|
| 1 | +# RFC-822 message manipulation class. |
| 2 | +# |
| 3 | +# XXX This is only a very rough sketch of a full RFC-822 parser; |
| 4 | +# additional methods are needed to parse addresses and dates, and to |
| 5 | +# tokenize lines according to various other syntax 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 init() method of Message: |
| 14 | +# m = Message().init(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 | +# See the class definition for lower level access methods. |
| 26 | +# |
| 27 | +# There are also some utility functions here. |
| 28 | + |
| 29 | + |
| 30 | +import regex |
| 31 | +import string |
| 32 | + |
| 33 | + |
| 34 | +class Message: |
| 35 | + |
| 36 | + # Initialize the class instance and read the headers. |
| 37 | + |
| 38 | + def init(self, fp): |
| 39 | + self.fp = fp |
| 40 | + # |
| 41 | + try: |
| 42 | + self.startofheaders = self.fp.tell() |
| 43 | + except IOError: |
| 44 | + self.startofheaders = None |
| 45 | + # |
| 46 | + self.readheaders() |
| 47 | + # |
| 48 | + try: |
| 49 | + self.startofbody = self.fp.tell() |
| 50 | + except IOError: |
| 51 | + self.startofbody = None |
| 52 | + # |
| 53 | + return self |
| 54 | + |
| 55 | + |
| 56 | + # Rewind the file to the start of the body (if seekable). |
| 57 | + |
| 58 | + def rewindbody(self): |
| 59 | + self.fp.seek(self.startofbody) |
| 60 | + |
| 61 | + |
| 62 | + # Read header lines up to the entirely blank line that |
| 63 | + # terminates them. The (normally blank) line that ends the |
| 64 | + # headers is skipped, but not included in the returned list. |
| 65 | + # If a non-header line ends the headers, (which is an error), |
| 66 | + # an attempt is made to backspace over it; it is never |
| 67 | + # included in the returned list. |
| 68 | + # |
| 69 | + # The variable self.status is set to the empty string if all |
| 70 | + # went well, otherwise it is an error message. |
| 71 | + # The variable self.headers is a completely uninterpreted list |
| 72 | + # of lines contained in the header (so printing them will |
| 73 | + # reproduce the header exactly as it appears in the file). |
| 74 | + |
| 75 | + def readheaders(self): |
| 76 | + self.headers = list = [] |
| 77 | + self.status = '' |
| 78 | + headerseen = 0 |
| 79 | + while 1: |
| 80 | + line = self.fp.readline() |
| 81 | + if not line: |
| 82 | + self.status = 'EOF in headers' |
| 83 | + break |
| 84 | + if self.islast(line): |
| 85 | + break |
| 86 | + elif headerseen and line[0] in ' \t': |
| 87 | + # It's a continuation line. |
| 88 | + list.append(line) |
| 89 | + elif regex.match('^[!-9;-~]+:', line): |
| 90 | + # It's a header line. |
| 91 | + list.append(line) |
| 92 | + headerseen = 1 |
| 93 | + else: |
| 94 | + # It's not a header line; stop here. |
| 95 | + if not headerseen: |
| 96 | + self.status = 'No headers' |
| 97 | + else: |
| 98 | + self.status = 'Bad header' |
| 99 | + # Try to undo the read. |
| 100 | + try: |
| 101 | + self.fp.seek(-len(line), 1) |
| 102 | + except IOError: |
| 103 | + self.status = \ |
| 104 | + self.status + '; bad seek' |
| 105 | + break |
| 106 | + |
| 107 | + |
| 108 | + # Method to determine whether a line is a legal end of |
| 109 | + # RFC-822 headers. You may override this method if your |
| 110 | + # application wants to bend the rules, e.g. to accept lines |
| 111 | + # ending in '\r\n', to strip trailing whitespace, or to |
| 112 | + # recognise MH template separators ('--------'). |
| 113 | + |
| 114 | + def islast(self, line): |
| 115 | + return line == '\n' |
| 116 | + |
| 117 | + |
| 118 | + # Look through the list of headers and find all lines matching |
| 119 | + # a given header name (and their continuation lines). |
| 120 | + # A list of the lines is returned, without interpretation. |
| 121 | + # If the header does not occur, an empty list is returned. |
| 122 | + # If the header occurs multiple times, all occurrences are |
| 123 | + # returned. Case is not important in the header name. |
| 124 | + |
| 125 | + def getallmatchingheaders(self, name): |
| 126 | + name = string.lower(name) + ':' |
| 127 | + n = len(name) |
| 128 | + list = [] |
| 129 | + hit = 0 |
| 130 | + for line in self.headers: |
| 131 | + if string.lower(line[:n]) == name: |
| 132 | + hit = 1 |
| 133 | + elif line[:1] not in string.whitespace: |
| 134 | + hit = 0 |
| 135 | + if hit: |
| 136 | + list.append(line) |
| 137 | + return list |
| 138 | + |
| 139 | + |
| 140 | + # Similar, but return only the first matching header (and its |
| 141 | + # continuation lines). |
| 142 | + |
| 143 | + def getfirstmatchingheader(self, name): |
| 144 | + name = string.lower(name) + ':' |
| 145 | + n = len(name) |
| 146 | + list = [] |
| 147 | + hit = 0 |
| 148 | + for line in self.headers: |
| 149 | + if string.lower(line[:n]) == name: |
| 150 | + hit = 1 |
| 151 | + elif line[:1] not in string.whitespace: |
| 152 | + if hit: |
| 153 | + break |
| 154 | + if hit: |
| 155 | + list.append(line) |
| 156 | + return list |
| 157 | + |
| 158 | + |
| 159 | + # A higher-level interface to getfirstmatchingheader(). |
| 160 | + # Return a string containing the literal text of the header |
| 161 | + # but with the keyword stripped. All leading, trailing and |
| 162 | + # embedded whitespace is kept in the string, however. |
| 163 | + # Return None if the header does not occur. |
| 164 | + |
| 165 | + def getrawheader(self, name): |
| 166 | + list = self.getfirstmatchingheader(name) |
| 167 | + if not list: |
| 168 | + return None |
| 169 | + list[0] = list[0][len(name) + 1:] |
| 170 | + return string.joinfields(list, '') |
| 171 | + |
| 172 | + |
| 173 | + # Going one step further: also strip leading and trailing |
| 174 | + # whitespace. |
| 175 | + |
| 176 | + def getheader(self, name): |
| 177 | + text = self.getrawheader(name) |
| 178 | + if text == None: |
| 179 | + return None |
| 180 | + return string.strip(text) |
| 181 | + |
| 182 | + |
| 183 | + # XXX The next step would be to define self.getaddr(name) |
| 184 | + # and self.getaddrlist(name) which would parse a header |
| 185 | + # consisting of a single mail address and a number of mail |
| 186 | + # addresses, respectively. Lower level functions would be |
| 187 | + # parseaddr(string) and parseaddrlist(string). |
| 188 | + |
| 189 | + # XXX Similar, there would be a function self.getdate(name) to |
| 190 | + # return a date in canonical form (perhaps a number compatible |
| 191 | + # to time.time()) and a function parsedate(string). |
| 192 | + |
| 193 | + # XXX The inverses of the parse functions may also be useful. |
| 194 | + |
| 195 | + |
| 196 | + |
| 197 | + |
| 198 | +# Utility functions |
| 199 | +# ----------------- |
| 200 | + |
| 201 | + |
| 202 | +# Remove quotes from a string. |
| 203 | +# XXX Should fix this to be really conformant. |
| 204 | + |
| 205 | +def unquote(str): |
| 206 | + if len(str) > 1: |
| 207 | + if str[0] == '"' and str[-1:] == '"': |
| 208 | + return str[1:-1] |
| 209 | + if str[0] == '<' and str[-1:] == '>': |
| 210 | + return str[1:-1] |
| 211 | + return str |
0 commit comments