parsing this message. See email.errors for a detailed description of the possible parsing defects.