def adjust_content_type(content_type, body=None, filename=None):
"""Adjust content type based on filename or body contents
"""
if filename and str(content_type) == 'application/octet-stream':
# check if our internal guess returns anything
guessed = _guess_type(filename)
if guessed:
return guessed
# our internal attempt didn't return anything, use mimetypes
guessed = mimetypes.guess_type(filename)[0]
if guessed:
main, sub = fix_content_type(
guessed, default=('application', 'octet-stream'))
content_type = ContentType(main, sub)
if content_type.main == 'image' and body:
image_preamble = body[:32]
if six.PY3 and isinstance(body, six.text_type):
image_preamble = image_preamble.encode('utf-8', 'ignore')
sub = imghdr.what(None, image_preamble)
if sub:
content_type = ContentType('image', sub)
elif content_type.main == 'audio' and body:
sub = _email.detect_audio_type(body)
if sub:
content_type = ContentType('audio', sub)
return content_type
def adjust_content_type(content_type, body=None, filename=None):
"""Adjust content type based on filename or body contents
"""
if filename and str(content_type) == 'application/octet-stream':
# check if our internal guess returns anything
guessed = _guess_type(filename)
if guessed:
return guessed
# our internal attempt didn't return anything, use mimetypes
guessed = mimetypes.guess_type(filename)[0]
if guessed:
main, sub = fix_content_type(
guessed, default=('application', 'octet-stream'))
content_type = ContentType(main, sub)
if content_type.main == 'image' and body:
sub = imghdr.what(None, body)
if sub:
content_type = ContentType('image', sub)
elif content_type.main == 'audio' and body:
sub = audio._whatsnd(body)
if sub:
content_type = ContentType('audio', sub)
return content_type
def _guess_type(filename):
"""
Internal content type guesser. This is used to hard code certain tricky content-types
that heuristic content type checker get wrong.
"""
if filename.endswith(".bz2"):
return ContentType("application", "x-bzip2")
if filename.endswith(".gz"):
return ContentType("application", "x-gzip")
return None
def _guess_type(filename):
"""
Internal content type guesser. This is used to hard code certain tricky content-types
that heuristic content type checker get wrong.
"""
if filename.endswith('.bz2'):
return ContentType('application', 'x-bzip2')
if filename.endswith('.gz'):
return ContentType('application', 'x-gzip')
return None
def charset(self, value):
charset = value.lower()
self.content_type.set_charset(value)
if 'Content-Type' not in self.headers:
self.headers['Content-Type'] = ContentType('text', 'plain', {})
self.headers['Content-Type'].params['charset'] = charset
self.headers.changed = True
def grab_headers(pointer, iterator, parent):
"""This function collects all tokens till the boundary
or the end of the message. Used to scan parts of the message
that contain random headers, e.g. text/rfc822-headers"""
content_type = None
while True:
iterator.check()
end = iterator.next()
# remember the first content-type we have met when grabbing
# the headers until the boundary or message end
if not content_type and end.is_content_type():
content_type = end
if not end.is_content_type():
break
return make_part(
content_type=content_type or ContentType("text", "plain"),
start=pointer,
end=end,
iterator=iterator,
parent=parent)
def decode_charset(ctype, body):
if ctype.main != 'text':
return body, ctype
mime_type = magic.from_buffer(body, mime=True)
mime_type = ContentType(*mime_type.split("/", 1))
if mime_type.main != 'text':
return body, mime_type
charset = ctype.get_charset()
body = charsets.convert_to_unicode(charset, body)
# for text/html unicode bodies make sure to replace
# the whitespace (0xA0) with Outlook is reported to
# have a bug there
if ctype.sub =='html' and charset == 'utf-8':
# Outlook bug
body = body.replace(u'\xa0', u' ')
return body, ctype
def make_part(content_type,
start,
end,
iterator,
parts=(),
enclosed=None,
parent=None,
badmime=False):
# here we detect where the message really starts
# the exact position in the string, at the end of the
# starting boundary and after the beginning of the end boundary
if start.is_boundary():
start = start.end + 1
else:
start = start.start
# if this is the message ending, end of part
# the position of the last symbol of the message
if end.is_end():
end = len(iterator.string) - 1
# for multipart boundaries
# consider the final boundary as the ending one
elif content_type.is_multipart():
end = end.end
# otherwise, end is position of the the symbol before
# the boundary start
else:
end = end.start - 1
# our tokenizer detected the beginning of the message container
# that is separated from the enclosed message by newlines
# here we find where the enclosed message begins by searching for the
# first newline
if parent and (parent.is_message_container()
or parent.is_headers_container()):
start = locate_first_newline(iterator.stream, start)
if badmime:
content_type = ContentType('application', 'octet-stream')
# ok, finally, create the MimePart.
# note that it does not parse anything, just remembers
# the position in the string
return MimePart(container=Stream(content_type=content_type,
start=start,
end=end,
stream=iterator.stream,
string=iterator.string),
badmime=badmime,
parts=parts,
enclosed=enclosed,
is_root=(parent == None))
def adjust_content_type(content_type, body=None, filename=None):
"""Adjust content type based on filename or body contents
"""
if filename and str(content_type) == 'application/octet-stream':
guessed = mimetypes.guess_type(filename)[0]
if guessed:
main, sub = fix_content_type(guessed,
default=('application',
'octet-stream'))
content_type = ContentType(main, sub)
if content_type.main == 'image' and body:
sub = imghdr.what(None, body)
if sub:
content_type = ContentType('image', sub)
elif content_type.main == 'audio' and body:
sub = audio._whatsnd(body)
if sub:
content_type = ContentType('audio', sub)
return content_type
def default_content_type():
return ContentType("text", "plain", {'charset': 'ascii'})
# and its body.
^(\r\n|\n)
)
""", re.IGNORECASE | re.MULTILINE | re.VERBOSE)
_CTYPE = 'ctype'
_BOUNDARY = 'boundary'
_END = End()
_MAX_OPS = 500
_SECTION_HEADERS = 'headers'
_SECTION_MULTIPART_PREAMBLE = 'multipart-preamble'
_SECTION_MULTIPART_EPILOGUE = 'multipart-epilogue'
_SECTION_BODY = 'body'
_DEFAULT_CONTENT_TYPE = ContentType('text', 'plain', {'charset': 'us-ascii'})
_EMPTY_LINE = '\r\n'
def tokenize(string):
"""
Scans the entire message to find all Content-Types and boundaries.
"""
tokens = deque()
for m in _RE_TOKENIZER.finditer(string):
if m.group(_CTYPE):
name, token = parsing.parse_header(m.group(_CTYPE))
elif m.group(_BOUNDARY):
token = Boundary(
m.group(_BOUNDARY).strip("\t\r\n"),
_grab_newline(m.start(), string, -1),