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 used to split header parameters. BAW: this may be too
23 # simple. It isn't strictly RFC 2045 (section 5.1) compliant, but it catches
24 # most headers found in the wild. We may eventually need a full fledged
26 paramre
= re
.compile(r
'\s*;\s*')
27 # Regular expression that matches `special' characters in parameters, the
28 # existance of which force quoting of the parameter value.
29 tspecials
= re
.compile(r
'[ \(\)<>@,;:\\"/\[\]\?=]')
34 def _formatparam(param
, value
=None, quote
=True):
35 """Convenience function to format and return a key=value pair.
37 This will quote the value if needed or if quote is true.
39 if value
is not None and len(value
) > 0:
40 # A tuple is used for RFC 2231 encoded parameter values where items
41 # are (charset, language, value). charset is a string, not a Charset
43 if isinstance(value
, tuple):
44 # Encode as per RFC 2231
46 value
= utils
.encode_rfc2231(value
[2], value
[0], value
[1])
47 # BAW: Please check this. I think that if quote is set it should
48 # force quoting even if not necessary.
49 if quote
or tspecials
.search(value
):
50 return '%s="%s"' % (param
, utils
.quote(value
))
52 return '%s=%s' % (param
, value
)
61 while end
> 0 and s
.count('"', 0, end
) % 2:
62 end
= s
.find(';', end
+ 1)
68 f
= f
[:i
].strip().lower() + '=' + f
[i
+1:].strip()
69 plist
.append(f
.strip())
74 def _unquotevalue(value
):
75 # This is different than utils.collapse_rfc2231_value() because it doesn't
76 # try to convert the value to a unicode. Message.get_param() and
77 # Message.get_params() are both currently defined to return the tuple in
78 # the face of RFC 2231 parameters.
79 if isinstance(value
, tuple):
80 return value
[0], value
[1], utils
.unquote(value
[2])
82 return utils
.unquote(value
)
87 """Basic message object.
89 A message object is defined as something that has a bunch of RFC 2822
90 headers and a payload. It may optionally have an envelope header
91 (a.k.a. Unix-From or From_ header). If the message is a container (i.e. a
92 multipart or a message/rfc822), then the payload is a list of Message
93 objects, otherwise it is a string.
95 Message objects implement part of the `mapping' interface, which assumes
96 there is exactly one occurrance of the header per message. Some headers
97 do in fact appear multiple times (e.g. Received) and for those headers,
98 you must use the explicit API to set or get all the headers. Not all of
99 the mapping methods are implemented.
103 self
._unixfrom
= None
106 # Defaults for multipart messages
107 self
.preamble
= self
.epilogue
= None
109 # Default content type
110 self
._default
_type
= 'text/plain'
113 """Return the entire formatted message as a string.
114 This includes the headers, body, and envelope header.
116 return self
.as_string(unixfrom
=True)
118 def as_string(self
, unixfrom
=False):
119 """Return the entire formatted message as a string.
120 Optional `unixfrom' when True, means include the Unix From_ envelope
123 This is a convenience method and may not generate the message exactly
124 as you intend because by default it mangles lines that begin with
125 "From ". For more flexibility, use the flatten() method of a
128 from email
.Generator
import Generator
131 g
.flatten(self
, unixfrom
=unixfrom
)
134 def is_multipart(self
):
135 """Return True if the message consists of multiple parts."""
136 return isinstance(self
._payload
, list)
141 def set_unixfrom(self
, unixfrom
):
142 self
._unixfrom
= unixfrom
144 def get_unixfrom(self
):
145 return self
._unixfrom
148 # Payload manipulation.
150 def attach(self
, payload
):
151 """Add the given payload to the current payload.
153 The current payload will always be a list of objects after this method
154 is called. If you want to set the payload to a scalar object, use
155 set_payload() instead.
157 if self
._payload
is None:
158 self
._payload
= [payload
]
160 self
._payload
.append(payload
)
162 def get_payload(self
, i
=None, decode
=False):
163 """Return a reference to the payload.
165 The payload will either be a list object or a string. If you mutate
166 the list object, you modify the message's payload in place. Optional
167 i returns that index into the payload.
169 Optional decode is a flag indicating whether the payload should be
170 decoded or not, according to the Content-Transfer-Encoding header
173 When True and the message is not a multipart, the payload will be
174 decoded if this header's value is `quoted-printable' or `base64'. If
175 some other encoding is used, or the header is missing, or if the
176 payload has bogus data (i.e. bogus base64 or uuencoded data), the
177 payload is returned as-is.
179 If the message is a multipart and the decode flag is True, then None
183 payload
= self
._payload
184 elif not isinstance(self
._payload
, list):
185 raise TypeError('Expected list, got %s' % type(self
._payload
))
187 payload
= self
._payload
[i
]
189 if self
.is_multipart():
191 cte
= self
.get('content-transfer-encoding', '').lower()
192 if cte
== 'quoted-printable':
193 return utils
._qdecode
(payload
)
194 elif cte
== 'base64':
196 return utils
._bdecode
(payload
)
197 except binascii
.Error
:
200 elif cte
in ('x-uuencode', 'uuencode', 'uue', 'x-uue'):
203 uu
.decode(StringIO(payload
+'\n'), sfp
, quiet
=True)
204 payload
= sfp
.getvalue()
206 # Some decoding problem
208 # Everything else, including encodings with 8bit or 7bit are returned
212 def set_payload(self
, payload
, charset
=None):
213 """Set the payload to the given value.
215 Optional charset sets the message's default character set. See
216 set_charset() for details.
218 self
._payload
= payload
219 if charset
is not None:
220 self
.set_charset(charset
)
222 def set_charset(self
, charset
):
223 """Set the charset of the payload to a given character set.
225 charset can be a Charset instance, a string naming a character set, or
226 None. If it is a string it will be converted to a Charset instance.
227 If charset is None, the charset parameter will be removed from the
228 Content-Type field. Anything else will generate a TypeError.
230 The message will be assumed to be of type text/* encoded with
231 charset.input_charset. It will be converted to charset.output_charset
232 and encoded properly, if needed, when generating the plain text
233 representation of the message. MIME headers (MIME-Version,
234 Content-Type, Content-Transfer-Encoding) will be added as needed.
238 self
.del_param('charset')
241 if isinstance(charset
, basestring
):
242 charset
= email
.charset
.Charset(charset
)
243 if not isinstance(charset
, email
.charset
.Charset
):
244 raise TypeError(charset
)
245 # BAW: should we accept strings that can serve as arguments to the
246 # Charset constructor?
247 self
._charset
= charset
248 if not self
.has_key('MIME-Version'):
249 self
.add_header('MIME-Version', '1.0')
250 if not self
.has_key('Content-Type'):
251 self
.add_header('Content-Type', 'text/plain',
252 charset
=charset
.get_output_charset())
254 self
.set_param('charset', charset
.get_output_charset())
255 if str(charset
) != charset
.get_output_charset():
256 self
._payload
= charset
.body_encode(self
._payload
)
257 if not self
.has_key('Content-Transfer-Encoding'):
258 cte
= charset
.get_body_encoding()
262 self
._payload
= charset
.body_encode(self
._payload
)
263 self
.add_header('Content-Transfer-Encoding', cte
)
265 def get_charset(self
):
266 """Return the Charset instance associated with the message's payload.
271 # MAPPING INTERFACE (partial)
274 """Return the total number of headers, including duplicates."""
275 return len(self
._headers
)
277 def __getitem__(self
, name
):
278 """Get a header value.
280 Return None if the header is missing instead of raising an exception.
282 Note that if the header appeared multiple times, exactly which
283 occurrance gets returned is undefined. Use get_all() to get all
284 the values matching a header field name.
286 return self
.get(name
)
288 def __setitem__(self
, name
, val
):
289 """Set the value of a header.
291 Note: this does not overwrite an existing header with the same field
292 name. Use __delitem__() first to delete any existing headers.
294 self
._headers
.append((name
, val
))
296 def __delitem__(self
, name
):
297 """Delete all occurrences of a header, if present.
299 Does not raise an exception if the header is missing.
303 for k
, v
in self
._headers
:
304 if k
.lower() != name
:
305 newheaders
.append((k
, v
))
306 self
._headers
= newheaders
308 def __contains__(self
, name
):
309 return name
.lower() in [k
.lower() for k
, v
in self
._headers
]
311 def has_key(self
, name
):
312 """Return true if the message contains the header."""
314 return self
.get(name
, missing
) is not missing
317 """Return a list of all the message's header field names.
319 These will be sorted in the order they appeared in the original
320 message, or were added to the message, and may contain duplicates.
321 Any fields deleted and re-inserted are always appended to the header
324 return [k
for k
, v
in self
._headers
]
327 """Return a list of all the message's header values.
329 These will be sorted in the order they appeared in the original
330 message, or were added to the message, and may contain duplicates.
331 Any fields deleted and re-inserted are always appended to the header
334 return [v
for k
, v
in self
._headers
]
337 """Get all the message's header fields and values.
339 These will be sorted in the order they appeared in the original
340 message, or were added to the message, and may contain duplicates.
341 Any fields deleted and re-inserted are always appended to the header
344 return self
._headers
[:]
346 def get(self
, name
, failobj
=None):
347 """Get a header value.
349 Like __getitem__() but return failobj instead of None when the field
353 for k
, v
in self
._headers
:
354 if k
.lower() == name
:
359 # Additional useful stuff
362 def get_all(self
, name
, failobj
=None):
363 """Return a list of all the values for the named field.
365 These will be sorted in the order they appeared in the original
366 message, and may contain duplicates. Any fields deleted and
367 re-inserted are always appended to the header list.
369 If no such fields exist, failobj is returned (defaults to None).
373 for k
, v
in self
._headers
:
374 if k
.lower() == name
:
380 def add_header(self
, _name
, _value
, **_params
):
381 """Extended header setting.
383 name is the header field to add. keyword arguments can be used to set
384 additional parameters for the header field, with underscores converted
385 to dashes. Normally the parameter will be added as key="value" unless
386 value is None, in which case only the key will be added.
390 msg.add_header('content-disposition', 'attachment', filename='bud.gif')
393 for k
, v
in _params
.items():
395 parts
.append(k
.replace('_', '-'))
397 parts
.append(_formatparam(k
.replace('_', '-'), v
))
398 if _value
is not None:
399 parts
.insert(0, _value
)
400 self
._headers
.append((_name
, SEMISPACE
.join(parts
)))
402 def replace_header(self
, _name
, _value
):
405 Replace the first matching header found in the message, retaining
406 header order and case. If no matching header was found, a KeyError is
409 _name
= _name
.lower()
410 for i
, (k
, v
) in zip(range(len(self
._headers
)), self
._headers
):
411 if k
.lower() == _name
:
412 self
._headers
[i
] = (k
, _value
)
415 raise KeyError(_name
)
418 # Use these three methods instead of the three above.
421 def get_content_type(self
):
422 """Return the message's content type.
424 The returned string is coerced to lower case of the form
425 `maintype/subtype'. If there was no Content-Type header in the
426 message, the default type as given by get_default_type() will be
427 returned. Since according to RFC 2045, messages always have a default
428 type this will always return a value.
430 RFC 2045 defines a message's default type to be text/plain unless it
431 appears inside a multipart/digest container, in which case it would be
435 value
= self
.get('content-type', missing
)
437 # This should have no parameters
438 return self
.get_default_type()
439 ctype
= paramre
.split(value
)[0].lower().strip()
440 # RFC 2045, section 5.2 says if its invalid, use text/plain
441 if ctype
.count('/') != 1:
445 def get_content_maintype(self
):
446 """Return the message's main content type.
448 This is the `maintype' part of the string returned by
451 ctype
= self
.get_content_type()
452 return ctype
.split('/')[0]
454 def get_content_subtype(self
):
455 """Returns the message's sub-content type.
457 This is the `subtype' part of the string returned by
460 ctype
= self
.get_content_type()
461 return ctype
.split('/')[1]
463 def get_default_type(self
):
464 """Return the `default' content type.
466 Most messages have a default content type of text/plain, except for
467 messages that are subparts of multipart/digest containers. Such
468 subparts have a default content type of message/rfc822.
470 return self
._default
_type
472 def set_default_type(self
, ctype
):
473 """Set the `default' content type.
475 ctype should be either "text/plain" or "message/rfc822", although this
476 is not enforced. The default content type is not stored in the
479 self
._default
_type
= ctype
481 def _get_params_preserve(self
, failobj
, header
):
482 # Like get_params() but preserves the quoting of values. BAW:
483 # should this be part of the public interface?
485 value
= self
.get(header
, missing
)
489 for p
in _parseparam(';' + value
):
491 name
, val
= p
.split('=', 1)
495 # Must have been a bare attribute
498 params
.append((name
, val
))
499 params
= utils
.decode_params(params
)
502 def get_params(self
, failobj
=None, header
='content-type', unquote
=True):
503 """Return the message's Content-Type parameters, as a list.
505 The elements of the returned list are 2-tuples of key/value pairs, as
506 split on the `=' sign. The left hand side of the `=' is the key,
507 while the right hand side is the value. If there is no `=' sign in
508 the parameter the value is the empty string. The value is as
509 described in the get_param() method.
511 Optional failobj is the object to return if there is no Content-Type
512 header. Optional header is the header to search instead of
513 Content-Type. If unquote is True, the value is unquoted.
516 params
= self
._get
_params
_preserve
(missing
, header
)
517 if params
is missing
:
520 return [(k
, _unquotevalue(v
)) for k
, v
in params
]
524 def get_param(self
, param
, failobj
=None, header
='content-type',
526 """Return the parameter value if found in the Content-Type header.
528 Optional failobj is the object to return if there is no Content-Type
529 header, or the Content-Type header has no such parameter. Optional
530 header is the header to search instead of Content-Type.
532 Parameter keys are always compared case insensitively. The return
533 value can either be a string, or a 3-tuple if the parameter was RFC
534 2231 encoded. When it's a 3-tuple, the elements of the value are of
535 the form (CHARSET, LANGUAGE, VALUE). Note that both CHARSET and
536 LANGUAGE can be None, in which case you should consider VALUE to be
537 encoded in the us-ascii charset. You can usually ignore LANGUAGE.
539 Your application should be prepared to deal with 3-tuple return
540 values, and can convert the parameter to a Unicode string like so:
542 param = msg.get_param('foo')
543 if isinstance(param, tuple):
544 param = unicode(param[2], param[0] or 'us-ascii')
546 In any case, the parameter value (either the returned string, or the
547 VALUE item in the 3-tuple) is always unquoted, unless unquote is set
550 if not self
.has_key(header
):
552 for k
, v
in self
._get
_params
_preserve
(failobj
, header
):
553 if k
.lower() == param
.lower():
555 return _unquotevalue(v
)
560 def set_param(self
, param
, value
, header
='Content-Type', requote
=True,
561 charset
=None, language
=''):
562 """Set a parameter in the Content-Type header.
564 If the parameter already exists in the header, its value will be
565 replaced with the new value.
567 If header is Content-Type and has not yet been defined for this
568 message, it will be set to "text/plain" and the new parameter and
569 value will be appended as per RFC 2045.
571 An alternate header can specified in the header argument, and all
572 parameters will be quoted as necessary unless requote is False.
574 If charset is specified, the parameter will be encoded according to RFC
575 2231. Optional language specifies the RFC 2231 language, defaulting
576 to the empty string. Both charset and language should be strings.
578 if not isinstance(value
, tuple) and charset
:
579 value
= (charset
, language
, value
)
581 if not self
.has_key(header
) and header
.lower() == 'content-type':
584 ctype
= self
.get(header
)
585 if not self
.get_param(param
, header
=header
):
587 ctype
= _formatparam(param
, value
, requote
)
589 ctype
= SEMISPACE
.join(
590 [ctype
, _formatparam(param
, value
, requote
)])
593 for old_param
, old_value
in self
.get_params(header
=header
,
596 if old_param
.lower() == param
.lower():
597 append_param
= _formatparam(param
, value
, requote
)
599 append_param
= _formatparam(old_param
, old_value
, requote
)
603 ctype
= SEMISPACE
.join([ctype
, append_param
])
604 if ctype
!= self
.get(header
):
608 def del_param(self
, param
, header
='content-type', requote
=True):
609 """Remove the given parameter completely from the Content-Type header.
611 The header will be re-written in place without the parameter or its
612 value. All values will be quoted as necessary unless requote is
613 False. Optional header specifies an alternative to the Content-Type
616 if not self
.has_key(header
):
619 for p
, v
in self
.get_params(header
=header
, unquote
=requote
):
620 if p
.lower() != param
.lower():
622 new_ctype
= _formatparam(p
, v
, requote
)
624 new_ctype
= SEMISPACE
.join([new_ctype
,
625 _formatparam(p
, v
, requote
)])
626 if new_ctype
!= self
.get(header
):
628 self
[header
] = new_ctype
630 def set_type(self
, type, header
='Content-Type', requote
=True):
631 """Set the main type and subtype for the Content-Type header.
633 type must be a string in the form "maintype/subtype", otherwise a
634 ValueError is raised.
636 This method replaces the Content-Type header, keeping all the
637 parameters in place. If requote is False, this leaves the existing
638 header's quoting as is. Otherwise, the parameters will be quoted (the
641 An alternative header can be specified in the header argument. When
642 the Content-Type header is set, we'll always also add a MIME-Version
645 # BAW: should we be strict?
646 if not type.count('/') == 1:
648 # Set the Content-Type, you get a MIME-Version
649 if header
.lower() == 'content-type':
650 del self
['mime-version']
651 self
['MIME-Version'] = '1.0'
652 if not self
.has_key(header
):
655 params
= self
.get_params(header
=header
, unquote
=requote
)
658 # Skip the first param; it's the old type.
659 for p
, v
in params
[1:]:
660 self
.set_param(p
, v
, header
, requote
)
662 def get_filename(self
, failobj
=None):
663 """Return the filename associated with the payload if present.
665 The filename is extracted from the Content-Disposition header's
666 `filename' parameter, and it is unquoted. If that header is missing
667 the `filename' parameter, this method falls back to looking for the
671 filename
= self
.get_param('filename', missing
, 'content-disposition')
672 if filename
is missing
:
673 filename
= self
.get_param('name', missing
, 'content-disposition')
674 if filename
is missing
:
676 return utils
.collapse_rfc2231_value(filename
).strip()
678 def get_boundary(self
, failobj
=None):
679 """Return the boundary associated with the payload if present.
681 The boundary is extracted from the Content-Type header's `boundary'
682 parameter, and it is unquoted.
685 boundary
= self
.get_param('boundary', missing
)
686 if boundary
is missing
:
688 # RFC 2046 says that boundaries may begin but not end in w/s
689 return utils
.collapse_rfc2231_value(boundary
).rstrip()
691 def set_boundary(self
, boundary
):
692 """Set the boundary parameter in Content-Type to 'boundary'.
694 This is subtly different than deleting the Content-Type header and
695 adding a new one with a new boundary parameter via add_header(). The
696 main difference is that using the set_boundary() method preserves the
697 order of the Content-Type header in the original message.
699 HeaderParseError is raised if the message has no Content-Type header.
702 params
= self
._get
_params
_preserve
(missing
, 'content-type')
703 if params
is missing
:
704 # There was no Content-Type header, and we don't know what type
705 # to set it to, so raise an exception.
706 raise errors
.HeaderParseError('No Content-Type header found')
709 for pk
, pv
in params
:
710 if pk
.lower() == 'boundary':
711 newparams
.append(('boundary', '"%s"' % boundary
))
714 newparams
.append((pk
, pv
))
716 # The original Content-Type header had no boundary attribute.
717 # Tack one on the end. BAW: should we raise an exception
719 newparams
.append(('boundary', '"%s"' % boundary
))
720 # Replace the existing Content-Type header with the new value
722 for h
, v
in self
._headers
:
723 if h
.lower() == 'content-type':
725 for k
, v
in newparams
:
729 parts
.append('%s=%s' % (k
, v
))
730 newheaders
.append((h
, SEMISPACE
.join(parts
)))
733 newheaders
.append((h
, v
))
734 self
._headers
= newheaders
736 def get_content_charset(self
, failobj
=None):
737 """Return the charset parameter of the Content-Type header.
739 The returned string is always coerced to lower case. If there is no
740 Content-Type header, or if that header has no charset parameter,
744 charset
= self
.get_param('charset', missing
)
745 if charset
is missing
:
747 if isinstance(charset
, tuple):
748 # RFC 2231 encoded, so decode it, and it better end up as ascii.
749 pcharset
= charset
[0] or 'us-ascii'
751 # LookupError will be raised if the charset isn't known to
752 # Python. UnicodeError will be raised if the encoded text
753 # contains a character not in the charset.
754 charset
= unicode(charset
[2], pcharset
).encode('us-ascii')
755 except (LookupError, UnicodeError):
757 # charset character must be in us-ascii range
759 if isinstance(charset
, str):
760 charset
= unicode(charset
, 'us-ascii')
761 charset
= charset
.encode('us-ascii')
764 # RFC 2046, $4.1.2 says charsets are not case sensitive
765 return charset
.lower()
767 def get_charsets(self
, failobj
=None):
768 """Return a list containing the charset(s) used in this message.
770 The returned list of items describes the Content-Type headers'
771 charset parameter for this message and all the subparts in its
774 Each item will either be a string (the value of the charset parameter
775 in the Content-Type header of that part) or the value of the
776 'failobj' parameter (defaults to None), if the part does not have a
777 main MIME type of "text", or the charset is not defined.
779 The list will contain one string for each part of the message, plus
780 one for the container message (i.e. self), so that a non-multipart
781 message will still return a list of length 1.
783 return [part
.get_content_charset(failobj
) for part
in self
.walk()]
785 # I.e. def walk(self): ...
786 from email
.Iterators
import walk