AiiDA Team Meeting 2024-04-04

# AiiDA Team Meeting 2024-04-04 ###### tags: `team meetings` ###### time: 14:00 CET [TOC] ### Present * Marnik * Xing * Miki * Edan * Jusong * Kristjan * Sebastiaan * Julian * Ali ### Catch-up round *Max. 3 minutes each* Sebastiaan: * **Open PRs**: * [CLI: Add the `verdi init` command](https://github.com/aiidateam/aiida-core/pull/6315): Big last open question is the logic to determine the location of the `.aiida` folder. Last suggestion was to have `AIIDA_PATH` have precedence over CWD. However, that essentially breaks `verdi init` if a user has `AIIDA_PATH` defined. Switching the precedence has its own problems though. Because if CWD is checked first, _if_ an `.aiida` folder exists in the home directory (which is likely for most installations) then it will always default to this, always ignoring the `AIIDA_PATH`. Only hacky way around is to exclude the home folder from the CWD hierarchy lookup. See [this comment](https://github.com/aiidateam/aiida-core/pull/6315#issuecomment-2017531751) for details * [Caching: Add `Node.CACHE_VERSION` to node attributes](https://github.com/aiidateam/aiida-core/pull/6328): See [this comment](https://github.com/aiidateam/aiida-core/pull/6328#issuecomment-2035584201) for details on current status. * [`verdi node graph generate` add `-N/--nodes` to accept multiple root nodes](https://github.com/aiidateam/aiida-core/pull/6338): Currently only accepts a single root node positionally. To properly deprecate it, a new option had to be added `-N/--nodes` instead of allowing multiple positional arguments. This is because the current command takes the output file as last optional positional argument. Allowing multiple root nodes positionally would break the output file. The latter is also deprecated in favor of `-O/--output-file`. * Working on improving the `pytest` fixtures that we provide. Most important one is to provide an `aiida_profile` fixture that creates a `core.sqlite_dos` profile without broker. This would make testing a lot easier for plugins as no services are needed. It will be configurable to easily use PostgreSQL and/or add RabbitMQ if that is needed for the tests. Testing this against `aiida-s3` and `aiida-shell` and almost done. Let me know if you are interested in discussing. Marnik: * Nothing to report on the `aiida-core` front. Xing: * Working on integrating the aiida-quantumespresso-hp to AiiDAlab as a QEApp plugin. This will allow user to calculate the Hubbard U and V in the GUI. * A new repo [aiidalab-qe-hp](https://github.com/superstar54/aiidalab-qe-hp) Edan: * [Docs: Where/how to document the backend key-alias mapping feature introduces in #6245](https://github.com/aiidateam/aiida-core/issues/6327) * In [#6245](https://github.com/aiidateam/aiida-core/pull/6245), a projection map was introduced to handle query builder aliases, e.g. `pk` -> `id`. This allows the user to use `pk` in the query builder instead of `id` regardless of the syntax. As such, I'm proposing we update the docs accordingly. * Resolution: Edan will scan the code for consistency and update codebase/documentation accordingly * @jusongyu what happened with the `aiida-core-dev` image? Jason: * Register as mentor for GSoC2024: https://forms.gle/CHV71nef1E3K7XeB6 * Question: is there rate limit for the AiiDA worker? I encounter the situation I have to increase the daemon to continue running things. Kristjan: * Gave an aiida talk in Tartu, Estonia: * one group will start using aiida-vasp, helped them set it up * GSOC: * One promising proposal on the "next-gen AiiDA provenance explorer" * main ideas from this comment: https://github.com/materialscloud-org/issues/issues/33#issuecomment-2022551485 * Do you see any issues with this as the first step? Ideas for "global provenance" browsing can potentially be implemented on top of this (e.g. ideas of Edan). * let's have a discussion during next aiida coding days * backup pr: * a small decision left: https://github.com/aiidateam/aiida-core/pull/6069#discussion_r1546578704 Miki: * Nothing to report Julian: * Nothing much to report * Writing tests for the dumping: * Added fixture to generate CalcJobNode with FolderData/SinglefileData inputs, repository content, `retrieved` output * Also fixture that creates nested WorkChain with one sub-WorkChain that has a CalcJob * Updating questionnaire and updated [aiida-statistics-query](https://github.com/aiidateam/aiida-statistics-query/pull/3) repo * Bring up discussion regarding `verdi init` (sorry, joined late) Ali: * Nothing much to report * working on: test writing for the new `verdi computer delete` [PR #6334](https://github.com/aiidateam/aiida-core/pull/6334) * Review with tests on `asyncfirecrest` [Practicality and Efficiency](https://github.com/eth-cscs/pyfirecrest/issues/94#issuecomment-2034033313) this will help `aiida-firecrest` ### Specific Agenda Items