Why does python urllib2 urlopen return something different from the browser on API call

Question

I need to routinely access and parse XML data from a website of the form:

https://api.website.com/stuff/getCurrentData?security_key=blah

I cannot post the actual connections because of the secure nature of the data. When I put this url into my browser (Safari), I get XML returned.

When I call this through urllib2, I get junk.

f = urllib2.urlopen("https://api.website.com/stuff/getCurrentData?security_key=blah") 
s = f.read()
f.close()
s
'\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\x03\xc5\x96mo\xda0\x10\xc7\xdf\xf7SX\xbc\xda4\x15\xc7y\x00R\xb9\xae\xfa\xb4U\x1a-\x150M{5y\xe1\x06V\x13\x079\x0e\x14>\xfd\x9c\x84\xb0\xd2\xa4S\xa4L\xe5\x95\xef\xeeo

This post Urllib's urlopen breaking on some sites (e.g. StackApps api): returns garbage results seems to be a similar problem, but it refers to JSON instead of XML. Following the instructions to look at headers, I think that I am getting GZIP data returned. {I did the test suggested, posted here}

req = urllib2.Request("https://api.website.com/stuff/getCurrentData?security_key=blah",
                      headers={'Accept-Encoding': 'gzip, identity'})
conn = urllib2.urlopen(req)
val = conn.read()
conn.close()
val[0:25]
'\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\x03\xc5\x96]o\xda0\x14\x86\xef\xfb+,\xae6M'

In that post, there was some suggestion that this could be a local problem, so I tried an example site.

f = urllib2.urlopen("http://www.python.org")
s = f.read()
f.close()
s
'<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">\n\n\n<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">\n\n<head>\n  <meta http-equiv="content-type" content="text/html; charset=utf-8" />\n  <title>Python Programming Language &ndash; Official Website</title>\n

This works just fine, so I think it has something to do with the site API that I am actually trying to access.

This post Why does text retrieved from pages sometimes look like gibberish? suggested that I might need to do something with "Selenium" but then the poster said the problem "fixed itself" which does not help me figure out what is wrong.

Am I not able to use python to download secure data? Do I need to use something different besides urlib2 and url open?

I am running python 2.7 on Mac OSX 10.7.5

Check the headers; does the server return compressed data? — Martijn Pieters
I don't know what you mean, @MartijnPieters -- I'm still new to this. But...this is what I get when I ask for headers print req.headers {'Accept-encoding': 'gzip, identity'} — jessi
Those are the headers you sent. Check conn.info().headers instead. — Martijn Pieters

Martijn Pieters Martijn Pieters · Accepted Answer · 2013-05-31T14:17:09

You are retrieving GZIPped, compressed data; the server expressly tells you it does with Content-Encoding: gzip. Either use the zlib library to decompress the data:

import zlib

decomp = zlib.decompressobj(16 + zlib.MAX_WBITS)
data = decomp.decompress(val)

or use a library that supports transparent decompression if the response headers indicate compression has been used, like requests.

Why does python urllib2 urlopen return something different from the browser on API call

2 Answers