How to decode the gzip compressed data returned in a HTTP Response in python?

Question

I have created a client/server architecture in python, I take HTTP request from the client which is served by requesting another HTTP server through my code.

When I get the response from the third server I am not able to decode the gzip compressed data, I first split the response data using \r\n as separation character which got me the data as the last item in the list then I tried decompressing it with

zlib.decompress(data[-1]) 

but it is giving me an error of incorrect headers. How should I go with this problem ?

Code

client_reply = ''
                 while 1:
                     chunk = server2.recv(512)
                     if len(chunk) :
                         client.send(chunk)
                         client_reply += chunk
                     else:
                         break
                 client_split = client_reply.split("\r\n")
                 print client_split[-1].decode('zlib')

I want to read the data that is been transferred between the client and the 2nd server.

Answer 1

Specify the wbits when using zlib.decompress(string, wbits, bufsize) see end of "troubleshooting" for example.

Troubleshooting

Lets start out with a a curl command that downloads a byte-range response with an unknown "content-encoding" (note: we know before hand it is some sort of compressed thing, mabye deflate maybe gzip):

export URL="https://commoncrawl.s3.amazonaws.com/crawl-data/CC-MAIN-2016-18/segments/1461860106452.21/warc/CC-MAIN-20160428161506-00007-ip-10-239-7-51.ec2.internal.warc.gz"
curl -r 266472196-266527075 $URL | gzip -dc | tee hello.txt

With the following response headers:

HTTP/1.1 206 Partial Content
x-amz-id-2: IzdPq3DAPfitkgdXhEwzBSwkxwJRx9ICtfxnnruPCLSMvueRA8j7a05hKr++Na6s
x-amz-request-id: 14B89CED698E0954
Date: Sat, 06 Aug 2016 01:26:03 GMT
Last-Modified: Sat, 07 May 2016 08:39:18 GMT
ETag: "144a93586a13abf27cb9b82b10a87787"
Accept-Ranges: bytes
Content-Range: bytes 266472196-266527075/711047506
Content-Type: application/octet-stream
Content-Length: 54880
Server: AmazonS3

So to the point.

Lets display the hex output of the first 10 bytes: curl -r 266472196-266472208 $URL | xxd

hex output:

0000000: 1f8b 0800 0000 0000 0000 ecbd eb

We can see some basics of what we are working with with the hex values.

Roughly meaning its probably a gzip ( 1f8b ) using deflate ( 0800 ) without a modification time ( 0000 0000 ), or any extra flags set ( 00 ), using a fat32 system( 00 ).

Please refer to section 2.3 / 2.3.1: https://www.rfc-editor.org/rfc/rfc1952#section-2.3.1

So onto the python:

>>> import requests
>>> url = 'https://commoncrawl.s3.amazonaws.com/crawl-data/CC-MAIN-2016-18/segments/1461860106452.21/warc/CC-MAIN-20160428161506-00006-ip-10-239-7-51.ec2.internal.warc.gz'
>>> response = requests.get(url, params={"range":"bytes=257173173-257248267"})
>>> unknown_compressed_data = response.content

notice anything similar?:

>>> unknown_compressed_data[:10]
'\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\x00'

And on to the decompression let's just try at random based on the (documentation):

>>> import zlib

"zlib.error: Error -2 while preparing to decompress data: inconsistent stream state":

>>> zlib.decompress(unknown_compressed_data, -31)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
zlib.error: Error -2 while preparing to decompress data: inconsistent stream state

"Error -3 while decompressing data: incorrect header check":

>>> zlib.decompress(unknown_compressed_data)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
zlib.error: Error -3 while decompressing data: incorrect header check

"zlib.error: Error -3 while decompressing data: invalid distance too far back":

>>> zlib.decompress(unknown_compressed_data, 30)
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
zlib.error: Error -3 while decompressing data: invalid distance too far back

Possible solution:

>>> zlib.decompress(unknown_compressed_data, 31)
'WARC/1.0\r\nWARC-Type: response\r\nWARC-Date: 2016-04-28T20:14:16Z\r\nWARC-Record-ID: <urn:uu

Answer 2

According to https://www.w3.org/Protocols/rfc2616/rfc2616-sec6.html the headers and the body are separated by an empty line containing only CRLF characters. You could try

client_split = client_reply.split("\r\n\r\n",1)
print client_split[1].decode('zlib')

The split finds the empty line and the additional parameter limits the number of splits - the result being array with two items, headers and body. But it is hard to recommend anything without knowing more about your code and the actual string being split.