1 # Copyright (C) 2001-2006 Python Software Foundation
3 # Contact: email-sig@python.org
5 """Basic message object for the email package object model."""
13 from cStringIO import StringIO
15 # Intrapackage imports
17 from email import utils
18 from email import errors
22 # Regular expression that matches `special' characters in parameters, the
23 # existence of which force quoting of the parameter value.
24 tspecials = re.compile(r'[ \(\)<>@,;:\\"/\[\]\?=]')
28 def _splitparam(param):
29 # Split header parameters. BAW: this may be too simple. It isn't
30 # strictly RFC 2045 (section 5.1) compliant, but it catches most headers
31 # found in the wild. We may eventually need a full fledged parser
33 a, sep, b = param.partition(';')
35 return a.strip(), None
36 return a.strip(), b.strip()
38 def _formatparam(param, value=None, quote=True):
39 """Convenience function to format and return a key=value pair.
41 This will quote the value if needed or if quote is true.
43 if value is not None and len(value) > 0:
44 # A tuple is used for RFC 2231 encoded parameter values where items
45 # are (charset, language, value). charset is a string, not a Charset
47 if isinstance(value, tuple):
48 # Encode as per RFC 2231
50 value = utils.encode_rfc2231(value[2], value[0], value[1])
51 # BAW: Please check this. I think that if quote is set it should
52 # force quoting even if not necessary.
53 if quote or tspecials.search(value):
54 return '%s="%s"' % (param, utils.quote(value))
56 return '%s=%s' % (param, value)
65 while end > 0 and (s.count('"', 0, end) - s.count('\\"', 0, end)) % 2:
66 end = s.find(';', end + 1)
72 f = f[:i].strip().lower() + '=' + f[i+1:].strip()
73 plist.append(f.strip())
78 def _unquotevalue(value):
79 # This is different than utils.collapse_rfc2231_value() because it doesn't
80 # try to convert the value to a unicode. Message.get_param() and
81 # Message.get_params() are both currently defined to return the tuple in
82 # the face of RFC 2231 parameters.
83 if isinstance(value, tuple):
84 return value[0], value[1], utils.unquote(value[2])
86 return utils.unquote(value)
91 """Basic message object.
93 A message object is defined as something that has a bunch of RFC 2822
94 headers and a payload. It may optionally have an envelope header
95 (a.k.a. Unix-From or From_ header). If the message is a container (i.e. a
96 multipart or a message/rfc822), then the payload is a list of Message
97 objects, otherwise it is a string.
99 Message objects implement part of the `mapping' interface, which assumes
100 there is exactly one occurrance of the header per message. Some headers
101 do in fact appear multiple times (e.g. Received) and for those headers,
102 you must use the explicit API to set or get all the headers. Not all of
103 the mapping methods are implemented.
107 self._unixfrom = None
110 # Defaults for multipart messages
111 self.preamble = self.epilogue = None
113 # Default content type
114 self._default_type = 'text/plain'
117 """Return the entire formatted message as a string.
118 This includes the headers, body, and envelope header.
120 return self.as_string(unixfrom=True)
122 def as_string(self, unixfrom=False):
123 """Return the entire formatted message as a string.
124 Optional `unixfrom' when True, means include the Unix From_ envelope
127 This is a convenience method and may not generate the message exactly
128 as you intend because by default it mangles lines that begin with
129 "From ". For more flexibility, use the flatten() method of a
132 from email.generator import Generator
135 g.flatten(self, unixfrom=unixfrom)
138 def is_multipart(self):
139 """Return True if the message consists of multiple parts."""
140 return isinstance(self._payload, list)
145 def set_unixfrom(self, unixfrom):
146 self._unixfrom = unixfrom
148 def get_unixfrom(self):
149 return self._unixfrom
152 # Payload manipulation.
154 def attach(self, payload):
155 """Add the given payload to the current payload.
157 The current payload will always be a list of objects after this method
158 is called. If you want to set the payload to a scalar object, use
159 set_payload() instead.
161 if self._payload is None:
162 self._payload = [payload]
164 self._payload.append(payload)
166 def get_payload(self, i=None, decode=False):
167 """Return a reference to the payload.
169 The payload will either be a list object or a string. If you mutate
170 the list object, you modify the message's payload in place. Optional
171 i returns that index into the payload.
173 Optional decode is a flag indicating whether the payload should be
174 decoded or not, according to the Content-Transfer-Encoding header
177 When True and the message is not a multipart, the payload will be
178 decoded if this header's value is `quoted-printable' or `base64'. If
179 some other encoding is used, or the header is missing, or if the
180 payload has bogus data (i.e. bogus base64 or uuencoded data), the
181 payload is returned as-is.
183 If the message is a multipart and the decode flag is True, then None
187 payload = self._payload
188 elif not isinstance(self._payload, list):
189 raise TypeError('Expected list, got %s' % type(self._payload))
191 payload = self._payload[i]
193 if self.is_multipart():
195 cte = self.get('content-transfer-encoding', '').lower()
196 if cte == 'quoted-printable':
197 return utils._qdecode(payload)
198 elif cte == 'base64':
200 return utils._bdecode(payload)
201 except binascii.Error:
204 elif cte in ('x-uuencode', 'uuencode', 'uue', 'x-uue'):
207 uu.decode(StringIO(payload+'\n'), sfp, quiet=True)
208 payload = sfp.getvalue()
210 # Some decoding problem
212 # Everything else, including encodings with 8bit or 7bit are returned
216 def set_payload(self, payload, charset=None):
217 """Set the payload to the given value.
219 Optional charset sets the message's default character set. See
220 set_charset() for details.
222 self._payload = payload
223 if charset is not None:
224 self.set_charset(charset)
226 def set_charset(self, charset):
227 """Set the charset of the payload to a given character set.
229 charset can be a Charset instance, a string naming a character set, or
230 None. If it is a string it will be converted to a Charset instance.
231 If charset is None, the charset parameter will be removed from the
232 Content-Type field. Anything else will generate a TypeError.
234 The message will be assumed to be of type text/* encoded with
235 charset.input_charset. It will be converted to charset.output_charset
236 and encoded properly, if needed, when generating the plain text
237 representation of the message. MIME headers (MIME-Version,
238 Content-Type, Content-Transfer-Encoding) will be added as needed.
242 self.del_param('charset')
245 if isinstance(charset, basestring):
246 charset = email.charset.Charset(charset)
247 if not isinstance(charset, email.charset.Charset):
248 raise TypeError(charset)
249 # BAW: should we accept strings that can serve as arguments to the
250 # Charset constructor?
251 self._charset = charset
252 if 'MIME-Version' not in self:
253 self.add_header('MIME-Version', '1.0')
254 if 'Content-Type' not in self:
255 self.add_header('Content-Type', 'text/plain',
256 charset=charset.get_output_charset())
258 self.set_param('charset', charset.get_output_charset())
259 if isinstance(self._payload, unicode):
260 self._payload = self._payload.encode(charset.output_charset)
261 if str(charset) != charset.get_output_charset():
262 self._payload = charset.body_encode(self._payload)
263 if 'Content-Transfer-Encoding' not in self:
264 cte = charset.get_body_encoding()
268 self._payload = charset.body_encode(self._payload)
269 self.add_header('Content-Transfer-Encoding', cte)
271 def get_charset(self):
272 """Return the Charset instance associated with the message's payload.
277 # MAPPING INTERFACE (partial)
280 """Return the total number of headers, including duplicates."""
281 return len(self._headers)
283 def __getitem__(self, name):
284 """Get a header value.
286 Return None if the header is missing instead of raising an exception.
288 Note that if the header appeared multiple times, exactly which
289 occurrance gets returned is undefined. Use get_all() to get all
290 the values matching a header field name.
292 return self.get(name)
294 def __setitem__(self, name, val):
295 """Set the value of a header.
297 Note: this does not overwrite an existing header with the same field
298 name. Use __delitem__() first to delete any existing headers.
300 self._headers.append((name, val))
302 def __delitem__(self, name):
303 """Delete all occurrences of a header, if present.
305 Does not raise an exception if the header is missing.
309 for k, v in self._headers:
310 if k.lower() != name:
311 newheaders.append((k, v))
312 self._headers = newheaders
314 def __contains__(self, name):
315 return name.lower() in [k.lower() for k, v in self._headers]
317 def has_key(self, name):
318 """Return true if the message contains the header."""
320 return self.get(name, missing) is not missing
323 """Return a list of all the message's header field names.
325 These will be sorted in the order they appeared in the original
326 message, or were added to the message, and may contain duplicates.
327 Any fields deleted and re-inserted are always appended to the header
330 return [k for k, v in self._headers]
333 """Return a list of all the message's header values.
335 These will be sorted in the order they appeared in the original
336 message, or were added to the message, and may contain duplicates.
337 Any fields deleted and re-inserted are always appended to the header
340 return [v for k, v in self._headers]
343 """Get all the message's header fields and values.
345 These will be sorted in the order they appeared in the original
346 message, or were added to the message, and may contain duplicates.
347 Any fields deleted and re-inserted are always appended to the header
350 return self._headers[:]
352 def get(self, name, failobj=None):
353 """Get a header value.
355 Like __getitem__() but return failobj instead of None when the field
359 for k, v in self._headers:
360 if k.lower() == name:
365 # Additional useful stuff
368 def get_all(self, name, failobj=None):
369 """Return a list of all the values for the named field.
371 These will be sorted in the order they appeared in the original
372 message, and may contain duplicates. Any fields deleted and
373 re-inserted are always appended to the header list.
375 If no such fields exist, failobj is returned (defaults to None).
379 for k, v in self._headers:
380 if k.lower() == name:
386 def add_header(self, _name, _value, **_params):
387 """Extended header setting.
389 name is the header field to add. keyword arguments can be used to set
390 additional parameters for the header field, with underscores converted
391 to dashes. Normally the parameter will be added as key="value" unless
392 value is None, in which case only the key will be added.
396 msg.add_header('content-disposition', 'attachment', filename='bud.gif')
399 for k, v in _params.items():
401 parts.append(k.replace('_', '-'))
403 parts.append(_formatparam(k.replace('_', '-'), v))
404 if _value is not None:
405 parts.insert(0, _value)
406 self._headers.append((_name, SEMISPACE.join(parts)))
408 def replace_header(self, _name, _value):
411 Replace the first matching header found in the message, retaining
412 header order and case. If no matching header was found, a KeyError is
415 _name = _name.lower()
416 for i, (k, v) in zip(range(len(self._headers)), self._headers):
417 if k.lower() == _name:
418 self._headers[i] = (k, _value)
421 raise KeyError(_name)
424 # Use these three methods instead of the three above.
427 def get_content_type(self):
428 """Return the message's content type.
430 The returned string is coerced to lower case of the form
431 `maintype/subtype'. If there was no Content-Type header in the
432 message, the default type as given by get_default_type() will be
433 returned. Since according to RFC 2045, messages always have a default
434 type this will always return a value.
436 RFC 2045 defines a message's default type to be text/plain unless it
437 appears inside a multipart/digest container, in which case it would be
441 value = self.get('content-type', missing)
443 # This should have no parameters
444 return self.get_default_type()
445 ctype = _splitparam(value)[0].lower()
446 # RFC 2045, section 5.2 says if its invalid, use text/plain
447 if ctype.count('/') != 1:
451 def get_content_maintype(self):
452 """Return the message's main content type.
454 This is the `maintype' part of the string returned by
457 ctype = self.get_content_type()
458 return ctype.split('/')[0]
460 def get_content_subtype(self):
461 """Returns the message's sub-content type.
463 This is the `subtype' part of the string returned by
466 ctype = self.get_content_type()
467 return ctype.split('/')[1]
469 def get_default_type(self):
470 """Return the `default' content type.
472 Most messages have a default content type of text/plain, except for
473 messages that are subparts of multipart/digest containers. Such
474 subparts have a default content type of message/rfc822.
476 return self._default_type
478 def set_default_type(self, ctype):
479 """Set the `default' content type.
481 ctype should be either "text/plain" or "message/rfc822", although this
482 is not enforced. The default content type is not stored in the
485 self._default_type = ctype
487 def _get_params_preserve(self, failobj, header):
488 # Like get_params() but preserves the quoting of values. BAW:
489 # should this be part of the public interface?
491 value = self.get(header, missing)
495 for p in _parseparam(';' + value):
497 name, val = p.split('=', 1)
501 # Must have been a bare attribute
504 params.append((name, val))
505 params = utils.decode_params(params)
508 def get_params(self, failobj=None, header='content-type', unquote=True):
509 """Return the message's Content-Type parameters, as a list.
511 The elements of the returned list are 2-tuples of key/value pairs, as
512 split on the `=' sign. The left hand side of the `=' is the key,
513 while the right hand side is the value. If there is no `=' sign in
514 the parameter the value is the empty string. The value is as
515 described in the get_param() method.
517 Optional failobj is the object to return if there is no Content-Type
518 header. Optional header is the header to search instead of
519 Content-Type. If unquote is True, the value is unquoted.
522 params = self._get_params_preserve(missing, header)
523 if params is missing:
526 return [(k, _unquotevalue(v)) for k, v in params]
530 def get_param(self, param, failobj=None, header='content-type',
532 """Return the parameter value if found in the Content-Type header.
534 Optional failobj is the object to return if there is no Content-Type
535 header, or the Content-Type header has no such parameter. Optional
536 header is the header to search instead of Content-Type.
538 Parameter keys are always compared case insensitively. The return
539 value can either be a string, or a 3-tuple if the parameter was RFC
540 2231 encoded. When it's a 3-tuple, the elements of the value are of
541 the form (CHARSET, LANGUAGE, VALUE). Note that both CHARSET and
542 LANGUAGE can be None, in which case you should consider VALUE to be
543 encoded in the us-ascii charset. You can usually ignore LANGUAGE.
545 Your application should be prepared to deal with 3-tuple return
546 values, and can convert the parameter to a Unicode string like so:
548 param = msg.get_param('foo')
549 if isinstance(param, tuple):
550 param = unicode(param[2], param[0] or 'us-ascii')
552 In any case, the parameter value (either the returned string, or the
553 VALUE item in the 3-tuple) is always unquoted, unless unquote is set
556 if header not in self:
558 for k, v in self._get_params_preserve(failobj, header):
559 if k.lower() == param.lower():
561 return _unquotevalue(v)
566 def set_param(self, param, value, header='Content-Type', requote=True,
567 charset=None, language=''):
568 """Set a parameter in the Content-Type header.
570 If the parameter already exists in the header, its value will be
571 replaced with the new value.
573 If header is Content-Type and has not yet been defined for this
574 message, it will be set to "text/plain" and the new parameter and
575 value will be appended as per RFC 2045.
577 An alternate header can specified in the header argument, and all
578 parameters will be quoted as necessary unless requote is False.
580 If charset is specified, the parameter will be encoded according to RFC
581 2231. Optional language specifies the RFC 2231 language, defaulting
582 to the empty string. Both charset and language should be strings.
584 if not isinstance(value, tuple) and charset:
585 value = (charset, language, value)
587 if header not in self and header.lower() == 'content-type':
590 ctype = self.get(header)
591 if not self.get_param(param, header=header):
593 ctype = _formatparam(param, value, requote)
595 ctype = SEMISPACE.join(
596 [ctype, _formatparam(param, value, requote)])
599 for old_param, old_value in self.get_params(header=header,
602 if old_param.lower() == param.lower():
603 append_param = _formatparam(param, value, requote)
605 append_param = _formatparam(old_param, old_value, requote)
609 ctype = SEMISPACE.join([ctype, append_param])
610 if ctype != self.get(header):
614 def del_param(self, param, header='content-type', requote=True):
615 """Remove the given parameter completely from the Content-Type header.
617 The header will be re-written in place without the parameter or its
618 value. All values will be quoted as necessary unless requote is
619 False. Optional header specifies an alternative to the Content-Type
622 if header not in self:
625 for p, v in self.get_params(header=header, unquote=requote):
626 if p.lower() != param.lower():
628 new_ctype = _formatparam(p, v, requote)
630 new_ctype = SEMISPACE.join([new_ctype,
631 _formatparam(p, v, requote)])
632 if new_ctype != self.get(header):
634 self[header] = new_ctype
636 def set_type(self, type, header='Content-Type', requote=True):
637 """Set the main type and subtype for the Content-Type header.
639 type must be a string in the form "maintype/subtype", otherwise a
640 ValueError is raised.
642 This method replaces the Content-Type header, keeping all the
643 parameters in place. If requote is False, this leaves the existing
644 header's quoting as is. Otherwise, the parameters will be quoted (the
647 An alternative header can be specified in the header argument. When
648 the Content-Type header is set, we'll always also add a MIME-Version
651 # BAW: should we be strict?
652 if not type.count('/') == 1:
654 # Set the Content-Type, you get a MIME-Version
655 if header.lower() == 'content-type':
656 del self['mime-version']
657 self['MIME-Version'] = '1.0'
658 if header not in self:
661 params = self.get_params(header=header, unquote=requote)
664 # Skip the first param; it's the old type.
665 for p, v in params[1:]:
666 self.set_param(p, v, header, requote)
668 def get_filename(self, failobj=None):
669 """Return the filename associated with the payload if present.
671 The filename is extracted from the Content-Disposition header's
672 `filename' parameter, and it is unquoted. If that header is missing
673 the `filename' parameter, this method falls back to looking for the
677 filename = self.get_param('filename', missing, 'content-disposition')
678 if filename is missing:
679 filename = self.get_param('name', missing, 'content-type')
680 if filename is missing:
682 return utils.collapse_rfc2231_value(filename).strip()
684 def get_boundary(self, failobj=None):
685 """Return the boundary associated with the payload if present.
687 The boundary is extracted from the Content-Type header's `boundary'
688 parameter, and it is unquoted.
691 boundary = self.get_param('boundary', missing)
692 if boundary is missing:
694 # RFC 2046 says that boundaries may begin but not end in w/s
695 return utils.collapse_rfc2231_value(boundary).rstrip()
697 def set_boundary(self, boundary):
698 """Set the boundary parameter in Content-Type to 'boundary'.
700 This is subtly different than deleting the Content-Type header and
701 adding a new one with a new boundary parameter via add_header(). The
702 main difference is that using the set_boundary() method preserves the
703 order of the Content-Type header in the original message.
705 HeaderParseError is raised if the message has no Content-Type header.
708 params = self._get_params_preserve(missing, 'content-type')
709 if params is missing:
710 # There was no Content-Type header, and we don't know what type
711 # to set it to, so raise an exception.
712 raise errors.HeaderParseError('No Content-Type header found')
715 for pk, pv in params:
716 if pk.lower() == 'boundary':
717 newparams.append(('boundary', '"%s"' % boundary))
720 newparams.append((pk, pv))
722 # The original Content-Type header had no boundary attribute.
723 # Tack one on the end. BAW: should we raise an exception
725 newparams.append(('boundary', '"%s"' % boundary))
726 # Replace the existing Content-Type header with the new value
728 for h, v in self._headers:
729 if h.lower() == 'content-type':
731 for k, v in newparams:
735 parts.append('%s=%s' % (k, v))
736 newheaders.append((h, SEMISPACE.join(parts)))
739 newheaders.append((h, v))
740 self._headers = newheaders
742 def get_content_charset(self, failobj=None):
743 """Return the charset parameter of the Content-Type header.
745 The returned string is always coerced to lower case. If there is no
746 Content-Type header, or if that header has no charset parameter,
750 charset = self.get_param('charset', missing)
751 if charset is missing:
753 if isinstance(charset, tuple):
754 # RFC 2231 encoded, so decode it, and it better end up as ascii.
755 pcharset = charset[0] or 'us-ascii'
757 # LookupError will be raised if the charset isn't known to
758 # Python. UnicodeError will be raised if the encoded text
759 # contains a character not in the charset.
760 charset = unicode(charset[2], pcharset).encode('us-ascii')
761 except (LookupError, UnicodeError):
763 # charset character must be in us-ascii range
765 if isinstance(charset, str):
766 charset = unicode(charset, 'us-ascii')
767 charset = charset.encode('us-ascii')
770 # RFC 2046, $4.1.2 says charsets are not case sensitive
771 return charset.lower()
773 def get_charsets(self, failobj=None):
774 """Return a list containing the charset(s) used in this message.
776 The returned list of items describes the Content-Type headers'
777 charset parameter for this message and all the subparts in its
780 Each item will either be a string (the value of the charset parameter
781 in the Content-Type header of that part) or the value of the
782 'failobj' parameter (defaults to None), if the part does not have a
783 main MIME type of "text", or the charset is not defined.
785 The list will contain one string for each part of the message, plus
786 one for the container message (i.e. self), so that a non-multipart
787 message will still return a list of length 1.
789 return [part.get_content_charset(failobj) for part in self.walk()]
791 # I.e. def walk(self): ...
792 from email.iterators import walk