---
tags: NumPy, BIDS
---

# 2019-3-13 BIDS/NumPy Development Meeting

- Join via Zoom at [https://berkeley.zoom.us/j/400054438](https://berkeley.zoom.us/j/400054438), or dial-in through a [local phone number](https://zoom.us/u/adQDmEc1wI).
- [Trello workboard](https://trello.com/b/Azg4fYZH/numpy-at-bids)
- [Previous meetings](https://github.com/BIDS-numpy/docs/tree/master/status_meetings)
- [NumPy at BIDS Calendar](https://calendar.google.com/calendar?cid=YmVya2VsZXkuZWR1X2lla2dwaWdtMjMyamJobGRzZmIyYzJqODFjQGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20)

**Present:** Tyler, Matti, Stefan, Chuck, Hameer, Ralf, Sebastian

## Follow-up from last meeting/discussions

- There's some initial content on the private [logistics gitlab repo](https://gitlab.com/numpy/logistics) for NumPy core devs. This content is primarily related to obtaining access to machines of various architectures for open source debugging purposes. We can probably continue to build this up slowly.
- Who has permissions to see it? (mattip) -- Ralf, Stefan, Tyler can add people
- Any progress with respect to [mixins discussion](https://github.com/numpy/numpy/issues/10446) for easier duck-typing of arrays?
- I brought it up when Dask was implementing `__array_function__`. https://github.com/dask/dask/issues/4583 (Hameer)
- There seems to be consensus that it should be separate from NumPy proper, but still a PyData community project. I'm willing to maintain it.

## Community Suggested Topics

- Responding to Outreachy applicants
- Matti replied privately, will cc the mailing list on the introduction mail that ended up there.
- Google Season of Docs https://developers.google.com/season-of-docs/ (ralf)
- NumFOCUS is going in the direction of being an organization, we can find a mentor if someone shows up. We should put some target goals on the NumFOCUS page.
- Random Number Generator work call (https://doodle.com/poll/qzqdygpq6nyvszgp) (Stéfan)
- Number of open PRs is climbing again (matti)
- Generate a html report of open PRS: names, last action, any labels (Nelle?)
- May 10/11 Developer Meeting (Stéfan)
- C-API changes: do we require a NEP before changing? (Chuck)
<what note/conclusion can we add here?>: ask on mailing list, Ralf suggested adding if it makes sense for non-breaking case?
- Reorganizing the `polynomial` namespace. Importing that namespace gives you the classes at the top level, but to get to the functions you need to dive deep into the namespaces. Is there documentation on how to use this all?
- Write a docstring for `numpy.polynomial.__init__.py` to describe recommmened use

## Work status

Matti:
Slowly getting back into it as I feel better.
- Worked out some of why there are so many documentation warnings when using `-n` with sphinx-build. My initial guess was only close, not correct: it is not the parameters of python functions, but missing parameter [declarations in c function signatures](https://github.com/numpy/numpy/pull/13104/files#diff-8d11b9e1fddd0ac2d5d7fe628fcc959f) as well as missing namespaces on polynomial and ma internal references.
- Merge the current PR and open an issue with for new contributors.
- worked a bit more on cython `@property` decorator for `cdef` functions. It would be nice to finish [PR 12284](https://github.com/numpy/numpy/pull/12284) - vendoring our own numpy.pxd file, in preperation for the new randomgen which uses cython extensively.
- [PR 12962](https://github.com/numpy/numpy/pull/12962) around `packbits`/`unpackbits` got a thorough review and I am working through the comments.
- Promoting Open Source in general and expanding the contributors to scientific python/NumPy in particular: I put some effort in networking with people I know who may be able to promote open source
- The head of the computer science dept at a local college (about 300 students) wants to move to python as a first language. I want to use this as an opportunity to market open source, python, and giving back to the community. Hopefully I will be invited to give a 90 minute "Python and Open Source: why and how?" kind of talk, here is a [first draft as slides](https://hackmd.io/p/By3NCkXwN#/) and [as text](https://hackmd.io/_90aBzOGSYqdj4LddFYWfQ?both#)
- A discussion with an industry trainer about promoting open source led to a discussion of presenting the case for contributing back to open source within the framework of helping companies develop a coherent open source use policy (license approval, tracking use of open source in projects that reach customers, how to manage multiple code sources, ...).

Tyler:
- I've put some effort into helping move a community-submitted `scipy.spatial` [ConvexHull enhancement PR](https://github.com/scipy/scipy/pull/9721) forward; proposed for merge on March 23:
- A few updates on my efforts on [Eric's `np.clip() ufunc` PR](https://github.com/numpy/numpy/pull/12519):
- looks like core dev reviews are currently slightly in favor of not requiring a code-level check for `amin < amax`
- I've written some [property-based tests](https://github.com/tylerjereddy/numpy-property/blob/master/tests/test_clip.py#L12) which demonstrate some weaknesses for enforcement of currently-preferred equivalence to `np.maximum(amin, np.minimum(arr, amax))`
- After discussion, we would prefer to raise an error if max is less than min.

---

Please remember to archive this file and commit it to github.com/bids-numpy/docs