splinepy welcomes and appreciates discussions, issues and pull requests!
Once the repo is forked, one possible starting point would be creating a new python environments, for example, using conda with python=3.10
conda create -n splinepyenv python=3.10
conda activate splinepyenv
git clone [email protected]:<path-to-your-fork>
cd splinepy # or <forkname>
git submodule update --init --recursive
git checkout -b new-feature0
pip install -e. --config-settings=cmake.args=-DSPLINEPY_MORE=OFF --config-settings=cmake.build-type="Debug"--config-settings=cmake.args=-DSPLINEPY_MORE build argument builds splines up to 3D (both parametric and physical dimensions if they are part of template parameters), and that way we can reduce compile time. --config-settings=cmake.build-type="Debug" build also reduces compile time. We are experimenting with ways to reduce compile time during development. Let us know if you have a great idea!
- use
ifandraiseinstead ofassert - no complex comphrehensions: preferably fits in a line, 2 lines max if it is totally necessary
- use first letter abbreviations in element loops:
for kv in knot_vectors - use
i,j,k,lfor pure index:for i, kv in enumerate(knot_vectors) - module local import with a leading underscore:
from splinepy import settings as _settings
For c++, we've prepared a .clang-format, with that you can just run clang-format. We closely follow naming scheme suggested by google stype guide, with a clear exception of file naming.
Here's a preference:
#pragma once
To check the format and style of your code use the following commands:
pip install pre-commit
precommit run -apip install -r docs/requirements.txt
sphinx-build -b html docs/source docs/build -E -j auto # `-E` to ignore existing files, -j auto for parallel build (or specify number of processes)
# you can now open `docs/build/index.html` with your browser