Google Colab and Jupyter Notebooks

If you are new to Google Colab or just need a refresher, we suggest going through Google’s introduction to Colaboratory.

Google Colab has the same functionality as a Jupyter Notebook. The main difference is that Jupyter Notebooks are run on your local machine. The Jupyter Notebook documentation details installation and use.

Updating Numpy

Google Colab is versatile and platform-independent, but some pre-installed libraries may be old versions. When running MRSimulator on Google Colab, an incompatible version of numpy will cause the following error

ValueError: numpy.ndarray size changed, which may indicate binary incompatibility. Expected 88 from C header got 80 from PyObject

To update numpy in Google Colab and execute the following command in a new cell.

!pip install -U numpy

This step may take a few seconds. After updating, a warning should pop up saying

WARNING: The following packages were previously imported in this runtime: [numpy]    You must restart the runtime to use newly installed versions.

Press the restart runtime button, and numpy should be up-to-date. Remember to re-run all code cells since all previous outputs are cleared after restarting the runtime.

Order of Cell Execution

For Jupyter Notebooks and Google Colab, the order of cell execution is important. If a variable foo is referenced before assignment, Python will throw NameError: name 'foo' is not defined.

If you get this error when working in a notebook, first check the cell where your variable was defined has been executed. A cell can be executed by pressing shift + enter while the text cursor is in that cell or by pressing the run button near the top-left of the cell.

Similarly, if you’ve reassigned a variable but the code doesn’t reflect that reassignment, check to ensure the cell where the variable was reassigned has been executed. Value changes are only recognized after cell execution.