encoding_gzip.py

一个java写的proxy的例子

# gunzip.py, [e-mail], March 2000a
#
# Implements the minimal amount of work needed to ungzip an input stream
#
# Based on gzip.py in the standard Python distribution,
# but exports a proxy4 encoding interface:
#      decode(string) => return as much of the string as can be decoded
#      flush()        => return everything else

# TEST CASE:
#    http://tv.excite.com/grid

import zlib
from string import find

class GunzipStream:
    # Flags in the gzip header
    FTEXT, FHCRC, FEXTRA, FNAME, FCOMMENT = 1, 2, 4, 8, 16
    
    def __init__(self):
        self.decompressor = zlib.decompressobj(-zlib.MAX_WBITS)
        self.buffer = ''
        self.header_seen = 0
        self.closed = 0

    def attempt_header_read(self):
        "Try to parse the header from buffer, and if we can, set flag"
        if len(self.buffer) < 10: # Incomplete fixed part of header
            return ''
        
        magic = self.buffer[:2]
        if magic != '\037\213':
            raise zlib.error, 'not gzip format'
            
        method = ord(self.buffer[2])
        if method != 8:
            raise zlib.error, 'unknown compression method'
        
        flag = ord(self.buffer[3])
        # Skip until byte 10
        string = self.buffer[10:]
        
        if flag & self.FEXTRA:
            # Read & discard the extra field, if present
            if len(string) < 2: return '' # Incomplete
            xlen=ord(string[0])
            xlen=xlen+256*ord(string[1])
            if len(string) < 2+xlen: return '' # Incomplete
            string = string[2+xlen:]
            
        if flag & self.FNAME:
            # Read and discard a null-terminated string containing the filename
            i = find(string, '\000')
            if i < 0: return '' # Incomplete
            string = string[i+1:]
            
        if flag & self.FCOMMENT:
            # Read and discard a null-terminated string containing a comment
            i = find(string, '\000')
            if i < 0: return '' # Incomplete
            string = string[i+1:]
            
        if flag & self.FHCRC:
            # Read & discard the 16-bit header CRC
            if len(string) < 2: return '' # Incomplete
            string = string[2:]

        # We actually got through the header
        self.buffer = string
        self.header_seen = 1

    def decode(self, string):
        if not self.header_seen:
            # Try to parse the header
            self.buffer = self.buffer + string
            string = ''
            self.attempt_header_read()
            if self.header_seen:
                # Put the rest of the buffer back into the string,
                # for zlib use
                string = self.buffer
                self.buffer = ''
            else:
                # We haven't finished parsing the header
                return ''
            
        # We have seen the header, so we can move on to zlib
        return self.decompressor.decompress(string)

    def flush(self):
        if not self.header_seen:
            # We still haven't finished parsing the header .. oh well
            return ''
        else:
            return self.decompressor.flush()