Feedback and Contribution#

We welcome any contribution via ROSS issue tracker. These include bug reports, problems on the documentation, feedback, enhancement proposals etc. You can use the repository Discussions section for questions and further information.

Code style: Black#

To format our code we use Black, which is the “uncompromising Python code formatter”. You can configure your development environment to use Black before a commit. More information on how to set this is given at Black’s documentation.

We also recommend using the pre-commit tool so that black is automatically run when doing a commit.

How to contribute to ROSS using git#

Git is a version control system (VCS) for tracking changes in code during software development. To download the ROSS source code and contribute to its development, you need Git installed in your machine. Refer to the Git website and follow the instructions to download and install it. Once you have Git installed, you will be able to follow the instructions in [How to contribute to ROSS using git], which explains how to download and contribute to ROSS.

To use git to contribute to ROSS project, follow the steps below: For Windows users: commands provided here can be executed using Git Bash instead of Git GUI.

Step 1: Make your copy (fork) of ROSS#

Go to petrobras/ross In the top-right corner of the page, click Fork, to fork it to your GitHub account.

From the command line:

git clone
cd ross
git remote add upstream

Step 2: Keep in sync with changes in ROSS#

Setup your local repository, so it pulls from upstream by default:

git config branch.main.remote upstream
git config branch.main.merge refs/heads/main

This can also be done by editing the config file inside your ross/.git directory. It should look like this:

        repositoryformatversion = 0
        filemode = true
        bare = false
        logallrefupdates = true
[remote "origin"]
        url =
        fetch = +refs/heads/*:refs/remotes/origin/*
[remote "upstream"]
        url =
        fetch = +refs/heads/*:refs/remotes/upstream/*
        fetch = +refs/pull/*/head:refs/remotes/upstream/pr/*
[branch "main"]
        remote = origin
        merge = refs/heads/main

The part fetch = +refs/pull/*/head:refs/remotes/upstream/pr/* will make pull requests available in your local repository after a git fetch.

For example, assuming $ID is the pull request number and $BRANCHNAME is the name of the new local branch you wish to create:

git fetch upstream pull/$ID/head:$BRANCHNAME

Switch to the newly created branch:

git switch $BRANCHNAME

Step 3: Set up development environment#

To set up a development environment you can create a conda environment:

conda create -n rs
conda activate rs

or a virtualenv:

python3 -m venv env
. env/bin/activate
# or "env\Scripts\activate" on Windows

and then install ROSS in editable mode with development dependencies:

pip install -e ".[dev]"

Step 4: Make a new feature branch#

git fetch upstream
git checkout -b my-new-feature upstream/main

Step 5: Testing the code#

We use pytest to test the code. Unit tests are placed in the ~/ross/ross/tests folder. We also test our docstrings to assure that the examples are working. If you want to run all the tests you can do it with (from the ~/ross/ross folder):


Code is only merged to main if tests pass. This is checked by services GitHub Actions, so make sure tests are passing before pushing your code to github.

Step 6: Push changes to your git repository#

After a complete working set of related changes are made:

git add modified_file
git commit
git push origin my-new-feature

The following blog posts have some good information on how to write commit messages:

A Note About Git Commit Messages

On commit messages

Step 7: Push changes to the main repo#

To create a Pull Request (PR), refer to the github PR guide.

Docstrings for class and methods#

A new method must have a docstring presenting a summary for what the method does. ROSS’ docstrings follows the Numpy docstring style. It’s important to follow the Numpy’s template due to the formatting that will be presented on the ROSS website.

Example of docstring:

def foo(arg1, arg2, arg3):
"""Title (First line should be in imperative mood and end with a period)

A brief explanation of what this method does. (Optional)

Parameters (if the method receives any arguments)
arg1 : TYPE
arg2 : TYPE
arg3 : TYPE

References (if applicable)
.. bibliography:: ../../../docs/refs.bib

Raises (if there's any error message raised)

Returns (if the method return something)
result : TYPE

Examples (if applicable)
>>> a = 1
>>> b = 2
>>> c = 3
>>> s = foo(a, b, c)
result = arg1 + arg2 + arg3
return result

It is possible to add other sections in addition to those previously presented (e.g. Notes, See Also, Warnings…). Just follow the same rules and it’s good to go.

When creating examples, be aware of code lines that return any result from a method or class. The example output must match what the method returns because GitHub Actions (the CI that runs tests for ROSS) checks the examples and raise errors, if the example output does not match the actual output.

Sometimes, it’s not possible to represent all the output (e.g. a figure, a large matrix, etc), so it’s recommended to use the comment # doctest: +ELLIPSIS, and then, truncate the function output with a ..., and add this comment beside the command line.


from bokeh.plotting import figure

def foo():
    """Plot a bokeh figure.

    figure : bokeh.figure
        A figure.

    >>> figure = foo()
    >>> figure # doctest: +ELLIPSIS
    fig = figure()
    fig.line([1, 2, 3], [1, 2, 3])

    return fig


We use sphinx to generate the project’s documentation. We keep the source files at ~/ross/docs, and the website is hosted here. The website tracks the documentation for the released version with the ‘Docs’ GitHub Action.

If you want to test the documentation locally:

Go to the ~/ross/docs folder and run:

make html

Optionally, if you don’t want run all notebooks you can use:

make EXECUTE_NOTEBOOKS='off' html

After building the docs, go to the _build/html directory (~/ross/docs/_build/html) and start a python http server:

python -m http.server

After that you can access your local server ( and see the generated docs.

Making new releases#

To make a new release we need only to create a tag using git and push to GitHub:

git tag <version number>
git push upstream --tags

Pushing the new tag to the GitHub repository will start a new build on GitHub actions. If all the tests succeed, GitHub will upload the new package to PyPI (see the deploy command on .github/workflows/publish-to-pypi.yml).

It is recommended to first use release candidates version (e.g. v1.1.2rc1). These will only be installed with:

pip install --pre ross-rotordynamics

and it is usefull to test the installation process before the final release.

ROSS structure#

To explain how ROSS is structured, we will describe the following building blocks:

  • Elements: represent the physical components of the rotor (e.g. shaft, disks, bearings, etc);

  • Rotor: represent the rotor itself, which is composed by elements;

  • Results: represent the results of the simulation (e.g. displacement, velocity, etc).


Elements are the building blocks of the rotors. They are the physical components of the rotor (e.g. shaft, disks, bearings, etc). Each element has its own class, which is responsible for calculating the element’s stiffness and damping matrices, and its gyroscopic effect.

All the elements classes inherit from the Element class, which is defined in the ross/ file.

The Element class is an abstract base class, which means that classes that inherit from it must implement the methods defined in the Element class. Some of these abstract methods are:

  • M(self): returns the mass matrix of the element;

  • K(self): returns the stiffness matrix of the element;

  • C(self): returns the damping matrix of the element;

  • G(self): returns the gyroscopic matrix of the element.

  • dof_mapping(self): returns the degree of freedom mapping of the element.

  • _patch(self): returns a plotly.graph_objects.Figure that will be used in the rotor plot.

If a new element is created, it must inherit from the Element class and implement the methods described above. With that, the element will be compatible with the rest of the code.


The Rotor class is defined in the ross/ file. It is responsible for assembling the rotor, which means that it will assemble the stiffness, damping and gyroscopic matrices of the rotor.

After having a Rotor object, it is possible to run different analysis which are available as methods with the prefix .run_.


The Results class is defined in the ross/ file. It is responsible for storing the results of each analysis executed with a .run_ method. A Results object will also have some methods to plot the results that are stored in the object.

Coding conventions#

Save and load data#

Each element or rotor has a save method that saves the object in a .toml file.

The load method is a class method that loads the object from a .toml file.

Here is a code example for saving and loading a rotor:

import ross as rs

# create a rotor
rotor = rs.rotor_example()

# save the rotor"rotor.toml")

# load the rotor using the class method
rotor_loaded = rs.Rotor.load("rotor.toml")

Additionally, if the calculation of a specific object is expensive, we implement a .run method that will check if the object was already calculated and saved in the .toml file.

This way we avoid recalculating when loading the object.