First things first

Image by David Whittaker

What is SciPy?

SciPy is a set of fundamental algorithms for scientific computing in Python. It includes:

• numerical integration
• interpolation
• signal processing
• linear algebra, including eigenvalue problems
• algebraic and differential equations
• statistics
• mathematical optimization

It is based on NumPy arrays, but also includes sparse matrices and k-dimensional trees manipulations.

The history of SciPy

• Initial release in 2001
• Over 600 unique code contributors
• Thousands of dependent packages
• Over 100,000 dependent repositories
• Millions of downloads per year

How is it organized?

• SciPy wraps highly-optimized implementations written in low-level languages like Fortran, C, and C++. (link)

• Code and documentation live in the same repository on github
• Website is developed in a separate repository (https://github.com/scipy/scipy.org)
• Teams structure https://scipy.org/teams/
• Meetings https://scientific-python.org/calendars/

Let's look at the code!

https://github.com/scipy/scipy

How can you contribute?

https://scipy.org/contribute/

• Writing code
• Reviewing Pull Requests
  • Summarize a long-running discussion on an issue or PR
• Triage issues and PRs
  • Check if older bugs are still present
  • Find duplicate issues and link related ones
  • Add good self-contained reproducers to issues
• Test proposed changes
• Developing educational materials
  • We’d also welcome worked examples, notebooks, and videos.
• Website development
• Graphic design
• Fundraising (SciPy 2019 talk)

Setting up your development environment

Start here: Contributor quickstart guide

Alternatively: Using GitPod

Start here: GitPod for SciPy development guide

What file should I edit?

• Generated/compiled files
  • A .pyx file is compiled by Cython to a .c file.
  • A .pxd file is a header for Cython code
  • A .pyf.src file is a signature file used to wrap a Fortran file with numpy.f2py
• _add_newdocs.py
• Narrative documentation

grep is your friend!

Building to see your changes

python dev.py

or

python dev.py --build-only

Building the documentation

python dev.py --doc

Submitting your Pull Request

• Commit messages
  • There are some guidelines for writing commit messages
• Pull request reviews
  • Reviews should be constructive and friendly; back-and-forths are expected and we can help you complete your PR!
• CI
  • An example (PR#16258)
  • It is not possible in general to run all these tests locally, so CI can catch it on GitHub.

Final thoughts

• https://scipy.org/community/
  • Mailing list
• https://scientific-python.org/calendars
  • Community and newcomer meetings
• Good first issues

Obrigada!

@melissawm