Recommended additional software

The following software is not required for the basic operation of PySB, but provides extra capabilities and features when installed.

• cython or weave Cython is a package for compiling Python code into C code on the fly. It is used by PySB's ScipyOdeSimulator to greatly improve the speed of ODE integration. PySB will detect and use Cython automatically, if available. To install with Anaconda, type conda install cython. With pip, type pip install cython.

  Weave performs the same job as Cython, and is slightly faster in some circumstances. It is only available on Python 2. To install with Anaconda, type conda install -c conda-forge weave. With pip, type pip install weave.

• matplotlib

  This Python package allows you to plot the results of your simulations. It is not a hard requirement of PySB but many of the example scripts use it. matplotlib is included with Anaconda. Otherwise, it can be installed with pip install matplotlib.

• pandas

  This Python package provides extra capabilities for examining large numerical datasets, with statistical summaries and database-like manipulation capabilities. It is not a hard requirement of PySB, but it is a useful addition, particularly with large sets of simulation results. pandas is included with Anaconda. Otherwise, it can be installed with pip install pandas.

• IPython

  An alternate interactive Python shell, much improved over the standard one. IPython is included with Anaconda. Otherwise, it can be installed with pip install ipython.

• Kappa 4.0

  Kappa is a rule-based modeling tool that can produce several useful model visualizations or perform an agent-based model simulation. PySB optionally interfaces with its KaSim simulator and KaSa static analyzer.

  To install Kappa for PySB use, put the KaSim executable (and optionally KaSa if you have it) in /usr/local/share/KaSim (Mac or Linux) or C:\\Program Files\\KaSim (Windows). If you would like to put it somewhere else, set the KAPPAPATH environment variable to the full path to the folder containing the KaSim and KaSa executables. Note that if you have downloaded the official binary build of KaSim, it will be named something like KaSim_4.0_winxp.exe or KaSim_4.0_mac_OSX_10.10. Regardless of where you install it, you will need to rename the file to strip out the version and operating system information so that you have just KaSim.exe (Windows) or KaSim (Mac or Linux).

  On Anaconda, Kappa can be installed with conda install -c alubbock kappa.

Background

Docker is a virtualization platform which encapsulates software within a container. It can be thought of like a virtual machine, only it contains just the application software (and supporting dependencies) and not a full operating system stack.

Install Docker and the PySB software stack

1. Install Docker

   To use PySB with Docker, first you'll need to install Docker, which can be obtained from https://www.docker.com/community-edition#/download (Windows and Mac). Linux users should use their package manager (e.g. apt-get).

2. Download the PySB software stack from the Docker Hub

   On the command line, this requires a single command:

   docker pull pysb/pysb

   This only needs to be done once, or when software updates are required.

3. Start the container

   Start the Docker container with the following command (on Linux, the command may need to be prefixed with sudo):

   docker run -it --rm -p 8888:8888 pysb/pysb

   This starts the PySB Docker container with Jupyter notebook and connects it to port 8888.

4. Open Jupyter Notebook in a web browser

   Open a web browser of your choice and enter the address http://localhost:8888 in the address bar. You should see a web page with the Jupyter notebook logo. Several example and tutorial notebooks are included to get you started.

Important notes for Docker installations

To see graphics from matplotlib within the Jupyter Notebook, you'll need to set the following option in your notebooks before calling any plot commands:

%matplotlib inline

Any Jupyter notebooks created will be saved in the container itself, rather than on the host computer. Notebooks can be downloaded using the Jupyter interface, or a directory on the host computer can be shared with the container.

The PySB container builds on the Jupyter SciPy notebook, which contains further information on the options available for the container (such as sharing a directory with the host computer to preserve notebooks, setting a password and more). Documentation from the Jupyter project is available at https://hub.docker.com/r/jupyter/scipy-notebook/