Quansight Labs 2024 Internship #713
Description
🙌🏼 Mentor(s): @hameerabbasi, @mtsokol
🧰 Skills needed: Basic Python/NumPy knowledge, Writing/organisation skills, basic scripting
🏷 Project description
Overhaul the sparse documentation, and as a stretch goal, perform some maintenance tasks.
🚀 Expected outcomes
In order of priority
- Convert docs to Material for MkDocs
- Expand documentation:
- Add docstrings to missing objects
- Use the documentation system proposed by Divio
- Find more examples like SDDMM to use in documentation, where we can get performance gains from fusion
- GitHub Labels for Issues/PRs
- Come up with a labelling system for issues/PRs
- Come up with issue/PR templates
- Implement that system
- (Stretch Goal) Use an auto-label bot of some sort
- (Stretch goal) Benchmarks
- Preparing real world examples for benchmarks
- Instead of generating random input with rand find some real dataset with reasonable size that can be used for benchmarking
- (Stretch goal) Expand Array API standard support
- Numba
- Finch
- (Stretch goal) Expand code coverage
- Numba now supports recording of coverage -- so find a way to track that during tests.
At the end of the project, the intern will have knowledge of sparse usage, Materials for MkDocs usage and setup, and optionally some real-world usage of sparse and the benefits and a rough understanding of fusion
🔗 Useful links
- Material for MkDocs: https://squidfunk.github.io/mkdocs-material/
- API Documentation Plugin: https://mkdocstrings.github.io
- Recipes to auto-generate docs from new objects: https://mkdocstrings.github.io/recipes/
- Documentation guidelines and types: https://docs.divio.com/documentation-system/
- Find more examples like SDDMM to use in documentation, where we can get performance gains
- Airspeed Velocity: https://asv.readthedocs.io/en/stable/using.html
- Array API standard: https://data-apis.org/array-api/latest/