Recurring OSS hacksessions

[mhauru]

Should we keep organising events like the one we had for hacktoberfest?

[llewelld]

Yes please.

[mastoffel]

Definitely!

[mhauru]

We'll run these monthly at least for now, every third Friday of a month, taking turns organising. At least @mhauru, @rwood-97, @mastoffel, @crangelsmith are volunteering to organise.

[mhauru]

Outputs from the first session (hacktoberfest) on 2023-10-20:

• Adding details to README for new users opendp/smartnoise-sdk#578
• docs: updating the TimeGAN notebook to support the new version of the API ydataai/ydata-synthetic#310
• Updating team model AIrsenal#561
• First attempt for individual hatching styles for stackplot matplotlib/matplotlib#27158
• Add warning message when scaling steps return NaN tidymodels/recipes#1252
• Bump python version to >=3.8 in pyproject.toml opendp/smartnoise-sdk#577
• feat: Add Cramer-rao uncertainties + covariance using autodiff to non-minuit fits by default scikit-hep/pyhf#2269 (comment)

Also:

• Ryan: "Found a bug in llama-index which I will raise an issue and PR to fix at another point"
• Markus: "Also made some progress towards adding some type hints to the package, but not ready to submit yet."

Might be of interest for @KatrionaGoldmann

[rwood-97]

Ryans llama-index PR is now merged, see here

[mhauru]

Also this: pygments/pygments#2552

[mastoffel]

Following discussions with @AoifeHughes and @martintoreilly , find a good place for people in the Turing to discover the regular hacksessions, such as the reading groups wiki, Almanac and/or Handbook.

[mhauru]

The information discovery function is now served by this wiki page: https://github.com/alan-turing-institute/research-engineering-group/wiki/Open-source-hacksessions Might link to it from at least the Handbook?

[rwood-97]

Outputs from December hacksession:

• Add AGPL as a license option scientific-python/cookie#339 (Markus)
• yongrenjie/sphinx-exec-directive@c0496ee (Jon)
• Add segmentation map processing to SAM Image Processor huggingface/transformers#27463 (Rosie)