AiiDA Team Meeting 2024-02-08

# AiiDA Team Meeting 2024-02-08 ###### tags: `team meetings` ###### time: 14:00 CET [TOC] ### Present * Edan * Giovanni * Sebastiaan ### Catch-up round *Max. 3 minutes each* Giovanni: * Welcome Julian and Ali! (On SwissTwins) We discussed some very first tasks for now, both to learn and to fix simple but useful things for usability/user friendliness (see also last meeting notes). Tasks are temporarily here: https://docs.google.com/document/d/1PnfVmDNOb0s668pLOny5gh-s_8V4_ue53aXzIPvOa6Y/edit and a few more will be discussed in ~2 weeks * I was migrating some old 1.6 DBs and I got the same error as Daniel Marchand in this thread: https://groups.google.com/g/aiidausers/c/F8APVrr8Ht4 * I'm not sure if it's worth fixing, but maybe we can * 1. put some more clear information that you have do do version by version (maybe explaining which versions are really needed, probalby for the 2.x one only needs 1 or 2 steps) * 2. prevent a too-recent AiiDA version to migrate the config file if the version is too old (luckily there are backups, but when I migrated to 2.5 it went to v9, and then when I went back to 2.0 to do the step-by-step migration it complained that it needed v8) * 3. If we have time, check what the actual problem is, mostly to have the same problem in the future Sebastiaan: * **PRs** * [`SlurmScheduler`: Make detailed job info fields dynamic](https://github.com/aiidateam/aiida-core/pull/6270): Use official `--long` option or use subshell to determine supported fields? Advantage of latter is that it is complete, but it is more fragile as it relies on the output format of `sacct --helpformat`. The `--long` option is more reliable but doesn't contain all information. * [CLI: Add the `verdi node list` command](https://github.com/aiidateam/aiida-core/pull/6267) * [Dependencies: Update requirements for kiwipy and plumpy](https://github.com/aiidateam/aiida-core/pull/5732): This has been open for a long time. The reason is that there is a single test failing in `aiida-core`, but it seems to be affecting just the test suite and not real usage. Essentially, there is a timeout when the communicator is shutdown, which happens when a runner is closed (daemon stops for example). Nevertheless, I think this is good to get it. The updated versions of underlying libraries fix a pernicious bug where a `ChannelInvalidStateError` would be raised by processes when the daemon is under heavy load. Have been debugging this for a long time with Lorenzo Bastonero and he said that with this branch, all of these problems have gone and he can run in high throughput without issues. So I would propose to merge this into `main` so we can start testing it more widely. Since we just released `v2.5` this would give us some time to test this properly in the wild before releasing it with `v2.6`. * [ORM: Switch to pydantic for code schema definition](https://github.com/aiidateam/aiida-core/pull/6190): * [ORM: Use pydantic to specify a schema for each ORM entity](https://github.com/aiidateam/aiida-core/pull/6255): Implementation of [this AEP](https://github.com/aiidateam/AEP/pull/40). Have gotten some feedback from Chris which I have addressed. Would be great to have others have a look at it as well. Edan: * QueryBuilder * Chris's PR [#6245](https://github.com/aiidateam/aiida-core/pull/6245) * Merging against Seb's pydantic changes delayed * AiiDAlab widget * Seb's new monitor feature (comments) * Backup test Marnik: * Nothing to report on the `aiida-core` front. * Next week 15-16 Feb: `aiida-core` coding days. Main topic is testing the `verdi backup` command so we can get this merged ASAP. (If the already discussed issues are resolved, else perhaps it doesn't make sense?) Other topic suggestions welcome! Will touch base with Seb/Jason early next week to flesh out the focus some more. Jason: * Two DOC PRs: * Doc: quick start to lighting dive user in [#6261](https://github.com/aiidateam/aiida-core/pull/6261). With this non-services profile, the daemon cannot start because the runner complain about the communicator not exist. But I think this can be workaround by not raise the exceptions and then can have multiple daemon to run processes. * DOC: Add more instructions on how to use docker image [#6249](https://github.com/aiidateam/aiida-core/pull/6261) * A quick demo on the possible way for AiiDA GUI that can directly use things we developed in AiiDAlab. * Update the [developers wiki](https://github.com/aiidateam/aiida-core/wiki/Development-environment) with using docker image for dev environment setup. * Do we need to find time to go through [team resposibility](https://hackmd.io/hYaz4oXyTfWDUWKIKc0HUg)? Miki * [StructureData](https://github.com/aiidateam/aiida-atomistic/blob/0bca290cfba70b51d532b252e0f41c212f5cfb22/docs/source/user_guide/structuredata/API_start.ipynb): * orm.StructureData class ported into [aiida-atomistic](https://github.com/aiidateam/aiida-atomistic/tree/main), start the cleanup * working on a j notebook as documentation for the API * working on having a working example with aiida-quantumespresso for the next-coding week * Testing the aiida-worktree for (aiida-)[koopmans](https://koopmans-functionals.org/en/latest/#) ### Specific Agenda Items