We are providing instructions to install the jupyter notebook and other python packages that we'll need for the course problem sets in your own computer, without a VM. This instructions will be modified in the future, as we will need more packages for other future assignments.
Pease keep in mind, though, that the official way of doing the problem sets is using the VM, and that we cannot support everyone trying to install the software in their own computers. Unfortunately, we don't have the resources to fix all the problems that people will encounter with different operating systems and hardware if following these instructions. You should only do this if you are comfortable installing complex software.
First, install miniconda (use the python2.7 version). It's a lightweight multiplatform Python distribution that can download binaries for libraries for all platforms, making installations of complicated libraries (numpy, matplotlib, scipy) extremely easy and fast. Basically, follow the instructions here or just download and execute the bash script from the first link.
Make sure to update your bash script to have your miniconda path be prepended to your PATH. If you've done this properly, executing python should read something like:
Python 2.7.10 |Continuum Analytics, Inc.|
(maybe with a different version number, but look for the Continuum Analytics thing)
Another nice thing about the miniconda system is that it installs everything into a user folder. That means that you don't need to use sudo for any install command, and that you can easily have multiple python installations with their own libraries if you need to.
Once miniconda is installed, install the required libraries with the following command:
conda install --yes jupyter ipython scipy numpy matplotlib seaborn networkx nose shapely pyyaml requests
Finally, install these two other libraries that can be installed with pip and are not in conda:
pip install pydot-ng nbgrader descartes
Again note that you don't need to use sudo.
That should give you everything you need for these problem sets, and it should be pretty fast since its downloading binaries and nothing needs to be compiled.
Since you are not using the VM, you'll need to obtain the problem sets from our repositories directly.
For that, you'll need an MIT Github account. Then you'll need to email the TA (me!) and ask to be added to the course repository. Again, we don't expect most people to be doing this.
You'll then have access to the problem set repositories. For the first problem set, for example, the repository is located at:
You can then clone this repository everywhere you want in your computer. If you don't know how to clone a git repository, you probably should be using the VM…
Navigate to the folder you have cloned in the terminal and open the Jupyter notebook with the following command:
Finally, follow the same instructions as everyone else to submit your problem set (compress the whole folder and upload it to Stellar).
P.S.: A word of warning: in the second half of the course we'll be using a Python package called Numberjack, which is known to not like Windows.