# SP dev summit 2024 work log

- Trying to update numpydoc to be more explicit about single backticks to refer to parameters https://github.com/numpy/numpydoc/pull/525 and https://github.com/scientific-python/summit-2024/issues/22

- PRs created:
- [pydata-sphinx-theme issue backlog](https://github.com/scientific-python/summit-2024/issues/10)
- https://github.com/pydata/pydata-sphinx-theme/pull/1849
- https://github.com/pydata/pydata-sphinx-theme/pull/1850
- https://github.com/pydata/pydata-sphinx-theme/pull/1867
- https://github.com/pydata/pydata-sphinx-theme/pull/1861
- https://github.com/scientific-python/upload-nightly-action/issues/82
- [Add prototype of static "intime" analysis -- docstub#2](https://github.com/lagru/docstub/pull/2)
- [Disable ruff/pyupgrade rule UP038 -- scikit-image#7430](https://github.com/scikit-image/scikit-image/pull/7430)
- [MAINT: sparse: maxprint now in func sig for constructors](https://github.com/scipy/scipy/pull/20719)
- [Fix handling of multi-module intersphinx registries](https://github.com/sphinx-gallery/sphinx-gallery/pull/1320) on sphinx-gallery
- [Fix bug with SciPy parallel doc build](https://github.com/scipy/scipy/pull/20897) on scipy
- [CI Use full length commit SHA for PyPI publishing](https://github.com/scikit-learn/scikit-learn/pull/29180) on scikit-learn
- [Parallel example building in sphinx-gallery](https://github.com/sphinx-gallery/sphinx-gallery/pull/877)
- [Be mindful of resources by canceling in-progress workflows -- skimage#7436](https://github.com/scikit-image/scikit-image/pull/7436)
- [Recommend editable install for dev setup -- skimage#7434](https://github.com/scikit-image/scikit-image/pull/7434)
- [Add optional dependencies section -- changelist#18](https://github.com/scientific-python/changelist/pull/18)

- PRs reviewed:

- PRs merged:
- [DOC: Remove definition lists from contribution guide](https://github.com/sphinx-gallery/sphinx-gallery/pull/1318) on sphinx-gallery
- [fix PyData Sphinx Theme overview diagram](https://github.com/pydata/pydata-sphinx-theme/pull/1849)
- [Write API reference titles in monospace font](https://github.com/scipy/scipy/pull/20890/files) on SciPy
- [Expose logo_link in theme.conf](https://github.com/pydata/pydata-sphinx-theme/pull/1850) in PyData Sphinx Theme
- [re-enable parallel-write safe](https://github.com/pydata/pydata-sphinx-theme/pull/1859) in PyData Sphinx Theme
- [DOC: Add version warning banner for docs versions different from stable](https://github.com/matplotlib/matplotlib/pull/28336) on Matplotlib
- [ENH: sparse: Add axis to sparse `count_nonzero(axis)`](https://github.com/scipy/scipy/pull/20857) in scipy.sparse
- [ENH: sparse: add dtype validation in __init__ and astype](https://github.com/scipy/scipy/pull/20754)
- [CI: Use full length commit SHA for PyPI publishing](https://github.com/scikit-image/scikit-image/pull/7433) on scikit-image
- [Add MeeseeksDev / Lumberbot (App) to GitHub page](https://github.com/scientific-python/tools.scientific-python.org/pull/49) on tools.scientific-python.org
- [CI: Add GitHub artifact attestations to package distribution -- skimage#7427](https://github.com/scikit-image/scikit-image/pull/7427)

- Issues created:
- [simple security upgrade to release actions](https://github.com/mne-tools/mne-python/issues/12641)
- [convert MNE from sparse matrix to sparse array](https://github.com/mne-tools/mne-python/issues/12642)
- [Sphinx parallel build env bug](https://github.com/sphinx-doc/sphinx/issues/12409)
- [DOC: Write API reference titles in monospace font #20890](https://github.com/scipy/scipy/pull/20890)
- [Add CI test for docs build of selected Scientific Python projects
#1851](https://github.com/pydata/pydata-sphinx-theme/issues/1851)
- [Bugs with sparse Dijkstra](https://github.com/scipy/scipy/issues/20904) found while working on sparse array (rather than matrix) adoption
- [Consistent use of monospaced font](https://github.com/pydata/pydata-sphinx-theme/issues/1852)
- [Create GitHub artifact attestations directly after wheel generation -- skimage#7432](https://github.com/scikit-image/scikit-image/issues/7432)

- Issues closed:
- https://github.com/scientific-python/specs/issues/230

- Repos created:
- [marray: backend agnostic, array API compatible masked arrays](https://github.com/mdhaber/marray)

Array API discussion: https://hackmd.io/TG6Uz2whSDmGi97czbVEPA?edit

Sparse WG: https://hackmd.io/9UCjKdRUTEeoS8KdQNll-A

## Spontaneous collaborations

### Exploring awkward-array for brain connection trajectories.

More information about brain tractography methods and dataset: https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2024.1389680/abstract

Notebook with example: https://gist.github.com/arokem/167666bf4017a740a5321f97f38718d5

The example downloads data from S3 and then tries out a couple of things. The last cell is an example of the kind of computation that would be great to accelerate with awkward and/or similar libraries.

### Exploring scipp and nested-pandas for HEP data

A notebook example [in this gist](https://gist.github.com/nsmith-/7e52a21b87873180ba9f9ba69ff3ce01) explores using [scipp](https://scipp.github.io/index.html) to do a basic "hello world" data analysis using open data from the Compact Muon Solenoid (CMS) to reconstruct a [Z boson peak](https://home.cern/science/physics/z-boson) at 91 GeV, and compares/contrasts the analysis user story with that for awkward-array.

Another notebook example [in this gist](https://gist.github.com/nsmith-/bdd3135fe0bdeac7d71164894e39cbb7) demonstrates reading the same CMS open data into [nested-pandas](https://nested-pandas.readthedocs.io/en/latest/index.html), a project by the LINCC frameworks team to handle ragged arrays.

### Best practices on running user surveys

- Review of the sk-learn user survey: https://hackmd.io/HAr8vBTPT_mncKC9u6z8Ig?both
https://hackmd.io/lxJDxtPKQ9KbB5u1VwIkJQ (shorter version)
- ADD LINK to the published guidebook

## Documentation

- [x] Finish and merge [numpy/numpydoc#525: DOC: Clarify recommendations regarding use of backticks](https://github.com/numpy/numpydoc/pull/525)
- [x] Use `intersphinx_registry` in SciPy [PR](https://github.com/scipy/scipy/pull/20876)
- [x] [using `intersphinx_registry` for numpydocs](https://github.com/numpy/numpydoc/pull/563)
- [x] [Add lookup functionality to instersphinx_registry](https://github.com/Quansight-Labs/intersphinx_registry/pull/14)
- [X] [Consistent monospace usage in API documentation page titles](https://github.com/scipy/scipy/pull/20890) (for SciPy)
- [ ] Move to pydata-sphinx-theme?
- [ ] Converting numpy-tutorials to MyST-JS (branch: https://github.com/numpy/numpy-tutorials/tree/mystjs)
- In progress

Interactive documentation discussion: https://hackmd.io/@melissawm/interactive_docs

## SciPy

- Follow up on the univariate distribution RFC and final call for action: https://github.com/scipy/scipy/issues/15928#issuecomment-2145952393
- [x] Added Dan Schult to the team, welcome! :tada:
- Doc final transition from mailing list to discourse forum [PR](https://github.com/scipy/scipy/pull/20881)
- Some deprecations https://github.com/scipy/scipy/pull/20858
- Const cython https://github.com/scipy/scipy/pull/20891
- Maint https://github.com/scipy/scipy/pull/20883

## Tooling

- A tools team was formed with 5 members, PRs are to a private repository so cannot be linked to.
- Migrate circleci-artifacts-redirector-action: [documentation PR](https://github.com/scientific-python/circleci-artifacts-redirector-action/pull/49), [website PR](https://github.com/scientific-python/tools.scientific-python.org/pull/48), [sanity check](https://github.com/mne-tools/mne-python/pull/12647), etc.

## SPECS

### SPEC-3: Dispatching (Spatch)

Discussions around this parallel to the Array API, exchange of information about the NetworkX work.
Started some thoughts/notes [here](https://hackmd.io/yI1iAqekQIq0a4jLS9WPyw) about different designs of dispatching/backend selection (seberg).

### SPEC-7: Seeding pseudo-random number generation (SPRaNG)

Draft getting there: https://github.com/scientific-python/specs/pull/180
We have a global solution for a smooth migration based on a decorator and check RNG helper.

* PR to SciPy to do the ball rolling (one place is actually changed): https://github.com/scipy/scipy/pull/20899

### SPEC-8: Supply Chain Security

TBD

### SPEC-9: Governance

PR: https://github.com/scientific-python/specs/pull/323

### SPEC-10: Changelog and release documentation

PR: https://github.com/scientific-python/specs/pull/321

### SPEC-12: Maths

PR: https://github.com/scientific-python/specs/pull/326

### SPEC-13: Naming conventions

PR: https://github.com/scientific-python/specs/pull/324

PR: https://github.com/scientific-python/specs/pull/321

## Backport bot worklog:

- Stefan and Matthew invited to both https://github.com/Meeseeksbox and invited to heroku where the bot runs:
- https://dashboard.heroku.com/apps/meeseeksbot/access

The code of the bot is in:
- https://github.com/meeseeksBox/meeseeksdev
When a new commit is made, this is deployed to heroku.

There is also a github app https://github.com/organizations/MeeseeksBox/settings/apps (Lumber bot app), this is the github side that when installed into a repository will give the right credentials to the heroku app.

Matthias made a new deployment of the bot as it hadn't been deployed in over a year.

## Parallelization/workers setup discussion

Discussed potential new API for sklearn, although we would need this to be a global variable (or module that everyone shares): https://hackmd.io/84thx0ucQ2ab17ZYrBhWRw