source file: /disk/u/t/dev/python-svn/Lib/gzip.py
file stats: 325 lines, 246 executed: 75.7% covered
   1. """Functions that read and write gzipped files.
   2. 
   3. The user of the file doesn't have to worry about the compression,
   4. but random access is not allowed."""
   5. 
   6. # based on Andrew Kuchling's minigzip.py distributed with the zlib module
   7. 
   8. import struct, sys, time
   9. import zlib
  10. import __builtin__
  11. 
  12. __all__ = ["GzipFile","open"]
  13. 
  14. FTEXT, FHCRC, FEXTRA, FNAME, FCOMMENT = 1, 2, 4, 8, 16
  15. 
  16. READ, WRITE = 1, 2
  17. 
  18. def U32(i):
  19.     """Return i as an unsigned integer, assuming it fits in 32 bits.
  20. 
  21.     If it's >= 2GB when viewed as a 32-bit unsigned int, return a long.
  22.     """
  23.     if i < 0:
  24.         i += 1L << 32
  25.     return i
  26. 
  27. def LOWU32(i):
  28.     """Return the low-order 32 bits of an int, as a non-negative int."""
  29.     return i & 0xFFFFFFFFL
  30. 
  31. def write32(output, value):
  32.     output.write(struct.pack("<l", value))
  33. 
  34. def write32u(output, value):
  35.     # The L format writes the bit pattern correctly whether signed
  36.     # or unsigned.
  37.     output.write(struct.pack("<L", value))
  38. 
  39. def read32(input):
  40.     return struct.unpack("<l", input.read(4))[0]
  41. 
  42. def open(filename, mode="rb", compresslevel=9):
  43.     """Shorthand for GzipFile(filename, mode, compresslevel).
  44. 
  45.     The filename argument is required; mode defaults to 'rb'
  46.     and compresslevel defaults to 9.
  47. 
  48.     """
  49.     return GzipFile(filename, mode, compresslevel)
  50. 
  51. class GzipFile:
  52.     """The GzipFile class simulates most of the methods of a file object with
  53.     the exception of the readinto() and truncate() methods.
  54. 
  55.     """
  56. 
  57.     myfileobj = None
  58.     max_read_chunk = 10 * 1024 * 1024   # 10Mb
  59. 
  60.     def __init__(self, filename=None, mode=None,
  61.                  compresslevel=9, fileobj=None):
  62.         """Constructor for the GzipFile class.
  63. 
  64.         At least one of fileobj and filename must be given a
  65.         non-trivial value.
  66. 
  67.         The new class instance is based on fileobj, which can be a regular
  68.         file, a StringIO object, or any other object which simulates a file.
  69.         It defaults to None, in which case filename is opened to provide
  70.         a file object.
  71. 
  72.         When fileobj is not None, the filename argument is only used to be
  73.         included in the gzip file header, which may includes the original
  74.         filename of the uncompressed file.  It defaults to the filename of
  75.         fileobj, if discernible; otherwise, it defaults to the empty string,
  76.         and in this case the original filename is not included in the header.
  77. 
  78.         The mode argument can be any of 'r', 'rb', 'a', 'ab', 'w', or 'wb',
  79.         depending on whether the file will be read or written.  The default
  80.         is the mode of fileobj if discernible; otherwise, the default is 'rb'.
  81.         Be aware that only the 'rb', 'ab', and 'wb' values should be used
  82.         for cross-platform portability.
  83. 
  84.         The compresslevel argument is an integer from 1 to 9 controlling the
  85.         level of compression; 1 is fastest and produces the least compression,
  86.         and 9 is slowest and produces the most compression.  The default is 9.
  87. 
  88.         """
  89. 
  90.         # guarantee the file is opened in binary mode on platforms
  91.         # that care about that sort of thing
  92.         if mode and 'b' not in mode:
  93.             mode += 'b'
  94.         if fileobj is None:
  95.             fileobj = self.myfileobj = __builtin__.open(filename, mode or 'rb')
  96.         if filename is None:
  97.             if hasattr(fileobj, 'name'): filename = fileobj.name
  98.             else: filename = ''
  99.         if mode is None:
 100.             if hasattr(fileobj, 'mode'): mode = fileobj.mode
 101.             else: mode = 'rb'
 102. 
 103.         if mode[0:1] == 'r':
 104.             self.mode = READ
 105.             # Set flag indicating start of a new member
 106.             self._new_member = True
 107.             self.extrabuf = ""
 108.             self.extrasize = 0
 109.             self.name = filename
 110.             # Starts small, scales exponentially
 111.             self.min_readsize = 100
 112. 
 113.         elif mode[0:1] == 'w' or mode[0:1] == 'a':
 114.             self.mode = WRITE
 115.             self._init_write(filename)
 116.             self.compress = zlib.compressobj(compresslevel,
 117.                                              zlib.DEFLATED,
 118.                                              -zlib.MAX_WBITS,
 119.                                              zlib.DEF_MEM_LEVEL,
 120.                                              0)
 121.         else:
 122.             raise IOError, "Mode " + mode + " not supported"
 123. 
 124.         self.fileobj = fileobj
 125.         self.offset = 0
 126. 
 127.         if self.mode == WRITE:
 128.             self._write_gzip_header()
 129. 
 130.     @property
 131.     def filename(self):
 132.         import warnings
 133.         warnings.warn("use the name attribute", DeprecationWarning)
 134.         if self.mode == WRITE and self.name[-3:] != ".gz":
 135.             return self.name + ".gz"
 136.         return self.name
 137. 
 138.     def __repr__(self):
 139.         s = repr(self.fileobj)
 140.         return '<gzip ' + s[1:-1] + ' ' + hex(id(self)) + '>'
 141. 
 142.     def _init_write(self, filename):
 143.         self.name = filename
 144.         self.crc = zlib.crc32("")
 145.         self.size = 0
 146.         self.writebuf = []
 147.         self.bufsize = 0
 148. 
 149.     def _write_gzip_header(self):
 150.         self.fileobj.write('\037\213')             # magic header
 151.         self.fileobj.write('\010')                 # compression method
 152.         fname = self.name
 153.         if fname.endswith(".gz"):
 154.             fname = fname[:-3]
 155.         flags = 0
 156.         if fname:
 157.             flags = FNAME
 158.         self.fileobj.write(chr(flags))
 159.         write32u(self.fileobj, long(time.time()))
 160.         self.fileobj.write('\002')
 161.         self.fileobj.write('\377')
 162.         if fname:
 163.             self.fileobj.write(fname + '\000')
 164. 
 165.     def _init_read(self):
 166.         self.crc = zlib.crc32("")
 167.         self.size = 0
 168. 
 169.     def _read_gzip_header(self):
 170.         magic = self.fileobj.read(2)
 171.         if magic != '\037\213':
 172.             raise IOError, 'Not a gzipped file'
 173.         method = ord( self.fileobj.read(1) )
 174.         if method != 8:
 175.             raise IOError, 'Unknown compression method'
 176.         flag = ord( self.fileobj.read(1) )
 177.         # modtime = self.fileobj.read(4)
 178.         # extraflag = self.fileobj.read(1)
 179.         # os = self.fileobj.read(1)
 180.         self.fileobj.read(6)
 181. 
 182.         if flag & FEXTRA:
 183.             # Read & discard the extra field, if present
 184.             xlen = ord(self.fileobj.read(1))
 185.             xlen = xlen + 256*ord(self.fileobj.read(1))
 186.             self.fileobj.read(xlen)
 187.         if flag & FNAME:
 188.             # Read and discard a null-terminated string containing the filename
 189.             while True:
 190.                 s = self.fileobj.read(1)
 191.                 if not s or s=='\000':
 192.                     break
 193.         if flag & FCOMMENT:
 194.             # Read and discard a null-terminated string containing a comment
 195.             while True:
 196.                 s = self.fileobj.read(1)
 197.                 if not s or s=='\000':
 198.                     break
 199.         if flag & FHCRC:
 200.             self.fileobj.read(2)     # Read & discard the 16-bit header CRC
 201. 
 202. 
 203.     def write(self,data):
 204.         if self.mode != WRITE:
 205.             import errno
 206.             raise IOError(errno.EBADF, "write() on read-only GzipFile object")
 207. 
 208.         if self.fileobj is None:
 209.             raise ValueError, "write() on closed GzipFile object"
 210.         if len(data) > 0:
 211.             self.size = self.size + len(data)
 212.             self.crc = zlib.crc32(data, self.crc)
 213.             self.fileobj.write( self.compress.compress(data) )
 214.             self.offset += len(data)
 215. 
 216.     def read(self, size=-1):
 217.         if self.mode != READ:
 218.             import errno
 219.             raise IOError(errno.EBADF, "read() on write-only GzipFile object")
 220. 
 221.         if self.extrasize <= 0 and self.fileobj is None:
 222.             return ''
 223. 
 224.         readsize = 1024
 225.         if size < 0:        # get the whole thing
 226.             try:
 227.                 while True:
 228.                     self._read(readsize)
 229.                     readsize = min(self.max_read_chunk, readsize * 2)
 230.             except EOFError:
 231.                 size = self.extrasize
 232.         else:               # just get some more of it
 233.             try:
 234.                 while size > self.extrasize:
 235.                     self._read(readsize)
 236.                     readsize = min(self.max_read_chunk, readsize * 2)
 237.             except EOFError:
 238.                 if size > self.extrasize:
 239.                     size = self.extrasize
 240. 
 241.         chunk = self.extrabuf[:size]
 242.         self.extrabuf = self.extrabuf[size:]
 243.         self.extrasize = self.extrasize - size
 244. 
 245.         self.offset += size
 246.         return chunk
 247. 
 248.     def _unread(self, buf):
 249.         self.extrabuf = buf + self.extrabuf
 250.         self.extrasize = len(buf) + self.extrasize
 251.         self.offset -= len(buf)
 252. 
 253.     def _read(self, size=1024):
 254.         if self.fileobj is None:
 255.             raise EOFError, "Reached EOF"
 256. 
 257.         if self._new_member:
 258.             # If the _new_member flag is set, we have to
 259.             # jump to the next member, if there is one.
 260.             #
 261.             # First, check if we're at the end of the file;
 262.             # if so, it's time to stop; no more members to read.
 263.             pos = self.fileobj.tell()   # Save current position
 264.             self.fileobj.seek(0, 2)     # Seek to end of file
 265.             if pos == self.fileobj.tell():
 266.                 raise EOFError, "Reached EOF"
 267.             else:
 268.                 self.fileobj.seek( pos ) # Return to original position
 269. 
 270.             self._init_read()
 271.             self._read_gzip_header()
 272.             self.decompress = zlib.decompressobj(-zlib.MAX_WBITS)
 273.             self._new_member = False
 274. 
 275.         # Read a chunk of data from the file
 276.         buf = self.fileobj.read(size)
 277. 
 278.         # If the EOF has been reached, flush the decompression object
 279.         # and mark this object as finished.
 280. 
 281.         if buf == "":
 282.             uncompress = self.decompress.flush()
 283.             self._read_eof()
 284.             self._add_read_data( uncompress )
 285.             raise EOFError, 'Reached EOF'
 286. 
 287.         uncompress = self.decompress.decompress(buf)
 288.         self._add_read_data( uncompress )
 289. 
 290.         if self.decompress.unused_data != "":
 291.             # Ending case: we've come to the end of a member in the file,
 292.             # so seek back to the start of the unused data, finish up
 293.             # this member, and read a new gzip header.
 294.             # (The number of bytes to seek back is the length of the unused
 295.             # data, minus 8 because _read_eof() will rewind a further 8 bytes)
 296.             self.fileobj.seek( -len(self.decompress.unused_data)+8, 1)
 297. 
 298.             # Check the CRC and file size, and set the flag so we read
 299.             # a new member on the next call
 300.             self._read_eof()
 301.             self._new_member = True
 302. 
 303.     def _add_read_data(self, data):
 304.         self.crc = zlib.crc32(data, self.crc)
 305.         self.extrabuf = self.extrabuf + data
 306.         self.extrasize = self.extrasize + len(data)
 307.         self.size = self.size + len(data)
 308. 
 309.     def _read_eof(self):
 310.         # We've read to the end of the file, so we have to rewind in order
 311.         # to reread the 8 bytes containing the CRC and the file size.
 312.         # We check the that the computed CRC and size of the
 313.         # uncompressed data matches the stored values.  Note that the size
 314.         # stored is the true file size mod 2**32.
 315.         self.fileobj.seek(-8, 1)
 316.         crc32 = read32(self.fileobj)
 317.         isize = U32(read32(self.fileobj))   # may exceed 2GB
 318.         if U32(crc32) != U32(self.crc):
 319.             raise IOError, "CRC check failed"
 320.         elif isize != LOWU32(self.size):
 321.             raise IOError, "Incorrect length of data produced"
 322. 
 323.     def close(self):
 324.         if self.mode == WRITE:
 325.             self.fileobj.write(self.compress.flush())
 326.             # The native zlib crc is an unsigned 32-bit integer, but
 327.             # the Python wrapper implicitly casts that to a signed C
 328.             # long.  So, on a 32-bit box self.crc may "look negative",
 329.             # while the same crc on a 64-bit box may "look positive".
 330.             # To avoid irksome warnings from the `struct` module, force
 331.             # it to look positive on all boxes.
 332.             write32u(self.fileobj, LOWU32(self.crc))
 333.             # self.size may exceed 2GB, or even 4GB
 334.             write32u(self.fileobj, LOWU32(self.size))
 335.             self.fileobj = None
 336.         elif self.mode == READ:
 337.             self.fileobj = None
 338.         if self.myfileobj:
 339.             self.myfileobj.close()
 340.             self.myfileobj = None
 341. 
 342.     def __del__(self):
 343.         try:
 344.             if (self.myfileobj is None and
 345.                 self.fileobj is None):
 346.                 return
 347.         except AttributeError:
 348.             return
 349.         self.close()
 350. 
 351.     def flush(self,zlib_mode=zlib.Z_SYNC_FLUSH):
 352.         if self.mode == WRITE:
 353.             # Ensure the compressor's buffer is flushed
 354.             self.fileobj.write(self.compress.flush(zlib_mode))
 355.         self.fileobj.flush()
 356. 
 357.     def fileno(self):
 358.         """Invoke the underlying file object's fileno() method.
 359. 
 360.         This will raise AttributeError if the underlying file object
 361.         doesn't support fileno().
 362.         """
 363.         return self.fileobj.fileno()
 364. 
 365.     def isatty(self):
 366.         return False
 367. 
 368.     def tell(self):
 369.         return self.offset
 370. 
 371.     def rewind(self):
 372.         '''Return the uncompressed stream file position indicator to the
 373.         beginning of the file'''
 374.         if self.mode != READ:
 375.             raise IOError("Can't rewind in write mode")
 376.         self.fileobj.seek(0)
 377.         self._new_member = True
 378.         self.extrabuf = ""
 379.         self.extrasize = 0
 380.         self.offset = 0
 381. 
 382.     def seek(self, offset, whence=0):
 383.         if whence:
 384.             if whence == 1:
 385.                 offset = self.offset + offset
 386.             else:
 387.                 raise ValueError('Seek from end not supported')
 388.         if self.mode == WRITE:
 389.             if offset < self.offset:
 390.                 raise IOError('Negative seek in write mode')
 391.             count = offset - self.offset
 392.             for i in range(count // 1024):
 393.                 self.write(1024 * '\0')
 394.             self.write((count % 1024) * '\0')
 395.         elif self.mode == READ:
 396.             if offset < self.offset:
 397.                 # for negative seek, rewind and do positive seek
 398.                 self.rewind()
 399.             count = offset - self.offset
 400.             for i in range(count // 1024):
 401.                 self.read(1024)
 402.             self.read(count % 1024)
 403. 
 404.     def readline(self, size=-1):
 405.         if size < 0:
 406.             size = sys.maxint
 407.             readsize = self.min_readsize
 408.         else:
 409.             readsize = size
 410.         bufs = []
 411.         while size != 0:
 412.             c = self.read(readsize)
 413.             i = c.find('\n')
 414. 
 415.             # We set i=size to break out of the loop under two
 416.             # conditions: 1) there's no newline, and the chunk is
 417.             # larger than size, or 2) there is a newline, but the
 418.             # resulting line would be longer than 'size'.
 419.             if (size <= i) or (i == -1 and len(c) > size):
 420.                 i = size - 1
 421. 
 422.             if i >= 0 or c == '':
 423.                 bufs.append(c[:i + 1])    # Add portion of last chunk
 424.                 self._unread(c[i + 1:])   # Push back rest of chunk
 425.                 break
 426. 
 427.             # Append chunk to list, decrease 'size',
 428.             bufs.append(c)
 429.             size = size - len(c)
 430.             readsize = min(size, readsize * 2)
 431.         if readsize > self.min_readsize:
 432.             self.min_readsize = min(readsize, self.min_readsize * 2, 512)
 433.         return ''.join(bufs) # Return resulting line
 434. 
 435.     def readlines(self, sizehint=0):
 436.         # Negative numbers result in reading all the lines
 437.         if sizehint <= 0:
 438.             sizehint = sys.maxint
 439.         L = []
 440.         while sizehint > 0:
 441.             line = self.readline()
 442.             if line == "":
 443.                 break
 444.             L.append(line)
 445.             sizehint = sizehint - len(line)
 446. 
 447.         return L
 448. 
 449.     def writelines(self, L):
 450.         for line in L:
 451.             self.write(line)
 452. 
 453.     def __iter__(self):
 454.         return self
 455. 
 456.     def next(self):
 457.         line = self.readline()
 458.         if line:
 459.             return line
 460.         else:
 461.             raise StopIteration
 462. 
 463. 
 464. def _test():
 465.     # Act like gzip; with -d, act like gunzip.
 466.     # The input file is not deleted, however, nor are any other gzip
 467.     # options or features supported.
 468.     args = sys.argv[1:]
 469.     decompress = args and args[0] == "-d"
 470.     if decompress:
 471.         args = args[1:]
 472.     if not args:
 473.         args = ["-"]
 474.     for arg in args:
 475.         if decompress:
 476.             if arg == "-":
 477.                 f = GzipFile(filename="", mode="rb", fileobj=sys.stdin)
 478.                 g = sys.stdout
 479.             else:
 480.                 if arg[-3:] != ".gz":
 481.                     print "filename doesn't end in .gz:", repr(arg)
 482.                     continue
 483.                 f = open(arg, "rb")
 484.                 g = __builtin__.open(arg[:-3], "wb")
 485.         else:
 486.             if arg == "-":
 487.                 f = sys.stdin
 488.                 g = GzipFile(filename="", mode="wb", fileobj=sys.stdout)
 489.             else:
 490.                 f = __builtin__.open(arg, "rb")
 491.                 g = open(arg + ".gz", "wb")
 492.         while True:
 493.             chunk = f.read(1024)
 494.             if not chunk:
 495.                 break
 496.             g.write(chunk)
 497.         if g is not sys.stdout:
 498.             g.close()
 499.         if f is not sys.stdin:
 500.             f.close()
 501. 
 502. if __name__ == '__main__':
 503.     _test()