Sockets in Python

Sockets are the lead pipes of computer networks: they let you connect with other devices so that information can flow freely. As you might expect, they're widely used on the Internet. Peyton McCullough explains how to code sockets in Python.

Introduction

One of the must-know features of any language is sockets, and, no, I’m not talking about the ones positioned on your ceiling. While many novice scripters will run away at the sound of the word, sockets in Python are not hard to learn and work with at all. In this article, I will explain the basics and even apply sockets in a real-world situation.

Please note that this article assumes that you have a basic command of the Python language. If not, you can either read the official tutorial or buy a book – or both. If you’re looking for a good book, you should head over to O’Reilly & Associates.

With that stated, let’s get started.

The Anatomy of a Socket

Let’s start off simple: what is a socket? The easiest way to learn is to compare computer sockets to lead pipes. Let’s say you have some sort of machine that has two pipes. One pipe pumps in water and the other pipe pumps out water. This will be your server. Now, imagine you have a simple box that has one pipe going through it. Let’s say that you hook it up to your pump machine. It will receive the water from the machine and then give water back. This will be your client. If you wanted to, you could hook up more machines and pipes to this to form a network of clients and servers.

Sockets are useful. The Web browser that you are viewing this page with is connecting to the server that this page is in by way of a socket. If you were to connect to one of your favorite online games or chat networks, you would be using sockets as well. As you can see, sockets are both both powerful and widely-used.

Using sockets in Python is quite easy. Open up your Python command line, and let’s get to some code.

The first thing we must do is import the socket library:

>>> import socket

The socket library contains all the tools we need to work with sockets. The next thing we need to do is create a socket. This is simple. Execute the following code:

>>> mySocket = socket.socket ( socket.AF_INET, socket.SOCK_STREAM )

This creates a stream socket. You can also work with datagrams by replacing SOCK_STREAM with SOCK_DGRAM. A socket stream is where there is a constant connection between the client and server that stays alive until it is closed, and both the client and the server know if the connection is still alive. With datagrams, however, that is not the case. The connection is not kept alive, and your data might not even be received. Although datagrams can sound like a bad idea at first, they have their purposes. It might be easier and faster to use datagrams in certain situations.

Writing a Simple Server

Let’s write a simple server. If you’ll remember, a server is anything that receives connections from other computers, clients. Create a new Python file named server.py and insert the following code into it:

import socket mySocket = socket.socket ( socket.AF_INET, socket.SOCK_STREAM ) mySocket.bind ( ( '', 2727 ) ) mySocket.listen ( 1 ) while True: channel, details = mySocket.accept() print 'We have opened a connection with', details print channel.recv ( 100 ) channel.send ( 'Green-eyed monster.' ) channel.close()

That’s quite a mouthful, so let’s split it up into something we can understand. The first two lines should already look familiar. We create a new socket to use. In the third line, we open port 2727 for connections.

To understand what a port is, let’s go back to our analogy. Picture the pump machine with thousands of pipes leading in and out of it. Each pipe would be a port, and clients would have the option of connecting to different ports. However, each port would be different – some might pump out green water, and others might pump out orange water.

The next line tells our socket to wait, or listen, for clients. Following that, there is an infinite loop. In this loop, we accept client connections, print out the client's address, print out the client's message and finally send a message back, closing the connection when we're done.

While the socket library is good for many things, there are several libraries built on the socket library that make development easier and faster. Say you wanted to connect to the official Python site to retrieve a piece of information. You could accomplish this task with the socket library, but there's no use working at such a low level when other libraries could do this in a shorter amount of lines. There are libraries for HTTP, FTP, Gopher, telnet and more.

A Real-World Problem and a Pythonic Solution

Let's say that you need to convert United States Dollars into Euros. This may seem complex at first, but it really isn't. It can be accomplished with sockets. We'll use the urllib library to connect to 'http://xe.com' and convert ten dollars into Euros.

Open up the Python command line and import the urllib module:

>>> import urllib

Connecting to the currency converter website I mentioned is simple with the urllib library -- a lot simpler than it would have been if I had used the socket library alone. It can be accomplished with the following line of code:

>>> currency = urllib.urlopen ( 'http://xe.com/ucc/convert.cgi?Amount=10&From=USD&To=EUR' ) >>> data = currency.read()

If you attempt to print out the data variable, you will get a good bit of junk that is not necessary in our script. However, here are the key lines:

<TD ALIGN=RIGHT><FONT FACE="Arial,Helvetica"> <FONT SIZE=+1><B>10.00 USD</B></FONT><FONT SIZE=-1><BR>United States Dollars </FONT></TD> <TD> </TD> <TD ALIGN=CENTER><FONT FACE="Arial,Helvetica" SIZE=+3><B>=</B></FONT></TD> <TD> </TD> <TD ALIGN=LEFT><FONT FACE="Arial,Helvetica"> <FONT SIZE=+1><B>7.46213 EUR</B></FONT><FONT SIZE=-1><BR>Euro </FONT></TD>

( Note that the values above are outdated. )

We can use regular expressions to sort through all the junk and take out the figures we need from the above code:

>>> import re >>> euros = re.search ( '(\d*)\.(\d*) EUR', data ) >>> print '10 United States Dollars are equal to...' >>> print euros.group ( 1 ) + '.' + euros.group ( 2 ), 'Euros.'

We also need to clean up and close the connection like we did earlier:

>>> currency.close()

That's it, we've solved our problem in only a few lines! If we were to do it with the socket library, however, it would have required a number of more lines to accomplish the same result.

You have learned that sockets are a valuable tool in the Python scripting language. They enable you to communicate with other computers, clients and servers, and to exchange data. Sockets are used to accomplish many common computer-related tasks, such as browsing Web pages. Though many novices assume that sockets are difficult, the opposite is true. Python also includes a number of libraries built on top of the socket library that allow you to easily write networking applications.

You now have the knowledge and tools to construct scripts that use sockets. Where you choose to go from here is entirely up to you. However, you may wish to browse through the standard library to find modules to aid development using sockets. Some modules that may help you include the select, asyncore, asynchat, ftplib ( FTP ), poplib ( mail ), imaplib ( mail ), smtplib ( mail ), telnetlib ( telnet ), nnttplib ( news ), SocketServer ( server ), BaseHTTPServer ( HTTP server ), SimpleHTTPServer ( HTTP server ) and CGIHTTPServer ( HTTP server ) modules.