Code Development Guidelines

This document describes the coding requirements and guidelines to be followed during the development of PlasmaPy and affiliated packages.

Code written for PlasmaPy must be compatible with Python 3.6 and later. Python 2 is not supported by PlasmaPy.

PlasmaPy requires

  • Python 3.6 or later
  • Astropy 2.0 or later
  • NumPy 1.13 or later
  • SciPy 0.19 or later
  • matplotlib 2.0 or later
  • Cython
  • mpmath 1.0 or later
  • lmfit 0.9.7 or later

Obtaining PlasmaPy source code

After creating your GitHub account, go to the [main repository]( and fork a copy of PlasmaPy to your account.

To access Git commands on Windows, try Git Bash.

Next you must clone your fork to your computer. Go to the directory that will host your PlasmaPy directory, and run one of the following commands (after changing your-username to your username). If you would like to use HTTPS (which is the default and easier to set up), then run:

git clone

SSH is a more secure option, but requires you to [set up an SSH key]( beforehand. The equivalent SSH command is:

git clone
cd PlasmaPy

After cloning, we must tell git where the development version of PlasmaPy is by running:

git remote add upstream git://

To check on which remotes exist, run git remote -v. You should get something like this:

origin (fetch)
origin (push)
upstream (fetch)
upstream (push)

Setting up an environment for development

Setup procedures for the two most popular virtual environments, conda and virtualenv, are listed below.


To set up a development environment for PlasmaPy, the Anaconda distribution is strongly recommended.

After installing Anaconda, launch any conda environment. By default, conda installs a root environment, which you should be able to activate via

source /home/user/anaconda3/bin/activate root

where /home/user/anaconda3/ can be swapped to wherever your anaconda installation was resides.

On Windows, the way to do this is via running Anaconda Prompt from the Start Menu. Git Bash may also work if you have added Anaconda to PATH.

Afterwards, enter PlasmaPy’s repository root directory and execute the following:

conda env create -f requirements/environment.yml

You may now enter the environment via

source activate plasmapy

On Windows, skip the source part of the previous command.


Create a directory for holding the PlasmaPy repository, move into it and create the virtual environment

virtualenv -p python3 .

You may need to make sure that this directory’s path doesn’t contain any spaces, otherwise virtualenv may throw an error.

Your virtual environment should now be created. If you run ls you will notice that virtualenv has created a number of subdirectories: bin/, lib/, and include/. This is why we’re not creating the virtualenv within the repository itself - so as to not pollute it. To activate the virtualenv you will run:

source ./bin/activate

You should now see that your shell session is prepended with (plasmapy), like so:

(plasmapy) user@name:~/programming/plasmapy$

This indicates that the virtualenv is running. Congratulations! When your’re done working on PlasmaPy, you can deactivate the virtualenv by running

source deactivate

Now that you have plasmapy on your local computer and you have a virtual environment, you will want to “install” this development version of PlasmaPy along with its dependencies. Start by activating your virtual environment. Next you want install the PlasmaPy dependencies. One way to do this is to do

(plasmapy) user@name:~/programming/plasmapy$ pip install -r requirements/environment.txt

Next, setup the development version of PlasmaPy which you just cloned by moving into the root directory of the cloned repo and running the script there:

(plasmapy) user@name:~/programming/plasmapy/PlasmaPy$ pip install -e .

You should now be all set to run development versions of PlasmaPy modules via import PlasmaPy in your test scripts!

Running anaconda with virtualenv

If you are running the Anaconda suite and want to use virtualenv to setup your virtual environment, you will have to let the system know where the Python interpreter can be found. On Linux this is done with (for example, assuming having installed Anaconda into /anaconda3):

export LD_LIBRARY_PATH="$HOME/anaconda3/lib/"

Exporting the library path to the dynamic linker will only last for the duration of the current shell session.

You will have to add the python library directory to LD_LIBRARY_PATH, as described in a previous step, prior to activating the virtualenv for every new shell session.

Installing your own dev version

To be able to import PlasmaPy from your source version:

pip install -e {plasmapy-repository-root}

Where {plasmapy-repository-root} is the directory resulting from git clone.

If you are not working within a virtual environment, this may end in a permission error - this can be avoided via also adding the --user flag.

Coding Style

  • PlasmaPy follows the PEP8 Style Guide for Python Code. This style choice helps ensure that the code will be consistent and readable.
    • The PEP 8 Speaks integration on GitHub will comment when there are any departures from the PEP 8 style guide.
    • PEP 8 compliance may be checked locally using [pycodestyle](
    • Line lengths should be chosen to maximize the readability and elegance of the code. The maximum line length for Python code in PlasmaPy is 99 characters.
    • Docstrings and comments should generally be limited to 72 characters.
  • Follow the existing coding style within a subpackage.
  • Use standard abbreviations for imported packages when possible, such as import numpy as np, import matplotlib as mpl, import matplotlib.pyplot as plt, and import astropy.units as u.
  • files for modules should not contain any significant implementation code, but it can contain a docstring describing the module and code related to importing the module. Any substantial functionality should be put into a separate file.
  • Use absolute imports, such as from plasmapy.mathematics import Fermi_integral, rather than relative imports such as from ..mathematics import Fermi_integral.
  • For multiline imports, instead of using \ to break lines, wrap the imported names in () parentheses and use consistent whitespace.
  • There should be at most one pun per 1284 lines of code.

Branches, commits, and pull requests

Before making any changes, it is prudent to update your local repository with the most recent changes from the development repository:

git fetch upstream

Changes to PlasmaPy should be made using branches. It is usually best to avoid making changes on your master branch so that it can be kept consistent with the upstream repository. Instead we can create a new branch for the specific feature that you would like to work on:

git branch *your-new-feature*

Descriptive branch names such as grad-shafranov or adding-eigenfunction-poetry are helpful, while vague names like edits are considered harmful. After creating your branch locally, let your fork of PlasmaPy know about it by running:

git push --set-upstream origin *your-new-feature*

It is also useful to configure git so that only the branch you are working on gets pushed to GitHub:

git config --global push.default simple

Once you have set up your fork and created a branch, you are ready to make edits to PlasmaPy. Switch to your new branch by running:

git checkout *your-new-feature*

Go ahead and modify files with your favorite text editor. Be sure to include tests and documentation with any new functionality. We recommend reading about best practices for scientific computing. PlasmaPy uses the PEP 8 style guide for Python code and the numpydoc format for docstrings to maintain consistency and readability. New contributors should not worry too much about precisely matching these styles when first submitting a pull request, as the PEP8 Speaks GitHub integration will check pull requests for PEP 8 compatibility, and further changes to the style can be suggested during code review.

You may periodically commit changes to your branch by running

git add
git commit -m "*brief description of changes*"

Committed changes may be pushed to the corresponding branch on your GitHub fork of PlasmaPy using

git push origin *your-new-feature*

or, more simply,

git push

Once you have completed your changes and pushed them to the branch on GitHub, you are ready to make a pull request. Go to your fork of PlasmaPy in GitHub. Select “Compare and pull request”. Add a descriptive title and some details about your changes. Then select “Create pull request”. Other contributors will then have a chance to review the code and offer contructive suggestions. You can continue to edit the pull request by changing the corresponding branch on your PlasmaPy fork on GitHub. After a pull request is merged into the code, you may delete the branch you created for that pull request.

Commit Messages

Good commit messages communicate context and intention to other developers and to our future selves. They provide insight into why we chose a particular implementation, and help us avoid past mistakes.

Suggestions on how to write a git commit message:

  • Separate subject from body with a blank line
  • Limit the subject line to 50 characters
  • Capitalize the subject line
  • Do not end the subject line with a period
  • Use the imperative mood in the subject line
  • Wrap the body at 72 characters
  • Use the body to explain what and why vs. how


  • All public classes, methods, and functions should have docstrings using the numpydoc format.
  • Docstrings may be checked locally using [pydocstyle](
  • These docstrings should include usage examples.

Warnings and Exceptions

  • Debugging can be intensely frustrating when problems arise and the associated error messages do not provide useful information on the source of the problem. Warnings and error messages must be helpful enough for new users to quickly understand any problems that arise.
  • “Errors should never pass silently.” Users should be notified when problems arise by either issuing a warning or raising an exception.
  • The exceptions raised by a method should be described in the method’s docstring. Documenting exceptions makes it easier for future developers to plan exception handling.


  • Code within PlasmaPy must use SI units to minimize the chance of ambiguity, and for consistency with the recognized international standard. Physical formulae and expressions should be in base SI units.
    • Functions should not accept floats when an Astropy Quantity is expected. In particular, functions should not accept floats and make the assumption that the value will be in SI units.
    • A common convention among plasma physicists is to use electron-volts (eV) as a unit of temperature. Strictly speaking, this unit corresponds not to temperature but is rather a measure of the thermal energy per particle. Code within PlasmaPy must use the kelvin (K) as the unit of temperature to avoid unnecessary ambiguity.
  • PlasmaPy uses the astropy.units package to give physical units to values.
    • All units packages available in Python presently have some limitations, including incompatibility with some NumPy and SciPy functions. These limitations are due to issues within NumPy itself. Many of these limitations are being resolved, but require upstream fixes.
  • Dimensionless units may be used when appropriate, such as for certain numerical simulations. The conventions and normalizations should be clearly described in docstrings.

Equations and Physical Formulae

  • If a quantity has several names, then the function name should be the one that provides the most physical insight into what the quantity represents. For example, gyrofrequency indicates gyration, whereas Larmor_frequency indicates that this frequency is somehow related to someone named Larmor. Similarly, using omega_ce as a function name will make the code less readable to people who are unfamiliar with this particular notation.
  • Physical formulae should be inputted without first evaluating all of the physical constants. For example, the following line of code obscures information about the physics being represented:
>>> omega_ce = 1.76e7*(B/u.G)*u.rad/u.s   
In contrast, the following line of code shows the exact formula which makes the code much more readable.
>>> omega_ce = (e * B) / (m_e * c)       
The origins of numerical coefficients in formulae should be documented.
  • Docstrings should describe the physics associated with these quantities in ways that are understandable to students who are taking their first course in plasma physics while still being useful to experienced plasma physicists.

  • SI units that were named after a person should not be capitalized except at the beginning of a sentence.

  • Some plasma parameters depend on more than one quantity with the same units. In the following line, it is difficult to discern which is the electron temperature and which is the ion temperature.

    >>> ion_sound_speed(1e6*u.K, 2e6*u.K)  

    Remembering that “explicit is better than implicit”, it is more readable and less prone to errors to write:

    >>> ion_sound_speed(T_i=1e6*u.K, T_e=2e6*u.K)    
  • SI units that were named after a person should be lower case except at the beginning of a sentence, even if their symbol is capitalized. For example, kelvin is a unit while Kelvin was a scientist.

Angular Frequencies

Unit conversions involving angles must be treated with care. Angles are dimensionless but do have units. Angular velocity is often given in units of radians per second, though dimensionally this is equivalent to inverse seconds. Astropy will treat radians dimensionlessly when using the dimensionless_angles equivalency, but dimensionless_angles does not account for the multiplicative factor of 2*pi that is used when converting between frequency (1 / s) and angular frequency (rad / s). An explicit way to do this conversion is to set up an equivalency between cycles/s and Hz:

>>> from astropy import units as u
>>> f_ce =, equivalencies=[(, u.Hz)])   

However, dimensionless_angles does work when dividing a velocity by an angular frequency to get a length scale:

>>> d_i = (c/omega_pi).to(u.m, equivalencies=u.dimensionless_angles())