University of Southampton, Hans Fangohr, Computational Methods

Download and install Python software

as used in teaching Computing at the University of Southampton. Here we provide the same version of Python as is provided by on the ISS work stations. You find the installation instructions below.

Table of Contents

Download and install Python software
Testing your installation

Recommended method for MS Windows, OS X and Linux

Download the free Enthought Canopy environment from https://www.enthought.com/downloads/

and install on your machine following instructions.

You don't need to understand this when installing, but the tools provided through the free Canopy include a graphical user interface, IPython, numpy, scipy and matplotlib.

Install Symbolic Python

Go to the package manager, for example through the menu Tools -> Package Manager.
Type sympy into the search box and press return
Click on the green flag reading FREE that should have appeared in the search results window, then on Install (or straight on Install if immediately visible).

Test that the installation has worked: go to the python prompt and type import sympy. If you do not get an error message, then all is well.

Try this for fun:

import sympy
x, y = sympy.symbols('xy')
b = x + x + 10 * y
print(b)
c = sympy.integrate(b, x)
print(c)
sympy.pretty_print(c)
sympy.integrate(x * x, (x, 0, 2))

Install py.test

Go to the package manager, for example through the menu Tools -> Package Manager.
Type pytest into the search box and press return
Click on the blue(ish) flag reading COMMUNITY that should have appeared in the search results window, then on Subscribe (or straight on Subscribe if immediately visible).

If you have not got an Enthought Canopy account yet, you may have to create one (see Create an Enthought Canopy account).

You may be logged in immediately after you create your Canopy account, or you may have to click on login (if visible).
If this method fails, you can use the easy_install methods as outlined in Installing py.test with easy_install.

Create an Enthought Canopy account

To install a community package (or a canopy package), you need to create an Enthought Canopy account. Just having the account gives access to the Community packages. If you are a student at Southampton University, you can upgrade this account by obtaining an academic license, and this gives you access to the Canopy packages

The canopy software may have taken you to the sign up page. If not, then the current URL for this seems to be https://enthought.com/accounts/register/ (25 April 2013), so you can go there directly.

You need to provide your details, email address (we recommend you use the university email address here) and a password of your choice.

Once you complete this, you will receive an email with a link you need to click to activate this subscription (this is to check that the provided email address is actually yours).

Once activated, your canopy account gives you access to the packages labelled ''Community packages''. You will then be using Canopy Express (which is the free version of Canopy). If you care, you can request an academic license which gives you free usage of all the Canopy packages as well. We will not need them for the first year Python teaching.

Installing py.test with easy_install

The py.test tool helps to run regression tests automatically.

Open a shell:
- On Windows: Start MS Prompt (i.e. executed CMD.exe)
- On Mac OS X: Start the 'Terminal' Application (in /Applications/Utilities)
- On Linux: Open a bash, or xterm or other shell.
Check whether py.test is installed already?

You can check for this by attempting to execute it (and asking it for the installed version) using this command:
```
C:\Users\fangohr>py.test --version
```
If the response is:
```
'py.test' is not recognized as an internal or external command,
operable program or batch file.
```
then you need to install the pytest package:
Installing the pytest package: Execute in the command prompt:
```
C:\Users\fangohr>easy_install pytest
```
For this to work, you need to have the easy_install executable installed already (which is provided by Enthought Python and most other Python distributions), and you need to have a connection to the Internet.

You can check whether the installation has been successful by running:
```
C:\Users\fangohr>py.test --version
```
If py.test can be found, it will report something like:
```
This is py.test version 2.3.4, imported from ...
```
[The version number may differ]

Python on Linux

There is the Enthought Canopy software available for Linux.

However, you may prefer to install the required Python and Python module yourself as Python itself is usually already part of a Linux distribution and very simple to install.

You will need the following programs and libararies:

Python 2.7

numpy

scipy

matplotlib (this is 'pylab')

sympy [for symbolic operations]

On Debian Linux ('etch') and other Debian-based distributions (such as Ubuntu and Knoppix, for example), you would install these packages with the following command (as root):

aptitude install python2.7 python2.7-numpy
  python2.7-scipy python2.7-matplotlib python-visual

On Debian Linux 5.0 ('lenny'), use:

aptitude install python python-numpy
  python-scipy python-matplotlib python-visual python-sympy

The Debian package that provides py.test is called python-codespeak-lib, so run:

aptitude install  python-codespeak-lib

to install py.test.

On other distributions (SuSE, RedHat, Gentoo, ...) other commands may be necessary to install the required packages.

Testing your installation

We have provided a number of test programs in this directory (which are applicable to all operating systems). You can download these program one after the other and run them on your machine to see whether the software is installed okay. (This is usually not necessary and may only be useful if you run into difficulties.)

Return to Top