Wrapping up 2025

Completing the streak

For some random reason, at the beginning of 2025 I decided that I was going to do an RDKit blog post every week in 2025. It was a struggle at times, but if you include updates of old blog posts (which I definitely do!), I made it! :tada: For the last post of the year, I’m going to do a short look back at 2025.

The RDKit

For me the biggest RDKit thing of each year is the UGM, and in 2025 there were actually two. The first North American edition of the UGM took place in April in Cambridge, MA. I didn’t manage the travel for that, but it sounded like it went really well and was a great success. Here’s hoping that we can make that a regular thing! The European UGM took place in September in Prague. I did a recap post back in September, but the UGM was great. I love the chance to spend a couple of days surrounded by the RDKit community. Registration for the 2026 European UGM, which will take place from 16-18 September in Darmstadt, Germany, will open early next year.

Some numbers

(GitHub could make this easier…):

• two major releases
• ten minor releases (well, nine so far, the tenth comes early next week, before the year ends)
• 348 commits
• The two largest contributions (in terms of lines of code) were the SCSR parser from Tad Hurst at CDD and the CDX/CDXML parser from Brian Kelley at Glysade (using code contributed by Revvity)
• 83 contributors (counting pull requests and resolved issues and enhancements)
• 147 closed bugs
• 122 completed enhancements
• 39 cleanup PRs
• 17 documentation PRs
• I’d include something here about the number of discussions posts, but I’m not willing to invest the time to figure out the GraphQL API for doing that.

Things that should go better

Reviewing pull requests, particularly larger ones, can take much longer than it really should. We need to figure out a better way to handle these.

Next year I would like to be a bit better at paying attention to the Discussions tab in GitHub. I’m not 100% sure that GitHub Discussions is the best option for questions and general discussion, so I’m also trying an experiment with zulip. If you’re interested, this link should work to allow you to join automatically for the next 30 days. There’s a chicken and egg problem with this (we need participants for it to be useful, but it needs to be useful to attract participants), but we’ll see how it goes.

Because of all the other stuff going on, I didn’t manage to complete (or even make a lot of progress on) some of my longer-term RDKit projects this year; let’s see what 2026 brings there.

Related stuff

We continued to use and improve our lightweight registration system lwreg (the Application Note is here). One of the big things this year is that Jessica cleaned up the docs and figured out how to get things working with readthedocs, so now we have really nice looking documentation.

Sereina and I taught a new course on Cheminformatics and Computer-Aided Drug Design in the fall semester this year. Of course we used a lot of RDKit for that. :-) As you would expect, putting together a new set of lectures and associated exercises (we did hands-on stuff in Jupyter) was a lot of work, but I really enjoyed both the preparation and doing the lectures. We’re not doing student evaluations this year, so we won’t get feedback from the students, but hopefully they found the material useful.

Personal things

Its my blog, so I’ll go ahead and include some non-RDKit stuff too.

According to Garmin I have (so far, there are still a few days left in the year) run 2563km with about 71km of elevation gain (that doesn’t count elevation when doing uphill treadmill workouts). Over the summer I ran two ultra-marathons: a 77km race that I finished and a 170km race that I DNF’ed after 100km (it turns out that running that distance is even more complicated than I thought it would be). Getting ready for those took a lot of time, but much of it was fun days moving “fast” in the mountains, so I enjoyed it. I also ran my usual stage for the research group’s team in the Zurich SOLA relay race and managed a PR there. Finally, at the end of the year I set myself the random goal of running a fast 5km (not a race, just a fast time on a stretch I run a few times a week). I didn’t manage to hit my time goal of 20 minutes, but I only missed by 16 seconds, so I’m still pretty happy with the time. It was definitely entertaining to change things up and do some focused training for going faster.

I hiked 653km with about 40km of elevation gain; this is down from a normal year because of the amount of running and because our summer vacation didn’t end up going as planned.

The climbing doesn’t lend itself to quantification (at least not the way we do it) but we ended up doing a fair amount this year. I did considerably less bouldering than last year since I mainly went running on the days I normally would have gone to the bouldering gym, so I didn’t make as much progress on technique as last year, but it looks like I didn’t lose too much.

Ok, it’s time to wrap this up and get back to my holiday project.