1 """Guess the MIME type of a file. 2 3 This module defines two useful functions: 4 5 guess_type(url, strict=1) -- guess the MIME type and encoding of a URL. 6 7 guess_extension(type, strict=1) -- guess the extension for a given MIME type. 8 9 It also contains the following, for tuning the behavior: 10 11 Data: 12 13 knownfiles -- list of files to parse 14 inited -- flag set when init() has been called 15 suffix_map -- dictionary mapping suffixes to suffixes 16 encodings_map -- dictionary mapping suffixes to encodings 17 types_map -- dictionary mapping suffixes to types 18 19 Functions: 20 21 init([files]) -- parse a list of files, default knownfiles (on Windows, the 22 default values are taken from the registry) 23 read_mime_types(file) -- parse one file, return a dictionary or None 24 """ 25 26 import os 27 import sys 28 import posixpath 29 import urllib 30 try: 31 import _winreg 32 except ImportError: 33 _winreg = None 34 35 __all__ = [ 36 "guess_type","guess_extension","guess_all_extensions", 37 "add_type","read_mime_types","init" 38 ] 39 40 knownfiles = [ 41 "/etc/mime.types", 42 "/etc/httpd/mime.types", # Mac OS X 43 "/etc/httpd/conf/mime.types", # Apache 44 "/etc/apache/mime.types", # Apache 1 45 "/etc/apache2/mime.types", # Apache 2 46 "/usr/local/etc/httpd/conf/mime.types", 47 "/usr/local/lib/netscape/mime.types", 48 "/usr/local/etc/httpd/conf/mime.types", # Apache 1.2 49 "/usr/local/etc/mime.types", # Apache 1.3 50 ] 51 52 inited = False 53 _db = None 54 55 56 class MimeTypes: 57 """MIME-types datastore. 58 59 This datastore can handle information from mime.types-style files 60 and supports basic determination of MIME type from a filename or 61 URL, and can guess a reasonable extension given a MIME type. 62 """ 63 64 def __init__(self, filenames=(), strict=True): 65 if not inited: 66 init() 67 self.encodings_map = encodings_map.copy() 68 self.suffix_map = suffix_map.copy() 69 self.types_map = ({}, {}) # dict for (non-strict, strict) 70 self.types_map_inv = ({}, {}) 71 for (ext, type) in types_map.items(): 72 self.add_type(type, ext, True) 73 for (ext, type) in common_types.items(): 74 self.add_type(type, ext, False) 75 for name in filenames: 76 self.read(name, strict) 77 78 def add_type(self, type, ext, strict=True): 79 """Add a mapping between a type and an extension. 80 81 When the extension is already known, the new 82 type will replace the old one. When the type 83 is already known the extension will be added 84 to the list of known extensions. 85 86 If strict is true, information will be added to 87 list of standard types, else to the list of non-standard 88 types. 89 """ 90 self.types_map[strict][ext] = type 91 exts = self.types_map_inv[strict].setdefault(type, []) 92 if ext not in exts: 93 exts.append(ext) 94 95 def guess_type(self, url, strict=True): 96 """Guess the type of a file based on its URL. 97 98 Return value is a tuple (type, encoding) where type is None if 99 the type can't be guessed (no or unknown suffix) or a string 100 of the form type/subtype, usable for a MIME Content-type 101 header; and encoding is None for no encoding or the name of 102 the program used to encode (e.g. compress or gzip). The 103 mappings are table driven. Encoding suffixes are case 104 sensitive; type suffixes are first tried case sensitive, then 105 case insensitive. 106 107 The suffixes .tgz, .taz and .tz (case sensitive!) are all 108 mapped to '.tar.gz'. (This is table-driven too, using the 109 dictionary suffix_map.) 110 111 Optional `strict' argument when False adds a bunch of commonly found, 112 but non-standard types. 113 """ 114 scheme, url = urllib.splittype(url) 115 if scheme == 'data': 116 # syntax of data URLs: 117 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data 118 # mediatype := [ type "/" subtype ] *( ";" parameter ) 119 # data := *urlchar 120 # parameter := attribute "=" value 121 # type/subtype defaults to "text/plain" 122 comma = url.find(',') 123 if comma < 0: 124 # bad data URL 125 return None, None 126 semi = url.find(';', 0, comma) 127 if semi >= 0: 128 type = url[:semi] 129 else: 130 type = url[:comma] 131 if '=' in type or '/' not in type: 132 type = 'text/plain' 133 return type, None # never compressed, so encoding is None 134 base, ext = posixpath.splitext(url) 135 while ext in self.suffix_map: 136 base, ext = posixpath.splitext(base + self.suffix_map[ext]) 137 if ext in self.encodings_map: 138 encoding = self.encodings_map[ext] 139 base, ext = posixpath.splitext(base) 140 else: 141 encoding = None 142 types_map = self.types_map[True] 143 if ext in types_map: 144 return types_map[ext], encoding 145 elif ext.lower() in types_map: 146 return types_map[ext.lower()], encoding 147 elif strict: 148 return None, encoding 149 types_map = self.types_map[False] 150 if ext in types_map: 151 return types_map[ext], encoding 152 elif ext.lower() in types_map: 153 return types_map[ext.lower()], encoding 154 else: 155 return None, encoding 156 157 def guess_all_extensions(self, type, strict=True): 158 """Guess the extensions for a file based on its MIME type. 159 160 Return value is a list of strings giving the possible filename 161 extensions, including the leading dot ('.'). The extension is not 162 guaranteed to have been associated with any particular data stream, 163 but would be mapped to the MIME type `type' by guess_type(). 164 165 Optional `strict' argument when false adds a bunch of commonly found, 166 but non-standard types. 167 """ 168 type = type.lower() 169 extensions = self.types_map_inv[True].get(type, []) 170 if not strict: 171 for ext in self.types_map_inv[False].get(type, []): 172 if ext not in extensions: 173 extensions.append(ext) 174 return extensions 175 176 def guess_extension(self, type, strict=True): 177 """Guess the extension for a file based on its MIME type. 178 179 Return value is a string giving a filename extension, 180 including the leading dot ('.'). The extension is not 181 guaranteed to have been associated with any particular data 182 stream, but would be mapped to the MIME type `type' by 183 guess_type(). If no extension can be guessed for `type', None 184 is returned. 185 186 Optional `strict' argument when false adds a bunch of commonly found, 187 but non-standard types. 188 """ 189 extensions = self.guess_all_extensions(type, strict) 190 if not extensions: 191 return None 192 return extensions[0] 193 194 def read(self, filename, strict=True): 195 """ 196 Read a single mime.types-format file, specified by pathname. 197 198 If strict is true, information will be added to 199 list of standard types, else to the list of non-standard 200 types. 201 """ 202 with open(filename) as fp: 203 self.readfp(fp, strict) 204 205 def readfp(self, fp, strict=True): 206 """ 207 Read a single mime.types-format file. 208 209 If strict is true, information will be added to 210 list of standard types, else to the list of non-standard 211 types. 212 """ 213 while 1: 214 line = fp.readline() 215 if not line: 216 break 217 words = line.split() 218 for i in range(len(words)): 219 if words[i][0] == '#': 220 del words[i:] 221 break 222 if not words: 223 continue 224 type, suffixes = words[0], words[1:] 225 for suff in suffixes: 226 self.add_type(type, '.' + suff, strict) 227 228 def read_windows_registry(self, strict=True): 229 """ 230 Load the MIME types database from Windows registry. 231 232 If strict is true, information will be added to 233 list of standard types, else to the list of non-standard 234 types. 235 """ 236 237 # Windows only 238 if not _winreg: 239 return 240 241 def enum_types(mimedb): 242 i = 0 243 while True: 244 try: 245 ctype = _winreg.EnumKey(mimedb, i) 246 except EnvironmentError: 247 break 248 try: 249 ctype = ctype.encode(default_encoding) # omit in 3.x! 250 except UnicodeEncodeError: 251 pass 252 else: 253 yield ctype 254 i += 1 255 256 default_encoding = sys.getdefaultencoding() 257 with _winreg.OpenKey(_winreg.HKEY_CLASSES_ROOT, 258 r'MIME\Database\Content Type') as mimedb: 259 for ctype in enum_types(mimedb): 260 try: 261 with _winreg.OpenKey(mimedb, ctype) as key: 262 suffix, datatype = _winreg.QueryValueEx(key, 263 'Extension') 264 except EnvironmentError: 265 continue 266 if datatype != _winreg.REG_SZ: 267 continue 268 try: 269 suffix = suffix.encode(default_encoding) # omit in 3.x! 270 except UnicodeEncodeError: 271 continue 272 self.add_type(ctype, suffix, strict) 273 274 275 def guess_type(url, strict=True): 276 """Guess the type of a file based on its URL. 277 278 Return value is a tuple (type, encoding) where type is None if the 279 type can't be guessed (no or unknown suffix) or a string of the 280 form type/subtype, usable for a MIME Content-type header; and 281 encoding is None for no encoding or the name of the program used 282 to encode (e.g. compress or gzip). The mappings are table 283 driven. Encoding suffixes are case sensitive; type suffixes are 284 first tried case sensitive, then case insensitive. 285 286 The suffixes .tgz, .taz and .tz (case sensitive!) are all mapped 287 to ".tar.gz". (This is table-driven too, using the dictionary 288 suffix_map). 289 290 Optional `strict' argument when false adds a bunch of commonly found, but 291 non-standard types. 292 """ 293 if _db is None: 294 init() 295 return _db.guess_type(url, strict) 296 297 298 def guess_all_extensions(type, strict=True): 299 """Guess the extensions for a file based on its MIME type. 300 301 Return value is a list of strings giving the possible filename 302 extensions, including the leading dot ('.'). The extension is not 303 guaranteed to have been associated with any particular data 304 stream, but would be mapped to the MIME type `type' by 305 guess_type(). If no extension can be guessed for `type', None 306 is returned. 307 308 Optional `strict' argument when false adds a bunch of commonly found, 309 but non-standard types. 310 """ 311 if _db is None: 312 init() 313 return _db.guess_all_extensions(type, strict) 314 315 def guess_extension(type, strict=True): 316 """Guess the extension for a file based on its MIME type. 317 318 Return value is a string giving a filename extension, including the 319 leading dot ('.'). The extension is not guaranteed to have been 320 associated with any particular data stream, but would be mapped to the 321 MIME type `type' by guess_type(). If no extension can be guessed for 322 `type', None is returned. 323 324 Optional `strict' argument when false adds a bunch of commonly found, 325 but non-standard types. 326 """ 327 if _db is None: 328 init() 329 return _db.guess_extension(type, strict) 330 331 def add_type(type, ext, strict=True): 332 """Add a mapping between a type and an extension. 333 334 When the extension is already known, the new 335 type will replace the old one. When the type 336 is already known the extension will be added 337 to the list of known extensions. 338 339 If strict is true, information will be added to 340 list of standard types, else to the list of non-standard 341 types. 342 """ 343 if _db is None: 344 init() 345 return _db.add_type(type, ext, strict) 346 347 348 def init(files=None): 349 global suffix_map, types_map, encodings_map, common_types 350 global inited, _db 351 inited = True # so that MimeTypes.__init__() doesn't call us again 352 db = MimeTypes() 353 if files is None: 354 if _winreg: 355 db.read_windows_registry() 356 files = knownfiles 357 for file in files: 358 if os.path.isfile(file): 359 db.read(file) 360 encodings_map = db.encodings_map 361 suffix_map = db.suffix_map 362 types_map = db.types_map[True] 363 common_types = db.types_map[False] 364 # Make the DB a global variable now that it is fully initialized 365 _db = db 366 367 368 def read_mime_types(file): 369 try: 370 f = open(file) 371 except IOError: 372 return None 373 db = MimeTypes() 374 db.readfp(f, True) 375 return db.types_map[True] 376 377 378 def _default_mime_types(): 379 global suffix_map 380 global encodings_map 381 global types_map 382 global common_types 383 384 suffix_map = { 385 '.tgz': '.tar.gz', 386 '.taz': '.tar.gz', 387 '.tz': '.tar.gz', 388 '.tbz2': '.tar.bz2', 389 '.txz': '.tar.xz', 390 } 391 392 encodings_map = { 393 '.gz': 'gzip', 394 '.Z': 'compress', 395 '.bz2': 'bzip2', 396 '.xz': 'xz', 397 } 398 399 # Before adding new types, make sure they are either registered with IANA, 400 # at http://www.isi.edu/in-notes/iana/assignments/media-types 401 # or extensions, i.e. using the x- prefix 402 403 # If you add to these, please keep them sorted! 404 types_map = { 405 '.a' : 'application/octet-stream', 406 '.ai' : 'application/postscript', 407 '.aif' : 'audio/x-aiff', 408 '.aifc' : 'audio/x-aiff', 409 '.aiff' : 'audio/x-aiff', 410 '.au' : 'audio/basic', 411 '.avi' : 'video/x-msvideo', 412 '.bat' : 'text/plain', 413 '.bcpio' : 'application/x-bcpio', 414 '.bin' : 'application/octet-stream', 415 '.bmp' : 'image/x-ms-bmp', 416 '.c' : 'text/plain', 417 # Duplicates :( 418 '.cdf' : 'application/x-cdf', 419 '.cdf' : 'application/x-netcdf', 420 '.cpio' : 'application/x-cpio', 421 '.csh' : 'application/x-csh', 422 '.css' : 'text/css', 423 '.dll' : 'application/octet-stream', 424 '.doc' : 'application/msword', 425 '.dot' : 'application/msword', 426 '.dvi' : 'application/x-dvi', 427 '.eml' : 'message/rfc822', 428 '.eps' : 'application/postscript', 429 '.etx' : 'text/x-setext', 430 '.exe' : 'application/octet-stream', 431 '.gif' : 'image/gif', 432 '.gtar' : 'application/x-gtar', 433 '.h' : 'text/plain', 434 '.hdf' : 'application/x-hdf', 435 '.htm' : 'text/html', 436 '.html' : 'text/html', 437 '.ico' : 'image/vnd.microsoft.icon', 438 '.ief' : 'image/ief', 439 '.jpe' : 'image/jpeg', 440 '.jpeg' : 'image/jpeg', 441 '.jpg' : 'image/jpeg', 442 '.js' : 'application/javascript', 443 '.ksh' : 'text/plain', 444 '.latex' : 'application/x-latex', 445 '.m1v' : 'video/mpeg', 446 '.man' : 'application/x-troff-man', 447 '.me' : 'application/x-troff-me', 448 '.mht' : 'message/rfc822', 449 '.mhtml' : 'message/rfc822', 450 '.mif' : 'application/x-mif', 451 '.mov' : 'video/quicktime', 452 '.movie' : 'video/x-sgi-movie', 453 '.mp2' : 'audio/mpeg', 454 '.mp3' : 'audio/mpeg', 455 '.mp4' : 'video/mp4', 456 '.mpa' : 'video/mpeg', 457 '.mpe' : 'video/mpeg', 458 '.mpeg' : 'video/mpeg', 459 '.mpg' : 'video/mpeg', 460 '.ms' : 'application/x-troff-ms', 461 '.nc' : 'application/x-netcdf', 462 '.nws' : 'message/rfc822', 463 '.o' : 'application/octet-stream', 464 '.obj' : 'application/octet-stream', 465 '.oda' : 'application/oda', 466 '.p12' : 'application/x-pkcs12', 467 '.p7c' : 'application/pkcs7-mime', 468 '.pbm' : 'image/x-portable-bitmap', 469 '.pdf' : 'application/pdf', 470 '.pfx' : 'application/x-pkcs12', 471 '.pgm' : 'image/x-portable-graymap', 472 '.pl' : 'text/plain', 473 '.png' : 'image/png', 474 '.pnm' : 'image/x-portable-anymap', 475 '.pot' : 'application/vnd.ms-powerpoint', 476 '.ppa' : 'application/vnd.ms-powerpoint', 477 '.ppm' : 'image/x-portable-pixmap', 478 '.pps' : 'application/vnd.ms-powerpoint', 479 '.ppt' : 'application/vnd.ms-powerpoint', 480 '.ps' : 'application/postscript', 481 '.pwz' : 'application/vnd.ms-powerpoint', 482 '.py' : 'text/x-python', 483 '.pyc' : 'application/x-python-code', 484 '.pyo' : 'application/x-python-code', 485 '.qt' : 'video/quicktime', 486 '.ra' : 'audio/x-pn-realaudio', 487 '.ram' : 'application/x-pn-realaudio', 488 '.ras' : 'image/x-cmu-raster', 489 '.rdf' : 'application/xml', 490 '.rgb' : 'image/x-rgb', 491 '.roff' : 'application/x-troff', 492 '.rtx' : 'text/richtext', 493 '.sgm' : 'text/x-sgml', 494 '.sgml' : 'text/x-sgml', 495 '.sh' : 'application/x-sh', 496 '.shar' : 'application/x-shar', 497 '.snd' : 'audio/basic', 498 '.so' : 'application/octet-stream', 499 '.src' : 'application/x-wais-source', 500 '.sv4cpio': 'application/x-sv4cpio', 501 '.sv4crc' : 'application/x-sv4crc', 502 '.swf' : 'application/x-shockwave-flash', 503 '.t' : 'application/x-troff', 504 '.tar' : 'application/x-tar', 505 '.tcl' : 'application/x-tcl', 506 '.tex' : 'application/x-tex', 507 '.texi' : 'application/x-texinfo', 508 '.texinfo': 'application/x-texinfo', 509 '.tif' : 'image/tiff', 510 '.tiff' : 'image/tiff', 511 '.tr' : 'application/x-troff', 512 '.tsv' : 'text/tab-separated-values', 513 '.txt' : 'text/plain', 514 '.ustar' : 'application/x-ustar', 515 '.vcf' : 'text/x-vcard', 516 '.wav' : 'audio/x-wav', 517 '.wiz' : 'application/msword', 518 '.wsdl' : 'application/xml', 519 '.xbm' : 'image/x-xbitmap', 520 '.xlb' : 'application/vnd.ms-excel', 521 # Duplicates :( 522 '.xls' : 'application/excel', 523 '.xls' : 'application/vnd.ms-excel', 524 '.xml' : 'text/xml', 525 '.xpdl' : 'application/xml', 526 '.xpm' : 'image/x-xpixmap', 527 '.xsl' : 'application/xml', 528 '.xwd' : 'image/x-xwindowdump', 529 '.zip' : 'application/zip', 530 } 531 532 # These are non-standard types, commonly found in the wild. They will 533 # only match if strict=0 flag is given to the API methods. 534 535 # Please sort these too 536 common_types = { 537 '.jpg' : 'image/jpg', 538 '.mid' : 'audio/midi', 539 '.midi': 'audio/midi', 540 '.pct' : 'image/pict', 541 '.pic' : 'image/pict', 542 '.pict': 'image/pict', 543 '.rtf' : 'application/rtf', 544 '.xul' : 'text/xul' 545 } 546 547 548 _default_mime_types() 549 550 551 if __name__ == '__main__': 552 import getopt 553 554 USAGE = """\ 555 Usage: mimetypes.py [options] type 556 557 Options: 558 --help / -h -- print this message and exit 559 --lenient / -l -- additionally search of some common, but non-standard 560 types. 561 --extension / -e -- guess extension instead of type 562 563 More than one type argument may be given. 564 """ 565 566 def usage(code, msg=''): 567 print USAGE 568 if msg: print msg 569 sys.exit(code) 570 571 try: 572 opts, args = getopt.getopt(sys.argv[1:], 'hle', 573 ['help', 'lenient', 'extension']) 574 except getopt.error, msg: 575 usage(1, msg) 576 577 strict = 1 578 extension = 0 579 for opt, arg in opts: 580 if opt in ('-h', '--help'): 581 usage(0) 582 elif opt in ('-l', '--lenient'): 583 strict = 0 584 elif opt in ('-e', '--extension'): 585 extension = 1 586 for gtype in args: 587 if extension: 588 guess = guess_extension(gtype, strict) 589 if not guess: print "I don't know anything about type", gtype 590 else: print guess 591 else: 592 guess, encoding = guess_type(gtype, strict) 593 if not guess: print "I don't know anything about type", gtype 594 else: print 'type:', guess, 'encoding:', encoding 595
