Installing PyNWB for Developers

PyNWB has the following minimum requirements, which must be installed before you can get started using PyNWB.

  1. Python 3.7, 3.8, 3.9, or 3.10

  2. pip

Set up a virtual environment

For development, we recommend installing PyNWB in a virtual environment in editable mode. You can use the virtualenv tool to create a new virtual environment. Or you can use the conda package and environment management system for managing virtual environments.

Option 1: Using virtualenv

First, install the latest version of the virtualenv tool and use it to create a new virtual environment. This virtual environment will be stored in the venv directory in the current directory.

pip install -U virtualenv
virtualenv venv

On macOS or Linux, run the following to activate your new virtual environment:

source venv/bin/activate

On Windows, run the following to activate your new virtual environment:

venv\Scripts\activate

This virtual environment is a space where you can install Python packages that are isolated from other virtual environments. This is especially useful when working on multiple Python projects that have different package requirements and for testing Python code with different sets of installed packages or versions of Python.

Activate your newly created virtual environment using the above command whenever you want to work on HDMF. You can also deactivate it using the deactivate command to return to the base environment.

Option 2: Using conda

First, install Anaconda to install the conda tool. Then create and activate a new virtual environment called “venv” with Python 3.8 installed.

conda create --name venv python=3.8
conda activate venv

Similar to a virtual environment created with virtualenv, a conda environment is a space where you can install Python packages that are isolated from other virtual environments. In general, you should use conda install instead of pip install to install packages in a conda environment.

Activate your newly created virtual environment using the above command whenever you want to work on HDMF. You can also deactivate it using the conda deactivate command to return to the base environment.

Install from Git repository

After you have created and activated a virtual environment, clone the PyNWB git repository from GitHub, install the package requirements using the pip Python package manager, and install PyNWB in editable mode.

$ git clone --recurse-submodules git@github.com:NeurodataWithoutBorders/pynwb.git
$ cd pynwb
$ pip install -r requirements.txt
$ pip install -e .

Run tests

For running the tests, it is required to install the development requirements. Again, first activate your virtualenv or conda environment.

$ git clone --recurse-submodules git@github.com:NeurodataWithoutBorders/pynwb.git
$ cd pynwb
$ pip install -r requirements.txt -r requirements-dev.txt
$ pip install -e .
$ tox

For debugging it can be useful to keep the intermediate NWB files created by the tests. To keep these files create the environment variables CLEAN_NWB/CLEAN_HDMF and set them to 1.

Following PyNWB Style Guide

Before you create a Pull Request, make sure you are following PyNWB style guide (PEP8). To do that simply run the following command in the project’s root directory.

$ flake8

FAQ

  1. I am using a git cloned copy of PyNWB and getting the error: RuntimeError: Unable to load a TypeMap - no namespace file found

    • The PyNWB repo uses git submodules that have to be checked out when cloning the repos. Please make sure you are using the --recurse-submodules flag when running git clone:

      git clone --recurse-submodules git@github.com:NeurodataWithoutBorders/pynwb.git
      

      You can also run the following on your existing cloned repo.

      git submodule init
      git submodule update --checkout --force
      
  2. I did a git pull but I’m getting errors that some neurodata_type does not exist.

    • The PyNWB repo uses git submodules that have to be updated as well. Please make sure you are using the git pull --recurse-submodules