- Attendees: @sloosvel @schlunma @bouweandela @ehogan @rbeucher @valeriupredoi
- Absentees: @headmetal (not TLT anymore) @flicj191
- Chair: @valeriupredoi
- Minutes taker: @valeriupredoi
Meeting | Assignees | Item | Status |
---|---|---|---|
2024-08-13 | @bouweandela @schlunma | Variable derivation for ERA5 in the fix | pending |
2024-08-13 | all | Review RTW pull request | pending |
2024-08-13 | all | Find a new release manager | done |
2024-09-03 | @schlunma | Re-ping UET with e-mail and ask for review of new configuration PR | done |
2024-09-03 | all | Brainstorm possible topics for new ESiWACE computational performance improvement proposal (call will be published here, likely towards the end of the year) | pending |
2024-09-03 | @schlunma | Circulate paper about ESMValTool Dask improvements | done |
- Review of Pending Action Items: one new set to DONE: next release manager @sloosvel (many thanks, Saskia!)
- @schlunma: Vote on potential new TLT members: @bettina-gier Tina: will be working on IntakeESM, Healpix (technical stuff), PR reviews - there was a vote, and all in favour! @valeriupredoi and @schlunma: add her to the mailinglist, Slack, Google Cal event for the TLT; also Mike Tetley @headmetal to be removed, and add Charles Taylor (email off Slack);
- @bouweandela: Can we disable coverage collection by default when running tests locally ESMValGroup/ESMValCore#2456 -> takes about 2x slower to run tests: we like it, so long as there is documentation on how to turn it back on;
- @bouweandela: Switch from yapf to ruff and fix our pre-commit set up? See e.g. ESMValGroup/ESMValTool#2161 and ESMValGroup/ESMValCore#2490; @schlunma any plans for esmvaltool? @bouweandela: we'll try see how it goes;
- @bouweandela: Consider adopting spec0 ESMValGroup/ESMValCore#2533 -> all happy about it, @bouweandela points us to a checker, also we should start preparing docs for it, and adopt the new standard;
- @ehogan: Would we like to help create a geospatial benchmark suite, as described in this blog post from Coiled (Matt Rocklin's dask-as-a-service company): Large Scale Geospatial Benchmarks? Martin and Bill in AVD suggested that ESMValTool might well be able to help out with one of the asks, which is code that does large-scale geospatial data processing; @ehogan promised to open an Issue or Discussion item about it;
- @schlunma: What do we about NCL? There are (again) problems with our environment, see ESMValGroup/ESMValCore#2535: separate module maybe; @bouweandela: leave it to user to install it; @schlunma: diagnostics are old, most devs switching to Python for new stuff -> topic to be brought at the workshop; @rbeucher: a few recipes (eg Perfmetrics) that are good, to be ported to Python: possible solutions include separate conda envs and docker containers; this item has a Workshop TODO;
- @schlunma: Very brief update on ESMValTool+Dask paper: to be re-circulated today(ish) and some of us still need to send comments; @bouweandela give feedback on current performance status: few benchmark discrepancies to look at, i.e. perfomance is even better (e.g. used dev branch of
iris
); - @valeriupredoi: work on modernizing and re-deploying our ESMValBot https://github.com/ESMValGroup/ESMValBot/pull/37
- @valeriupredoi:
codacy-ruff
plugin to Codacy: ESMValGroup/ESMValTool#3739 - @rbeucher:
intake-esm
catalog support: more and more interest in adding datasets to the ESM catalog; @bouwenadela: looking at XCube also, a proposal in the pipes, so there may be time to implementintake
support; coordinate with Tina, so there is no duplication -> a meeting to be set up; - @schlunma wants to merge PR that introduces new (dask-like) configuration system today (ESMValGroup/ESMValCore#2448): this needs be merged ASAP;
- @ehogan: It's International RSE Day on October 10! 😊 yay!
- AOB: good news! We have our next Release Manager (for v2.12.0): @sloosvel - thank you, Saskia! ETA for release January 2025.