How do I modify this download function in Python?

Question

Right now, it's iffy. Gzip, images, sometimes it doesn't work.

How do I modify this download function so that it can work with anything? (Regardless of gzip or any header?)

How do I automatically " Detect " if it's gzip? I don't want to always pass True/False, like I do right now.

def download(source_url, g = False, correct_url = True):
    try:
        socket.setdefaulttimeout(10)
        agents = ['Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)','Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 5.1)','Microsoft Internet Explorer/4.0b1 (Windows 95)','Opera/8.00 (Windows NT 5.1; U; en)']
        ree = urllib2.Request(source_url)
        ree.add_header('User-Agent',random.choice(agents))
        ree.add_header('Accept-encoding', 'gzip')
        opener = urllib2.build_opener()
        h = opener.open(ree).read()
        if g:
            compressedstream = StringIO(h)
            gzipper = gzip.GzipFile(fileobj=compressedstream)
            data = gzipper.read()
            return data
        else:
            return h
    except Exception, e:
        return ""

Answer 1

Check the Content-Encoding header:

import urllib2
import socket
import random
import StringIO
import gzip

def download(source_url):
    try:
        socket.setdefaulttimeout(10)
        agents = ['Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)','Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 5.1)','Microsoft Internet Explorer/4.0b1 (Windows 95)','Opera/8.00 (Windows NT 5.1; U; en)']
        ree = urllib2.Request(source_url)
        ree.add_header('User-Agent',random.choice(agents))
        ree.add_header('Accept-encoding', 'gzip')
        opener = urllib2.build_opener()
        response = opener.open(ree)
        encoding=response.headers.getheader('Content-Encoding')
        content = response.read()
        if encoding and 'gzip' in encoding:
            compressedstream = StringIO.StringIO(content)
            gzipper = gzip.GzipFile(fileobj=compressedstream)
            data = gzipper.read()
            return data
        else:
            return content
    except urllib2.URLError as e:
        return ""

data=download('http://api.stackoverflow.com/1.0/questions/3708418?type=jsontext')
print(data)

If you are dealing with a server which does not report the Content-Encoding as gzip, then you could be more aggressive by just try ing first:

def download(source_url):
    try:
        socket.setdefaulttimeout(10)
        agents = ['Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)','Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 5.1)','Microsoft Internet Explorer/4.0b1 (Windows 95)','Opera/8.00 (Windows NT 5.1; U; en)']
        ree = urllib2.Request(source_url)
        ree.add_header('User-Agent',random.choice(agents))
        ree.add_header('Accept-encoding', 'gzip')
        opener = urllib2.build_opener()
        response = opener.open(ree)
        content = response.read()
        compressedstream = StringIO.StringIO(content)
        gzipper = gzip.GzipFile(fileobj=compressedstream)
        try:
            data = gzipper.read()
        except IOError:
            data = content
        return data        
    except urllib2.URLError as e:
        return ""

Answer 2

To detect what type of data you are downloading you should replace the h = opener.open(ree).read() for h = opener.open(ree) .

Now in h you have the response object. You can analyze the headers by using the h.headers (dict-like) object. Particularly you will be interested in the headers 'content-type' and 'content-encoding'. You can determine the content being sent analyzing it.

def download(source_url, correct_url = True):
    try:
        socket.setdefaulttimeout(10)
        agents = ['Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)','Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 5.1)','Microsoft Internet Explorer/4.0b1 (Windows 95)','Opera/8.00 (Windows NT 5.1; U; en)']
        ree = urllib2.Request(source_url)
        ree.add_header('User-Agent',random.choice(agents))
        ree.add_header('Accept-encoding', 'gzip')
        opener = urllib2.build_opener()
        h = opener.open(ree)
        if 'gzip' in h.headers.get('content-type', '') or
           'gzip' in h.headers.get('content-encoding', ''):
            compressedstream = StringIO(h.read())
            gzipper = gzip.GzipFile(fileobj=compressedstream)
            data = gzipper.read()
            return data
        else:
            return h.read()
    except Exception, e:
        return ""

Answer 3

import urllib2
import StringIO
import gzip

req = urllib2.Request('http:/foo/')
h = urllib2.urlopen(req)
data = resp.read()
if 'gzip' in resp.headers['Content-Encoding']:
    compressedstream = StringIO(h)
    gzipper = gzip.GzipFile(fileobj=compressedstream)
    data = gzipper.read()

# etc...