import time import string import zlib import StringIO import __builtin__ # implements a python function that reads and writes a gzipped file # the user of the file doesn't have to worry about the compression, # but sequential access is not allowed # based on Andrew Kuchling's minigzip.py distributed with the zlib module FTEXT, FHCRC, FEXTRA, FNAME, FCOMMENT = 1, 2, 4, 8, 16 READ, WRITE = 1, 2 def write32(output, value): t = divmod(value, 256) b1 = chr(t[1]) t = divmod(t[0], 256) b2 = chr(t[1]) t = divmod(t[0], 256) b3 = chr(t[1]) t = divmod(t[0], 256) b4 = chr(t[1]) buf = b1 + b2 + b3 + b4 output.write(buf) def read32(input): buf = input.read(4) v = ord(buf[0]) v = v + (ord(buf[1]) << 8) v = v + (ord(buf[2]) << 16) v = v + (ord(buf[3]) << 24) return v def open(filename, mode="r", compresslevel=9): return GzipFile(filename, mode, compresslevel) class GzipFile: myfileobj = None def __init__(self, filename=None, mode=None, compresslevel=9, fileobj=None): if fileobj is None: fileobj = self.myfileobj = __builtin__.open(filename, mode or 'r') if filename is None: if hasattr(fileobj, 'name'): filename = fileobj.name else: filename = 'GzippedFile' if mode is None: if hasattr(fileobj, 'mode'): mode = fileobj.mode else: mode = 'r' if mode[0:1] == 'r': self.mode = READ self._init_read() self.filename = filename self.decompress = zlib.decompressobj(-zlib.MAX_WBITS) elif mode[0:1] == 'w': self.mode = WRITE self._init_write(filename) self.compress = zlib.compressobj(compresslevel, zlib.DEFLATED, -zlib.MAX_WBITS, zlib.DEF_MEM_LEVEL, 0) else: raise ValueError, "Mode " + mode + " not supported" self.fileobj = fileobj if self.mode == WRITE: self._write_gzip_header() elif self.mode == READ: self._read_gzip_header() def __repr__(self): s = repr(self.fileobj) return '' def _init_write(self, filename): if filename[-3:] != '.gz': filename = filename + '.gz' self.filename = filename self.crc = zlib.crc32("") self.size = 0 self.writebuf = [] self.bufsize = 0 def _write_gzip_header(self): self.fileobj.write('\037\213') # magic header self.fileobj.write('\010') # compression method self.fileobj.write(chr(FNAME)) write32(self.fileobj, int(time.time())) self.fileobj.write('\002') self.fileobj.write('\377') self.fileobj.write(self.filename[:-3] + '\000') def _init_read(self): self.crc = zlib.crc32("") self.size = 0 self.extrabuf = "" self.extrasize = 0 def _read_gzip_header(self): magic = self.fileobj.read(2) if magic != '\037\213': raise RuntimeError, 'Not a gzipped file' method = ord( self.fileobj.read(1) ) if method != 8: raise RuntimeError, 'Unknown compression method' flag = ord( self.fileobj.read(1) ) # modtime = self.fileobj.read(4) # extraflag = self.fileobj.read(1) # os = self.fileobj.read(1) self.fileobj.read(6) if flag & FEXTRA: # Read & discard the extra field, if present xlen=ord(self.fileobj.read(1)) xlen=xlen+256*ord(self.fileobj.read(1)) self.fileobj.read(xlen) if flag & FNAME: # Read and discard a null-terminated string containing the filename while (1): s=self.fileobj.read(1) if s=='\000': break if flag & FCOMMENT: # Read and discard a null-terminated string containing a comment while (1): s=self.fileobj.read(1) if s=='\000': break if flag & FHCRC: self.fileobj.read(2) # Read & discard the 16-bit header CRC def write(self,data): if self.fileobj is None: raise ValueError, "write() on closed GzipFile object" if len(data) > 0: self.size = self.size + len(data) self.crc = zlib.crc32(data, self.crc) self.fileobj.write( self.compress.compress(data) ) def writelines(self,lines): self.write(string.join(lines)) def read(self,size=None): if self.extrasize <= 0 and self.fileobj is None: return '' if not size: # get the whole thing try: while 1: self._read() except EOFError: size = self.extrasize else: # just get some more of it try: while size > self.extrasize: self._read() except EOFError: pass chunk = self.extrabuf[:size] self.extrabuf = self.extrabuf[size:] self.extrasize = self.extrasize - size return chunk def _read(self): buf = self.fileobj.read(1024) if buf == "": uncompress = self.decompress.flush() if uncompress == "": self._read_eof() self.fileobj = None raise EOFError, 'Reached EOF' else: uncompress = self.decompress.decompress(buf) self.crc = zlib.crc32(uncompress, self.crc) self.extrabuf = self.extrabuf + uncompress self.extrasize = self.extrasize + len(uncompress) self.size = self.size + len(uncompress) def _read_eof(self): # Andrew writes: ## We've read to the end of the file, so we have to rewind in order ## to reread the 8 bytes containing the CRC and the file size. The ## decompressor is smart and knows when to stop, so feeding it ## extra data is harmless. self.fileobj.seek(-8, 2) crc32 = read32(self.fileobj) isize = read32(self.fileobj) if crc32 != self.crc: self.error = "CRC check failed" elif isize != self.size: self.error = "Incorrect length of data produced" def close(self): if self.mode == WRITE: self.fileobj.write(self.compress.flush()) write32(self.fileobj, self.crc) write32(self.fileobj, self.size) self.fileobj = None elif self.mode == READ: self.fileobj = None if self.myfileobj: self.myfileobj.close() self.myfileobj = None def flush(self): self.fileobj.flush() def seek(self): raise IOError, 'Random access not allowed in gzip files' def tell(self): raise IOError, 'I won\'t tell() you for gzip files' def isatty(self): return 0 def readline(self): # XXX This function isn't implemented in a very efficient way line="" while 1: c = self.read(1) line = line + c if c=='\n' or c=="": break return line def readlines(self): L=[] line = self.readline() while line!="": L.append(line) line = self.readline() return L def writelines(self, L): for line in L: self.write(line)