Skip to content

Latest commit

 

History

History
53 lines (46 loc) · 6.49 KB

20231107.md

File metadata and controls

53 lines (46 loc) · 6.49 KB

Technical Lead Call 7 November 2023 (online; regular schedule meeting)

Participants

  • Attendees: @ehogan @valeriupredoi @bouweandela @zklaus

  • Absentees: @sloosvel @schlunma

  • Chairs and minutes takers: @valeriupredoi

Proposed topics for discussion

added once here since we have actually gone through and discussed them

[UFMM] == undiscussed for many (previous) meetings

  • Agenda section at the start of the meeting)
  • [UFMM - discussed below] @ehogan: would it be possible to make a decision on whether we should use black, please? See ESMValTool: discussion: #2161 for more details. I would be very happy to help with this, given the benefits! 😊
  • [UFMM - discussed below] @ehogan: I would also really like to discuss the use of pre-commit.ci, and I realise I am not here for this meeting, but I wanted to add it to the list so I don't forget about it / maybe we can discuss it in the October meeting? See ESMValTool: discussion: #3206 for more details.
  • [UFMM - discussed below] @ehogan: Should we add a "Security policy" to ESMValTool? See https://github.com/ESMValGroup/ESMValTool/community for more details.
  • [UFMM - discussed below] @ehogan: I have added this to the September agenda as I wanted to bring it to the attention of the TLT before October (happy to discuss further on Slack when I'm back from RSECon23) 😊 Would it be possible to discuss supporting those at the MO looking to develop ESMValTool to support more CORDEX data? I am happy to help, but I'm not that familiar with CORDEX / regional data. Here is some information from Katy Richardson, a Senior Scientist at the MO:
    • Met Office has a small amount of funding to support capability development for loading and processing CORDEX datasets in ESMValTool, with an initial focus on the CORDEX Africa model simulations
    • We are aware of numerous issues encountered with loading RCM data, mainly linked with specific coordinate systems, projections and metadata, and associated issues with e.g. area averaging calculations. There are varying issues across the different RCM models used in the CORDEX project - some more complex than others.
    • We are looking for support in prioritising a solution to these issues and incorporating a solution into an ESMValTool release to support future RCM and CORDEX analysis with ESMValTool.
    • The funding is from September 2023 - March 2024. Current plans are for this strand of the funded work to run between October 2023 and January 2024.
  • [still ongoing - discussed below] @all @zklaus: how are the releases going?
  • [still ongoing] @bouweandela: Who will maintain the ESMValBot and ESMValTool VM now that Remi is gone? - we should talk about having a talk
  • [still ongoing] @bouweandela: Plan future releases - we decided to talk about possible RMs at the wider ESMValTool call, just a reminder
  • [still ongoing, updated] @valeriupredoi: non-contiguous bounds when crossing 360deg ESMValGroup/ESMValCore#2229 - it'd be good to have some tests done with Rebecca's PR, as suggested by @schlunma
  • Versions of observational data (obs4MIPS, discussed below)

Meeting notes

  • @ehogan: use of black, ruff, precommit.ci:
    • @zklaus: ruff has a standalone library of linters, but prospector uses existing standards; K prefers black to yapf; there could be risks due to different standards implementations
    • @ehogan: better to finish up a draft proposal, that includes pros/cons
    • @bouweandela: precommit.ci has a cmd-line interface/outlook; Codacy is more user-friendly; the issues with different versions are the different pins in different envs (.pre-commit-config.yaml, environment.yml, setup.py, Codacy); pre-commit only allows the analysis of complete files (instead of just the changed lines), resulting in many issues getting raised that were present before, this may cause people to feel unjustly blamed for those issues if they are raised on their pull requests.
  • @bouweandela: we may need to move to ruff at some point, e.g. pycodestyle, one of the linters we use through prospector, is already deprecated because all of its functionality is available in ruff. I would prefer to wait until Codacy supports ruff though.
    • fair consensus: change from yapf to black yes but should be discussed with community, but not so much Codacy to precommit.ci; bring it up to monthly meetings;
    • @ehogan will work on a white paper-like distilled analysis of these tools;
  • @ehogan: security policy: we should add at least a line, @bouweandela brought up other packages SP: if you find anything send an email to Tech Lead Team etc etc
  • @ehogan: CORDEX@MetOffice work going on, but it appears the issues are mostly iris-related; @zklaus: don't look at it mostly as an in-house issue, share away with people that are interested; funding there is!
  • @zklaus: 2.10 releases: Core RC is out (yay!); Tool is next, outstanding issue - obs4MIPS changed standards, we have not picked up their latest version; Romain did update the version, but recipes don't work:
    • @bouweandela: code need version information from the directory structure
    • question at hand: shall we update the obs4MIPs version for the release or after? - there are small and big issues; @bouweandela prefers Romain finishes the PR fully; consesnus: we wait for after the release!
    • another issue is the coexistence of frequency and time_frequency facets for the obs4MIPs project on ESGF, one is wrong, but hard to get to fix it upstream (at ESGF!); also the institute for CCCma ESMValGroup/ESMValCore#1937 and various others;
    • TODO for 2.10: missing bits to be readded: ESMValGroup/ESMValTool#3405
  • @valeriupredoi: Corectness conference at NCAR, presentation to be written by V with much support from @bouwenadela (many thanks!)
  • @valeriupredoi: Python 3.12: almost there, we need to raise the flake8 issue at https://github.com/PyCQA/flake8/issues - @ehogan volunteered to open that issue (many thanks!)