This course is best followed if you can reproduce the examples and tutorials provided with it. To do so, you will need to install in your machine a series of software packages. These are all open-source and available for free to download. Although there are several ways to approach this process, we first enumerate here the list of dependencies and then show two simple approaches to install them in different platforms.
This guide assumes you have the following additional files, available to download by clicking:
install_gds_stack(Windows,
Mac/Linux)check_gds_stack.ipynb[Required for the Vagrant option only]
The easiest way to install locally and natively the software stack required is to install a full scientific Python distribution. Although other good alternatives are also available (e.g. Canopy, Sage), we recommend to install Anaconda. Make sure, whichever option you pick, to install a Python 2 version, not Python 3. Please follow the instructions provided in the link for installation.
Once you have a fully working Anaconda distribution installed in your
computer, you can setup an isolated environment that contains all the required
libraries by simply running the install script install_gds_stack_X.sh provided
with this guide. These scripts vary depending on the platform you are on.
Open up a terminal (OSX and Linux) or Powershell/cmd prompt (Windows) and run the following commands:
Navigate to the folder where this file is:
cd /path/to/folder/
Make the script executable:
chmod +x install_gds_stack_unix.sh
Execute the script:
./install_gds_stack_unix.sh
Once this has run, you should be able to activate the environment:
Simply double-click on the install_gds_stack_win.bat file, this will set in
motion the installation.
NOTE: this assumes your Anaconda installation is placed under
C:\Anaconda. If this is not the case, you will have to open the file with a
text editor and replace the path in line 8, where it says:
set ANACONDA_DIR=C:\Anaconda
By the path to your Anaconda installation.
[NOTE: this is an alternative to the option above]
A fully automated and reproducible approach, albeit less “native” to the local machine, is to install a virtual machine using Vagrant. This takes some of the complexities away, but requires an extra layer of computation on top of the native OS. To install a virtual machine, follow these steps:
Install Vagrant from the following link:
Navigate to the folder where this file is:
cd /path/to/folder/
Run the following command. The first time, it will take a relative long time as it has to download a large amount of files (in fact, an entirely OS). Subsequent times, it is much faster:
vagrant up
To check things have installed correctly, an additional file is included, check_gds_stack.ipynb. To run it, open a terminal (PowerShell), navigate to the folder as showed above and type the following in Mac/Linux:
source activate gds
Or the following for Windows:
activate gds
And then:
jupyter notebook
This will open up a browser window with a list of the files in the folder. Click on check_gds_stack.ipynb, which will open a new tab in the browser. Navigate to Cell –> Run All and click on it. If you do not get any error and all the cells except for the first one return True, things went well.