base64.py revision 3d731c5994297d2786ccd1a6df1f8dc4cc46cde5
1#! /usr/bin/env python 2 3"""RFC 3548: Base16, Base32, Base64 Data Encodings""" 4 5# Modified 04-Oct-1995 by Jack Jansen to use binascii module 6# Modified 30-Dec-2003 by Barry Warsaw to add full RFC 3548 support 7 8import re 9import struct 10import string 11import binascii 12 13 14__all__ = [ 15 # Legacy interface exports traditional RFC 1521 Base64 encodings 16 'encode', 'decode', 'encodestring', 'decodestring', 17 # Generalized interface for other encodings 18 'b64encode', 'b64decode', 'b32encode', 'b32decode', 19 'b16encode', 'b16decode', 20 # Standard Base64 encoding 21 'standard_b64encode', 'standard_b64decode', 22 # Some common Base64 alternatives. As referenced by RFC 3458, see thread 23 # starting at: 24 # 25 # http://zgp.org/pipermail/p2p-hackers/2001-September/000316.html 26 'urlsafe_b64encode', 'urlsafe_b64decode', 27 ] 28 29_translation = [chr(_x) for _x in range(256)] 30EMPTYSTRING = '' 31 32 33def _translate(s, altchars): 34 translation = _translation[:] 35 for k, v in altchars.items(): 36 translation[ord(k)] = v 37 return s.translate(''.join(translation)) 38 39 40 41# Base64 encoding/decoding uses binascii 42 43def b64encode(s, altchars=None): 44 """Encode a string using Base64. 45 46 s is the string to encode. Optional altchars must be a string of at least 47 length 2 (additional characters are ignored) which specifies an 48 alternative alphabet for the '+' and '/' characters. This allows an 49 application to e.g. generate url or filesystem safe Base64 strings. 50 51 The encoded string is returned. 52 """ 53 # Strip off the trailing newline 54 encoded = binascii.b2a_base64(s)[:-1] 55 if altchars is not None: 56 return encoded.translate(string.maketrans(b'+/', altchars[:2])) 57 return encoded 58 59 60def b64decode(s, altchars=None): 61 """Decode a Base64 encoded string. 62 63 s is the string to decode. Optional altchars must be a string of at least 64 length 2 (additional characters are ignored) which specifies the 65 alternative alphabet used instead of the '+' and '/' characters. 66 67 The decoded string is returned. A TypeError is raised if s were 68 incorrectly padded or if there are non-alphabet characters present in the 69 string. 70 """ 71 if altchars is not None: 72 s = s.translate(string.maketrans(altchars[:2], '+/')) 73 try: 74 return binascii.a2b_base64(s) 75 except binascii.Error, msg: 76 # Transform this exception for consistency 77 raise TypeError(msg) 78 79 80def standard_b64encode(s): 81 """Encode a string using the standard Base64 alphabet. 82 83 s is the string to encode. The encoded string is returned. 84 """ 85 return b64encode(s) 86 87def standard_b64decode(s): 88 """Decode a string encoded with the standard Base64 alphabet. 89 90 s is the string to decode. The decoded string is returned. A TypeError 91 is raised if the string is incorrectly padded or if there are non-alphabet 92 characters present in the string. 93 """ 94 return b64decode(s) 95 96_urlsafe_encode_translation = string.maketrans(b'+/', b'-_') 97_urlsafe_decode_translation = string.maketrans(b'-_', b'+/') 98 99def urlsafe_b64encode(s): 100 """Encode a string using a url-safe Base64 alphabet. 101 102 s is the string to encode. The encoded string is returned. The alphabet 103 uses '-' instead of '+' and '_' instead of '/'. 104 """ 105 return b64encode(s).translate(_urlsafe_encode_translation) 106 107def urlsafe_b64decode(s): 108 """Decode a string encoded with the standard Base64 alphabet. 109 110 s is the string to decode. The decoded string is returned. A TypeError 111 is raised if the string is incorrectly padded or if there are non-alphabet 112 characters present in the string. 113 114 The alphabet uses '-' instead of '+' and '_' instead of '/'. 115 """ 116 return b64decode(s.translate(_urlsafe_decode_translation)) 117 118 119 120# Base32 encoding/decoding must be done in Python 121_b32alphabet = { 122 0: 'A', 9: 'J', 18: 'S', 27: '3', 123 1: 'B', 10: 'K', 19: 'T', 28: '4', 124 2: 'C', 11: 'L', 20: 'U', 29: '5', 125 3: 'D', 12: 'M', 21: 'V', 30: '6', 126 4: 'E', 13: 'N', 22: 'W', 31: '7', 127 5: 'F', 14: 'O', 23: 'X', 128 6: 'G', 15: 'P', 24: 'Y', 129 7: 'H', 16: 'Q', 25: 'Z', 130 8: 'I', 17: 'R', 26: '2', 131 } 132 133_b32tab = _b32alphabet.items() 134_b32tab.sort() 135_b32tab = [v for k, v in _b32tab] 136_b32rev = dict([(v, long(k)) for k, v in _b32alphabet.items()]) 137 138 139def b32encode(s): 140 """Encode a string using Base32. 141 142 s is the string to encode. The encoded string is returned. 143 """ 144 parts = [] 145 quanta, leftover = divmod(len(s), 5) 146 # Pad the last quantum with zero bits if necessary 147 if leftover: 148 s += ('\0' * (5 - leftover)) 149 quanta += 1 150 for i in range(quanta): 151 # c1 and c2 are 16 bits wide, c3 is 8 bits wide. The intent of this 152 # code is to process the 40 bits in units of 5 bits. So we take the 1 153 # leftover bit of c1 and tack it onto c2. Then we take the 2 leftover 154 # bits of c2 and tack them onto c3. The shifts and masks are intended 155 # to give us values of exactly 5 bits in width. 156 c1, c2, c3 = struct.unpack('!HHB', s[i*5:(i+1)*5]) 157 c2 += (c1 & 1) << 16 # 17 bits wide 158 c3 += (c2 & 3) << 8 # 10 bits wide 159 parts.extend([_b32tab[c1 >> 11], # bits 1 - 5 160 _b32tab[(c1 >> 6) & 0x1f], # bits 6 - 10 161 _b32tab[(c1 >> 1) & 0x1f], # bits 11 - 15 162 _b32tab[c2 >> 12], # bits 16 - 20 (1 - 5) 163 _b32tab[(c2 >> 7) & 0x1f], # bits 21 - 25 (6 - 10) 164 _b32tab[(c2 >> 2) & 0x1f], # bits 26 - 30 (11 - 15) 165 _b32tab[c3 >> 5], # bits 31 - 35 (1 - 5) 166 _b32tab[c3 & 0x1f], # bits 36 - 40 (1 - 5) 167 ]) 168 encoded = EMPTYSTRING.join(parts) 169 # Adjust for any leftover partial quanta 170 if leftover == 1: 171 return encoded[:-6] + '======' 172 elif leftover == 2: 173 return encoded[:-4] + '====' 174 elif leftover == 3: 175 return encoded[:-3] + '===' 176 elif leftover == 4: 177 return encoded[:-1] + '=' 178 return encoded 179 180 181def b32decode(s, casefold=False, map01=None): 182 """Decode a Base32 encoded string. 183 184 s is the string to decode. Optional casefold is a flag specifying whether 185 a lowercase alphabet is acceptable as input. For security purposes, the 186 default is False. 187 188 RFC 3548 allows for optional mapping of the digit 0 (zero) to the letter O 189 (oh), and for optional mapping of the digit 1 (one) to either the letter I 190 (eye) or letter L (el). The optional argument map01 when not None, 191 specifies which letter the digit 1 should be mapped to (when map01 is not 192 None, the digit 0 is always mapped to the letter O). For security 193 purposes the default is None, so that 0 and 1 are not allowed in the 194 input. 195 196 The decoded string is returned. A TypeError is raised if s were 197 incorrectly padded or if there are non-alphabet characters present in the 198 string. 199 """ 200 quanta, leftover = divmod(len(s), 8) 201 if leftover: 202 raise TypeError('Incorrect padding') 203 # Handle section 2.4 zero and one mapping. The flag map01 will be either 204 # False, or the character to map the digit 1 (one) to. It should be 205 # either L (el) or I (eye). 206 if map01: 207 s = s.translate(string.maketrans(b'01', b'O' + map01)) 208 if casefold: 209 s = s.upper() 210 # Strip off pad characters from the right. We need to count the pad 211 # characters because this will tell us how many null bytes to remove from 212 # the end of the decoded string. 213 padchars = 0 214 mo = re.search('(?P<pad>[=]*)$', s) 215 if mo: 216 padchars = len(mo.group('pad')) 217 if padchars > 0: 218 s = s[:-padchars] 219 # Now decode the full quanta 220 parts = [] 221 acc = 0 222 shift = 35 223 for c in s: 224 val = _b32rev.get(c) 225 if val is None: 226 raise TypeError('Non-base32 digit found') 227 acc += _b32rev[c] << shift 228 shift -= 5 229 if shift < 0: 230 parts.append(binascii.unhexlify('%010x' % acc)) 231 acc = 0 232 shift = 35 233 # Process the last, partial quanta 234 last = binascii.unhexlify('%010x' % acc) 235 if padchars == 0: 236 last = '' # No characters 237 elif padchars == 1: 238 last = last[:-1] 239 elif padchars == 3: 240 last = last[:-2] 241 elif padchars == 4: 242 last = last[:-3] 243 elif padchars == 6: 244 last = last[:-4] 245 else: 246 raise TypeError('Incorrect padding') 247 parts.append(last) 248 return EMPTYSTRING.join(parts) 249 250 251 252# RFC 3548, Base 16 Alphabet specifies uppercase, but hexlify() returns 253# lowercase. The RFC also recommends against accepting input case 254# insensitively. 255def b16encode(s): 256 """Encode a string using Base16. 257 258 s is the string to encode. The encoded string is returned. 259 """ 260 return binascii.hexlify(s).upper() 261 262 263def b16decode(s, casefold=False): 264 """Decode a Base16 encoded string. 265 266 s is the string to decode. Optional casefold is a flag specifying whether 267 a lowercase alphabet is acceptable as input. For security purposes, the 268 default is False. 269 270 The decoded string is returned. A TypeError is raised if s were 271 incorrectly padded or if there are non-alphabet characters present in the 272 string. 273 """ 274 if casefold: 275 s = s.upper() 276 if re.search('[^0-9A-F]', s): 277 raise TypeError('Non-base16 digit found') 278 return binascii.unhexlify(s) 279 280 281 282# Legacy interface. This code could be cleaned up since I don't believe 283# binascii has any line length limitations. It just doesn't seem worth it 284# though. 285 286MAXLINESIZE = 76 # Excluding the CRLF 287MAXBINSIZE = (MAXLINESIZE//4)*3 288 289def encode(input, output): 290 """Encode a file.""" 291 while True: 292 s = input.read(MAXBINSIZE) 293 if not s: 294 break 295 while len(s) < MAXBINSIZE: 296 ns = input.read(MAXBINSIZE-len(s)) 297 if not ns: 298 break 299 s += ns 300 line = binascii.b2a_base64(s) 301 output.write(line) 302 303 304def decode(input, output): 305 """Decode a file.""" 306 while True: 307 line = input.readline() 308 if not line: 309 break 310 s = binascii.a2b_base64(line) 311 output.write(s) 312 313 314def encodestring(s): 315 """Encode a string into multiple lines of base-64 data.""" 316 pieces = [] 317 for i in range(0, len(s), MAXBINSIZE): 318 chunk = s[i : i + MAXBINSIZE] 319 pieces.append(binascii.b2a_base64(chunk)) 320 return "".join(pieces) 321 322 323def decodestring(s): 324 """Decode a string.""" 325 return binascii.a2b_base64(s) 326 327 328 329# Useable as a script... 330def test(): 331 """Small test program""" 332 import sys, getopt 333 try: 334 opts, args = getopt.getopt(sys.argv[1:], 'deut') 335 except getopt.error, msg: 336 sys.stdout = sys.stderr 337 print msg 338 print """usage: %s [-d|-e|-u|-t] [file|-] 339 -d, -u: decode 340 -e: encode (default) 341 -t: encode and decode string 'Aladdin:open sesame'"""%sys.argv[0] 342 sys.exit(2) 343 func = encode 344 for o, a in opts: 345 if o == '-e': func = encode 346 if o == '-d': func = decode 347 if o == '-u': func = decode 348 if o == '-t': test1(); return 349 if args and args[0] != '-': 350 with open(args[0], 'rb') as f: 351 func(f, sys.stdout) 352 else: 353 func(sys.stdin, sys.stdout) 354 355 356def test1(): 357 s0 = "Aladdin:open sesame" 358 s1 = encodestring(s0) 359 s2 = decodestring(s1) 360 print s0, repr(s1), s2 361 362 363if __name__ == '__main__': 364 test() 365