February 01, 2012

Making HTTP Requests with Python

Python has a built in library for making HTTP requests. Well, it has two of them actually: urllib and urllib2. Unfortunately, both of them are not ideal. The API is thoroughly broken, and certain things are extremely hard.

Enter requests a fantastic third party python library. You can install it by typing pip install requests in your terminal if you have Pip installed.

In this tutorial, we’ll cover making simple HTTP requests and dealing with cookies, URL query string parameters and POST data.

Making a Simple GET Request

>>> import requests<br /><br />
>>> resp = requests.get(‘http://httpbin.org’)<br /><br />
>>> resp<br /><br />
<Response [200]>

>>> import requests

>>> resp = requests.get(‘http://httpbin.org’)

>>> resp

That’s it! resp is a Response object, which has several useful attributes. resp.content would be the actually content of the page, for instance. and resp.headers is a Python dictionary of the headers from the response. You can even see your request object with resp.request.

Why do I love requests? Because it’s simple. The above is extremely intuitive, and there are very Pythonic ways of doing many other things. Dealing with cookies is a good example.

Making Requests with Cookies

import requests</p><br />
<p>with requests.session() as s:<br /><br />
    # set a cookie<br /><br />
    s.get(‘http://httpbin.org/cookies/set/pmgcookie/mmmmmmcookie’)</p><br />
<p>    #is the cookie still there?<br /><br />
    resp = s.get(‘http://httpbin.org/cookies’)<br /><br />
    print resp.content # yup

import requests

with requests.session() as s:

# set a cookie

s.get(‘http://httpbin.org/cookies/set/pmgcookie/mmmmmmcookie’)

#is the cookie still there?

resp = s.get(‘http://httpbin.org/cookies’)

print resp.content # yup

The result

{<br /><br />
  “cookies”: {<br /><br />
    “pmgcookie”: “mmmmmmcookie”<br /><br />
  }<br /><br />
}

{

“cookies”: {

“pmgcookie”: “mmmmmmcookie”

}

Requests introduces a Session object, which acts as a sort of storage bin for data from previous requests. When you use requests.get() it actually creates a session for you. Using the with statement lets you make multiple requests with the same session easily. Meaning you can keep cookies around. You can also feed requests.get() a session object using the session keyword argument: requests.get('http://example.com', session=some_session)

Proxies and Proxy Authentication

Need to make a request through proxy? No problem. The proxies keyword argument of requests.get (and requests.post) makes it easy.

import requests</p><br />
<p>proxy = {‘http’: ‘http://some-proxy.com’}<br /><br />
resp = requests.get(‘http://httpbin.org’, proxies=proxy)

import requests

proxy = {‘http’: ‘http://some-proxy.com’}

resp = requests.get(‘http://httpbin.org’, proxies=proxy)

The value of proxies keyword arg should be a dictionary with the various protocols (eg. http and https) as keys and the proxy URI’s as values.

What about Proxy Authentication? If your proxy uses HTTP basic auth, you can use the auth keyword argument.

import requests</p><br />
<p>proxy = {‘http’: ‘http://some-proxy.com’}<br /><br />
auth = (‘username’, ‘password’,)<br /><br />
resp = requests.get(‘http://httpbin.org’, proxies=proxy, auth=auth)

import requests

proxy = {‘http’: ‘http://some-proxy.com’}

auth = (‘username’, ‘password’,)

resp = requests.get(‘http://httpbin.org’, proxies=proxy, auth=auth)

What about using the Proxy-Authentication header? Just use a the username and password in the URL like normal: http://user:password@some-proxy.com.

import requests</p><br />
<p>proxy = {‘http’: ‘http://user:password@some-proxy.com’}<br /><br />
resp = requests.get(‘http://httpbin.org’, proxies=proxy)

import requests

proxy = {‘http’: ‘http://user:password@some-proxy.com’}

resp = requests.get(‘http://httpbin.org’, proxies=proxy)

Query Strings & POST data

Need to send a URL query string with your request? You could just stick them on the end of the URI, but that’s no fun (and you’d have to worry about encoding them). Just use the params keyword arg.

import json<br /><br />
import requests</p><br />
<p># fetch!<br /><br />
params = {‘arg1’: ‘value1’, ‘arg2’: ‘value2’}<br /><br />
resp = requests.get(‘http://httpbin.org/get’, params=params)</p><br />
<p># let’s see what we got back<br /><br />
j = json.loads(resp.content)<br /><br />
print j.get(‘args’)

import json

import requests

# fetch!

params = {‘arg1’: ‘value1’, ‘arg2’: ‘value2’}

resp = requests.get(‘http://httpbin.org/get’, params=params)

# let’s see what we got back

j = json.loads(resp.content)

print j.get(‘args’)

The result:

http://httpbin.org/get?arg1=value1&arg2=value2<br /><br />
{u’arg1′: u’value1′, u’arg2′: u’value2′}

1 2	http://httpbin.org/get?arg1=value1&arg2=value2 {u‘arg1’: u‘value1’, u‘arg2’: u‘value2’}

Making POST requests is just as simple. Use the data keyword argument to send whatever you need to the remote server.

import json<br /><br />
import requests</p><br />
<p># fetch!<br /><br />
params = {‘arg1’: ‘value1’, ‘arg2’: ‘value2’}<br /><br />
resp = requests.post(‘http://httpbin.org/post’, data=params)</p><br />
<p># what did we get back?<br /><br />
j = json.loads(resp.content)<br /><br />
print j.get(‘form’)

import json

import requests

# fetch!

params = {‘arg1’: ‘value1’, ‘arg2’: ‘value2’}

resp = requests.post(‘http://httpbin.org/post’, data=params)

# what did we get back?

j = json.loads(resp.content)

print j.get(‘form’)

And the result:

{u’arg1′: u’value1′, u’arg2′: u’value2′}

1	{u‘arg1’: u‘value1’, u‘arg2’: u‘value2’}

Requests is, by far, one of the most well though out python HTTP libraries today. Use it, and you’ll love its simplicity and easy. Head over to the requests docs to learn more!

Making HTTP Requests with Python

Making a Simple GET Request

Making Requests with Cookies

Proxies and Proxy Authentication

Query Strings & POST data

Latest White Papers

Featured Posts

Categories

Subscribe to the PMG Newsletter