continued maintenance of `pygam`
@dswah, @adiezolivan, @lyessi78, it does look like this package has lapsed, or is in the process of lapsing maintenance. It would be very much a shame, GAM are very useful for a number of applications.
We have also seen the [HELP WANTED] signs, so here we are!
Some other signs of lapsed maintenance:
- no release since Feb 2024
- sensible PR such as python 3.13 compatibility https://github.com/dswah/pyGAM/pull/372 or fixes do not get reviews or other maintainer interaction
At the German Center for Open Source AI, we would be happy to help with maintenance - some of our core members have requested explicitly we consider helping out pygam, or taking on maintenance entirely.
We are non-profit and will ensure to keep the packages open and maintained. We are also happy to give continuous credit to you and not change the branding (depending on current owner preferences).
We have done this previously in high-profile cases such as pytorch-forecasting, see
https://github.com/sktime/pytorch-forecasting/issues/1592
We would like to keep the packages open (permissive license), and are happy to review contracts you would like signed.
From an engineering perspective, we usually look for opportunities to simplify maintenance process, integrate software interfaces. From an operations perspective, we look towards professionalization, and onboarding new maintainers to ensure quality, and redundancy, to avoid future lapses of maintenance.
ping, @dswah
FYI, I have opened a PEP 541 request now: https://github.com/pypi/support/issues/6611
Hi @fkiraly @adiezolivan @Lyessi78, and all, Indeed I no longer have the time to be the (sole) maintainer for this pygam.
I would love your help to keep this pygam alive!
What steps can we take?
Brilliant, @dswah!
To break this down into concrete steps, I would suggest:
- direct call with GC.OS to map out the parameters of the collaboration, e.g., is this going to be transfer or co-maintenance - email me at fkiraly - at - gcos - dot - ai and I will coordinate finding a time slot
- think about your preferred status, where you would like to see the package go, how (or whether) you would like to remain involved, whether we need to sign any contracts from your side
- you can also join the discord server for quick discussions with GC.OS maintainers (the name is "sktime" but we are currently using this discord for GC.OS as well, simply due to critical mass related to time series) https://discord.com/invite/54ACzaFsn7
- we should think about a maintenance or handover plan, what the priorities are - imo most important:
- python 3.11 and later version, compatibility
- going through open/stalled PR and getting them merged
- homogenizing maintenance model with GC.OS packages
@fkiraly agreed. these seem like perfectly reasonable first steps. my goal is for this package to continue to be useful to the community (and indeed to continue to improve in key functionality) and to continue to be open source, so i am totally open to any discussion about ownership, branding etc since the latter seem orthogonal to the former.
sending you an email...
Hi @dswah! I would like to help to maintain this package alive as well, on behalf of my Company. Can you please provide me your email so I can directly contact you?
see https://github.com/dswah/pyGAM/issues/379 for next steps