Lib/rfc822.py

   1 # RFC-822 message manipulation class.
   2 #
   3 # XXX This is only a very rough sketch of a full RFC-822 parser;
   4 # in particular the tokenizing of addresses does not adhere to all the
   5 # quoting rules.
   6 #
   7 # Directions for use:
   8 #
   9 # To create a Message object: first open a file, e.g.:
  10 #   fp = open(file, 'r')
  11 # (or use any other legal way of getting an open file object, e.g. use
  12 # sys.stdin or call os.popen()).
  13 # Then pass the open file object to the Message() constructor:
  14 #   m = Message(fp)
  15 #
  16 # To get the text of a particular header there are several methods:
  17 #   str = m.getheader(name)
  18 #   str = m.getrawheader(name)
  19 # where name is the name of the header, e.g. 'Subject'.
  20 # The difference is that getheader() strips the leading and trailing
  21 # whitespace, while getrawheader() doesn't.  Both functions retain
  22 # embedded whitespace (including newlines) exactly as they are
  23 # specified in the header, and leave the case of the text unchanged.
  24 #
  25 # For addresses and address lists there are functions
  26 #   realname, mailaddress = m.getaddr(name) and
  27 #   list = m.getaddrlist(name)
  28 # where the latter returns a list of (realname, mailaddr) tuples.
  29 #
  30 # There is also a method
  31 #   time = m.getdate(name)
  32 # which parses a Date-like field and returns a time-compatible tuple,
  33 # i.e. a tuple such as returned by time.localtime() or accepted by
  34 # time.mktime().
  35 #
  36 # See the class definition for lower level access methods.
  37 #
  38 # There are also some utility functions here.
  39
  40
  41 import regex
  42 import string
  43 import time
  44
  45
  46 class Message:
  47
  48         # Initialize the class instance and read the headers.
  49
  50         def __init__(self, fp):
  51                 self.fp = fp
  52                 #
  53                 try:
  54                         self.startofheaders = self.fp.tell()
  55                 except IOError:
  56                         self.startofheaders = None
  57                 #
  58                 self.readheaders()
  59                 #
  60                 try:
  61                         self.startofbody = self.fp.tell()
  62                 except IOError:
  63                         self.startofbody = None
  64
  65
  66         # Rewind the file to the start of the body (if seekable).
  67
  68         def rewindbody(self):
  69                 self.fp.seek(self.startofbody)
  70
  71
  72         # Read header lines up to the entirely blank line that
  73         # terminates them.  The (normally blank) line that ends the
  74         # headers is skipped, but not included in the returned list.
  75         # If a non-header line ends the headers, (which is an error),
  76         # an attempt is made to backspace over it; it is never
  77         # included in the returned list.
  78         #
  79         # The variable self.status is set to the empty string if all
  80         # went well, otherwise it is an error message.
  81         # The variable self.headers is a completely uninterpreted list
  82         # of lines contained in the header (so printing them will
  83         # reproduce the header exactly as it appears in the file).
  84
  85         def readheaders(self):
  86                 self.headers = list = []
  87                 self.status = ''
  88                 headerseen = 0
  89                 while 1:
  90                         line = self.fp.readline()
  91                         if not line:
  92                                 self.status = 'EOF in headers'
  93                                 break
  94                         if self.islast(line):
  95                                 break
  96                         elif headerseen and line[0] in ' \t':
  97                                 # It's a continuation line.
  98                                 list.append(line)
  99                         elif regex.match('^[!-9;-~]+:', line) >= 0:
 100                                 # It's a header line.
 101                                 list.append(line)
 102                                 headerseen = 1
 103                         else:
 104                                 # It's not a header line; stop here.
 105                                 if not headerseen:
 106                                         self.status = 'No headers'
 107                                 else:
 108                                         self.status = 'Bad header'
 109                                 # Try to undo the read.
 110                                 try:
 111                                         self.fp.seek(-len(line), 1)
 112                                 except IOError:
 113                                         self.status = \
 114                                                 self.status + '; bad seek'
 115                                 break
 116
 117
 118         # Method to determine whether a line is a legal end of
 119         # RFC-822 headers.  You may override this method if your
 120         # application wants to bend the rules, e.g. to strip trailing
 121         # whitespace, or to recognise MH template separators
 122         # ('--------').  For convenience (e.g. for code reading from
 123         # sockets) a line consisting of \r\n also matches.
 124
 125         def islast(self, line):
 126                 return line == '\n' or line == '\r\n'
 127
 128
 129         # Look through the list of headers and find all lines matching
 130         # a given header name (and their continuation lines).
 131         # A list of the lines is returned, without interpretation.
 132         # If the header does not occur, an empty list is returned.
 133         # If the header occurs multiple times, all occurrences are
 134         # returned.  Case is not important in the header name.
 135
 136         def getallmatchingheaders(self, name):
 137                 name = string.lower(name) + ':'
 138                 n = len(name)
 139                 list = []
 140                 hit = 0
 141                 for line in self.headers:
 142                         if string.lower(line[:n]) == name:
 143                                 hit = 1
 144                         elif line[:1] not in string.whitespace:
 145                                 hit = 0
 146                         if hit:
 147                                 list.append(line)
 148                 return list
 149
 150
 151         # Similar, but return only the first matching header (and its
 152         # continuation lines).
 153
 154         def getfirstmatchingheader(self, name):
 155                 name = string.lower(name) + ':'
 156                 n = len(name)
 157                 list = []
 158                 hit = 0
 159                 for line in self.headers:
 160                         if hit:
 161                                 if line[:1] not in string.whitespace:
 162                                         break
 163                         elif string.lower(line[:n]) == name:
 164                                 hit = 1
 165                         if hit:
 166                                 list.append(line)
 167                 return list
 168
 169
 170         # A higher-level interface to getfirstmatchingheader().
 171         # Return a string containing the literal text of the header
 172         # but with the keyword stripped.  All leading, trailing and
 173         # embedded whitespace is kept in the string, however.
 174         # Return None if the header does not occur.
 175
 176         def getrawheader(self, name):
 177                 list = self.getfirstmatchingheader(name)
 178                 if not list:
 179                         return None
 180                 list[0] = list[0][len(name) + 1:]
 181                 return string.joinfields(list, '')
 182
 183
 184         # Going one step further: also strip leading and trailing
 185         # whitespace.
 186
 187         def getheader(self, name):
 188                 text = self.getrawheader(name)
 189                 if text == None:
 190                         return None
 191                 return string.strip(text)
 192
 193
 194         # Retrieve a single address from a header as a tuple, e.g.
 195         # ('Guido van Rossum', 'guido@cwi.nl').
 196
 197         def getaddr(self, name):
 198                 data = self.getheader(name)
 199                 if not data:
 200                         return None, None
 201                 return parseaddr(data)
 202
 203         # Retrieve a list of addresses from a header, where each
 204         # address is a tuple as returned by getaddr().
 205
 206         def getaddrlist(self, name):
 207                 # XXX This function is not really correct.  The split
 208                 # on ',' might fail in the case of commas within
 209                 # quoted strings.
 210                 data = self.getheader(name)
 211                 if not data:
 212                         return []
 213                 data = string.splitfields(data, ',')
 214                 for i in range(len(data)):
 215                         data[i] = parseaddr(data[i])
 216                 return data
 217
 218         # Retrieve a date field from a header as a tuple compatible
 219         # with time.mktime().
 220
 221         def getdate(self, name):
 222                 data = self.getheader(name)
 223                 if not data:
 224                         return None
 225                 return parsedate(data)
 226
 227
 228         # Access as a dictionary (only finds first header of each type):
 229
 230         def __len__(self):
 231                 types = {}
 232                 for line in self.headers:
 233                         if line[0] in string.whitespace: continue
 234                         i = string.find(line, ':')
 235                         if i > 0:
 236                                 name = string.lower(line[:i])
 237                                 types[name] = None
 238                 return len(types)
 239
 240         def __getitem__(self, name):
 241                 value = self.getheader(name)
 242                 if value is None: raise KeyError, name
 243                 return value
 244
 245         def has_key(self, name):
 246                 value = self.getheader(name)
 247                 return value is not None
 248
 249         def keys(self):
 250                 types = {}
 251                 for line in self.headers:
 252                         if line[0] in string.whitespace: continue
 253                         i = string.find(line, ':')
 254                         if i > 0:
 255                                 name = line[:i]
 256                                 key = string.lower(name)
 257                                 types[key] = name
 258                 return types.values()
 259
 260         def values(self):
 261                 values = []
 262                 for name in self.keys():
 263                         values.append(self[name])
 264                 return values
 265
 266         def items(self):
 267                 items = []
 268                 for name in self.keys():
 269                         items.append(name, self[name])
 270                 return items
 271
 272
 273
 274 # Utility functions
 275 # -----------------
 276
 277 # XXX Should fix these to be really conformant.
 278 # XXX The inverses of the parse functions may also be useful.
 279
 280
 281 # Remove quotes from a string.
 282
 283 def unquote(str):
 284         if len(str) > 1:
 285                 if str[0] == '"' and str[-1:] == '"':
 286                         return str[1:-1]
 287                 if str[0] == '<' and str[-1:] == '>':
 288                         return str[1:-1]
 289         return str
 290
 291
 292 # Parse an address into (name, address) tuple
 293
 294 def parseaddr(address):
 295         # This is probably not perfect
 296         address = string.strip(address)
 297         # Case 1: part of the address is in <xx@xx> form.
 298         pos = regex.search('<.*>', address)
 299         if pos >= 0:
 300                 name = address[:pos]
 301                 address = address[pos:]
 302                 length = regex.match('<.*>', address)
 303                 name = name + address[length:]
 304                 address = address[:length]
 305         else:
 306                 # Case 2: part of the address is in (comment) form
 307                 pos = regex.search('(.*)', address)
 308                 if pos >= 0:
 309                         name = address[pos:]
 310                         address = address[:pos]
 311                         length = regex.match('(.*)', name)
 312                         address = address + name[length:]
 313                         name = name[:length]
 314                 else:
 315                         # Case 3: neither. Only an address
 316                         name = ''
 317         name = string.strip(name)
 318         address = string.strip(address)
 319         if address and address[0] == '<' and address[-1] == '>':
 320                 address = address[1:-1]
 321         if name and name[0] == '(' and name[-1] == ')':
 322                 name = name[1:-1]
 323         return name, address
 324
 325
 326 # Parse a date field
 327
 328 _monthnames = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul',
 329           'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
 330
 331 def parsedate(data):
 332         # XXX This still mostly ignores timezone matters at the moment...
 333         data = string.split(data)
 334         if data[0][-1] == ',':
 335                 # There's a dayname here. Skip it
 336                 del data[0]
 337         if len(data) == 4:
 338                 s = data[3]
 339                 i = string.find(s, '+')
 340                 if i > 0:
 341                         data[3:] = [s[:i], s[i+1:]]
 342                 else:
 343                         data.append('') # Dummy tz
 344         if len(data) < 5:
 345                 return None
 346         data = data[:5]
 347         [dd, mm, yy, tm, tz] = data
 348         if not mm in _monthnames:
 349                 return None
 350         mm = _monthnames.index(mm)+1
 351         tm = string.splitfields(tm, ':')
 352         if len(tm) == 2:
 353                 [thh, tmm] = tm
 354                 tss = '0'
 355         else:
 356                 [thh, tmm, tss] = tm
 357         try:
 358                 yy = string.atoi(yy)
 359                 dd = string.atoi(dd)
 360                 thh = string.atoi(thh)
 361                 tmm = string.atoi(tmm)
 362                 tss = string.atoi(tss)
 363         except string.atoi_error:
 364                 return None
 365         tuple = (yy, mm, dd, thh, tmm, tss, 0, 0, 0)
 366         return tuple
 367
 368
 369 # When used as script, run a small test program.
 370 # The first command line argument must be a filename containing one
 371 # message in RFC-822 format.
 372
 373 if __name__ == '__main__':
 374         import sys
 375         file = '/ufs/guido/Mail/drafts/,1'
 376         if sys.argv[1:]: file = sys.argv[1]
 377         f = open(file, 'r')
 378         m = Message(f)
 379         print 'From:', m.getaddr('from')
 380         print 'To:', m.getaddrlist('to')
 381         print 'Subject:', m.getheader('subject')
 382         print 'Date:', m.getheader('date')
 383         date = m.getdate('date')
 384         if date:
 385                 print 'ParsedDate:', time.asctime(date)
 386         else:
 387                 print 'ParsedDate:', None
 388         m.rewindbody()
 389         n = 0
 390         while f.readline():
 391                 n = n + 1
 392         print 'Lines:', n
 393         print '-'*70
 394         print 'len =', len(m)
 395         if m.has_key('Date'): print 'Date =', m['Date']
 396         if m.has_key('X-Nonsense'): pass
 397         print 'keys =', m.keys()
 398         print 'values =', m.values()
 399         print 'items =', m.items()
 400