Improved some error messages for command line processing.
[python/dscho.git] / Lib / gzip.py
blobe25464caf9127db7e29192b30696f377c5f340dc
1 import time
2 import string
3 import zlib
4 import __builtin__
6 # implements a python function that reads and writes a gzipped file
7 # the user of the file doesn't have to worry about the compression,
8 # but random access is not allowed
10 # based on Andrew Kuchling's minigzip.py distributed with the zlib module
12 FTEXT, FHCRC, FEXTRA, FNAME, FCOMMENT = 1, 2, 4, 8, 16
14 READ, WRITE = 1, 2
16 def write32(output, value):
17 t = divmod(value, 256)
18 b1 = chr(t[1])
20 t = divmod(t[0], 256)
21 b2 = chr(t[1])
23 t = divmod(t[0], 256)
24 b3 = chr(t[1])
26 t = divmod(t[0], 256)
27 b4 = chr(t[1])
29 buf = b1 + b2 + b3 + b4
30 output.write(buf)
33 def read32(input):
34 buf = input.read(4)
35 v = ord(buf[0])
36 v = v + (ord(buf[1]) << 8)
37 v = v + (ord(buf[2]) << 16)
38 v = v + (ord(buf[3]) << 24)
39 return v
41 def open(filename, mode="r", compresslevel=9):
42 return GzipFile(filename, mode, compresslevel)
44 class GzipFile:
46 myfileobj = None
48 def __init__(self, filename=None, mode=None,
49 compresslevel=9, fileobj=None):
50 if fileobj is None:
51 fileobj = self.myfileobj = __builtin__.open(filename, mode or 'r')
52 if filename is None:
53 if hasattr(fileobj, 'name'): filename = fileobj.name
54 else: filename = ''
55 if mode is None:
56 if hasattr(fileobj, 'mode'): mode = fileobj.mode
57 else: mode = 'r'
59 if mode[0:1] == 'r':
60 self.mode = READ
61 self._init_read()
62 self.filename = filename
63 self.decompress = zlib.decompressobj(-zlib.MAX_WBITS)
65 elif mode[0:1] == 'w':
66 self.mode = WRITE
67 self._init_write(filename)
68 self.compress = zlib.compressobj(compresslevel,
69 zlib.DEFLATED,
70 -zlib.MAX_WBITS,
71 zlib.DEF_MEM_LEVEL,
73 else:
74 raise ValueError, "Mode " + mode + " not supported"
76 self.fileobj = fileobj
78 if self.mode == WRITE:
79 self._write_gzip_header()
80 elif self.mode == READ:
81 self._read_gzip_header()
83 def __repr__(self):
84 s = repr(self.fileobj)
85 return '<gzip ' + s[1:-1] + ' ' + hex(id(self)) + '>'
87 def _init_write(self, filename):
88 if filename[-3:] != '.gz':
89 filename = filename + '.gz'
90 self.filename = filename
91 self.crc = zlib.crc32("")
92 self.size = 0
93 self.writebuf = []
94 self.bufsize = 0
96 def _write_gzip_header(self):
97 self.fileobj.write('\037\213') # magic header
98 self.fileobj.write('\010') # compression method
99 fname = self.filename[:-3]
100 flags = 0
101 if fname:
102 flags = FNAME
103 self.fileobj.write(chr(flags))
104 write32(self.fileobj, int(time.time()))
105 self.fileobj.write('\002')
106 self.fileobj.write('\377')
107 if fname:
108 self.fileobj.write(fname + '\000')
110 def _init_read(self):
111 self.crc = zlib.crc32("")
112 self.size = 0
113 self.extrabuf = ""
114 self.extrasize = 0
116 def _read_gzip_header(self):
117 magic = self.fileobj.read(2)
118 if magic != '\037\213':
119 raise RuntimeError, 'Not a gzipped file'
120 method = ord( self.fileobj.read(1) )
121 if method != 8:
122 raise RuntimeError, 'Unknown compression method'
123 flag = ord( self.fileobj.read(1) )
124 # modtime = self.fileobj.read(4)
125 # extraflag = self.fileobj.read(1)
126 # os = self.fileobj.read(1)
127 self.fileobj.read(6)
129 if flag & FEXTRA:
130 # Read & discard the extra field, if present
131 xlen=ord(self.fileobj.read(1))
132 xlen=xlen+256*ord(self.fileobj.read(1))
133 self.fileobj.read(xlen)
134 if flag & FNAME:
135 # Read and discard a null-terminated string containing the filename
136 while (1):
137 s=self.fileobj.read(1)
138 if not s or s=='\000': break
139 if flag & FCOMMENT:
140 # Read and discard a null-terminated string containing a comment
141 while (1):
142 s=self.fileobj.read(1)
143 if not s or s=='\000': break
144 if flag & FHCRC:
145 self.fileobj.read(2) # Read & discard the 16-bit header CRC
148 def write(self,data):
149 if self.fileobj is None:
150 raise ValueError, "write() on closed GzipFile object"
151 if len(data) > 0:
152 self.size = self.size + len(data)
153 self.crc = zlib.crc32(data, self.crc)
154 self.fileobj.write( self.compress.compress(data) )
156 def writelines(self,lines):
157 self.write(string.join(lines))
159 def read(self, size=None):
160 if self.extrasize <= 0 and self.fileobj is None:
161 return ''
163 readsize = 1024
164 if not size: # get the whole thing
165 try:
166 while 1:
167 self._read(readsize)
168 readsize = readsize * 2
169 except EOFError:
170 size = self.extrasize
171 else: # just get some more of it
172 try:
173 while size > self.extrasize:
174 self._read(readsize)
175 readsize = readsize * 2
176 except EOFError:
177 if size > self.extrasize:
178 size = self.extrasize
180 chunk = self.extrabuf[:size]
181 self.extrabuf = self.extrabuf[size:]
182 self.extrasize = self.extrasize - size
184 return chunk
186 def _unread(self, buf):
187 self.extrabuf = buf + self.extrabuf
188 self.extrasize = len(buf) + self.extrasize
190 def _read(self, size=1024):
191 try:
192 buf = self.fileobj.read(size)
193 except AttributeError:
194 raise EOFError, "Reached EOF"
195 if buf == "":
196 uncompress = self.decompress.flush()
197 if uncompress == "":
198 self._read_eof()
199 self.fileobj = None
200 raise EOFError, 'Reached EOF'
201 else:
202 uncompress = self.decompress.decompress(buf)
203 self.crc = zlib.crc32(uncompress, self.crc)
204 self.extrabuf = self.extrabuf + uncompress
205 self.extrasize = self.extrasize + len(uncompress)
206 self.size = self.size + len(uncompress)
208 def _read_eof(self):
209 # Andrew writes:
210 ## We've read to the end of the file, so we have to rewind in order
211 ## to reread the 8 bytes containing the CRC and the file size. The
212 ## decompressor is smart and knows when to stop, so feeding it
213 ## extra data is harmless.
214 self.fileobj.seek(-8, 2)
215 crc32 = read32(self.fileobj)
216 isize = read32(self.fileobj)
217 if crc32 != self.crc:
218 self.error = "CRC check failed"
219 elif isize != self.size:
220 self.error = "Incorrect length of data produced"
222 def close(self):
223 if self.mode == WRITE:
224 self.fileobj.write(self.compress.flush())
225 write32(self.fileobj, self.crc)
226 write32(self.fileobj, self.size)
227 self.fileobj = None
228 elif self.mode == READ:
229 self.fileobj = None
230 if self.myfileobj:
231 self.myfileobj.close()
232 self.myfileobj = None
234 def flush(self):
235 self.fileobj.flush()
237 def seek(self):
238 raise IOError, 'Random access not allowed in gzip files'
240 def tell(self):
241 raise IOError, 'I won\'t tell() you for gzip files'
243 def isatty(self):
244 return 0
246 def readline(self):
247 bufs = []
248 readsize = 100
249 while 1:
250 c = self.read(readsize)
251 i = string.find(c, '\n')
252 if i >= 0 or c == '':
253 bufs.append(c[:i+1])
254 self._unread(c[i+1:])
255 return string.join(bufs, '')
256 bufs.append(c)
257 readsize = readsize * 2
259 def readlines(self):
260 buf = self.read()
261 lines = string.split(buf, '\n')
262 for i in range(len(lines)-1):
263 lines[i] = lines[i] + '\n'
264 if lines and not lines[-1]:
265 del lines[-1]
266 return lines
268 def writelines(self, L):
269 for line in L:
270 self.write(line)
273 def _test():
274 # Act like gzip; with -d, act like gunzip.
275 # The input file is not deleted, however, nor are any other gzip
276 # options or features supported.
277 import sys
278 args = sys.argv[1:]
279 decompress = args and args[0] == "-d"
280 if decompress:
281 args = args[1:]
282 if not args:
283 args = ["-"]
284 for arg in args:
285 if decompress:
286 if arg == "-":
287 f = GzipFile(filename="", mode="rb", fileobj=sys.stdin)
288 g = sys.stdout
289 else:
290 if arg[-3:] != ".gz":
291 print "filename doesn't end in .gz:", `arg`
292 continue
293 f = open(arg, "rb")
294 g = __builtin__.open(arg[:-3], "wb")
295 else:
296 if arg == "-":
297 f = sys.stdin
298 g = GzipFile(filename="", mode="wb", fileobj=sys.stdout)
299 else:
300 f = __builtin__.open(arg, "rb")
301 g = open(arg + ".gz", "wb")
302 while 1:
303 chunk = f.read(1024)
304 if not chunk:
305 break
306 g.write(chunk)
307 if g is not sys.stdout:
308 g.close()
309 if f is not sys.stdin:
310 f.close()
312 if __name__ == '__main__':
313 _test()