Skip to content
Noah C. Benson edited this page Sep 21, 2018 · 3 revisions

Basic installation of neuropythy can be accomplished with pip or via the github archive. Both of these should automatically install the core dependencies of neuropythy. Additional dependencies are recommended for plotting, registration, and Human Connectome Project integration. These are documented below.

Make sure to read the page on configuring neuropythy once you've installed it.

Installation via Pip

# install neuropythy itself
pip install neuropythy
# install the optional dependencies
pip install matplotlib ipyvolume s3fs h5py

Note: If you are a conda user, you may want to use conda install matplotlib ipyvolume s3fs h5py for the dependencies (you may want to install neuropythy's core dependencies this way also).

Installation via Github

# Check out neuropythy
git clone https://github.com/noahbenson/neuropythy
# Install it
cd neuropythy
python setup.py install

Dependencies

The following software packages are not required for neuropythy's core features, but are required for certain functions or modules to work.

  • s3fs ≥ 0.1.5. The HCP dataset can be accessed automatically using neuropythy's hcp_subject() function. If configured correctly (see below), neuropythy will silently download the relevant HCP data from its Amazon S3 bucket as it is requested. Doing this requires the s3fs library.
  • h5py ≥ 2.8.0. The h5py file is used to import the HCP retinotopy data if it is found or configured for automatic-downloading (see below).
  • Java ≥ 1.8. The registration algorithm employed by the register_retinotopy command is performed by a Java library embedded in the neuropythy Python library. This library is the nben library, and is included as a submodule of this GitHub repository, found in neuropythy/lib/nben; a standalone jar-file is also distributed as part of the PyPI neuropythy distribution. However, in order for the Py4j library, which allows Python to execute Java routines, to use this jar-file, you must have a working version of Java installed; accordingly, the register_retinotopy command is only available if you have Java installed and working. For help getting Java configured to work with Py4j, see the Py4j installation page.
  • Jupyter. Jupyter is a Python notebook-based REPL environment. Neuropythy can interact with Jupyter to produce 2D and 3D graphical representations of neuro-imaging data, particularly the cortical surface meshes.
  • matplotlib ≥ 1.5.3. A few functions for plotting cortical maps are defined in the neuropythy.graphics package. These are not defined if matplotlib is not imported successfully. The primary interface to this functionality is the neuropythy.cortex_plot as well as some helper functions and colormaps.
  • ipyvolume ≥ 0.5.1. If you wish to make 3D graphics plots, you will need to install and use the ipyvolume library. The neuropythy function neuropythy.cortex_plot will handle most of the details, assuming you have ipyvolume installed.
Clone this wiki locally