# AiiDA Team Meeting 2024-03-21

###### tags: `team meetings`
###### time: 14:00 CET

[TOC]

### Present

* Marnik
* Julian
* Sebastiaan
* Xing
* Ali
* Edan
* Jusong
* Kristjan

### Catch-up round

*Max. 3 minutes each*

Marnik:
* Slack workspace: who are the current admins? Can we have some more please?
* [Discussion on shared queueing system interface](https://github.com/Matgenix/qtoolkit/issues/39): Raised by Jan Janssen from the PyIron workflow manager. Idea is to explore developing a common package for interfacing with queueing systems.
* Note: I've been working with developers of several other workflow managers (Jan Janssen from PyIron, Janine George from JobFlow/Atomate/Fireworks and Joerg Schaarschmidt from SimStack) on making a comparison between the various engines. One thing we're doing here is comparing a full implementation of an EOS workflow, along with the interface with the external code (Quantum ESPRESSO), see https://github.com/materialdigital/ADIS2023/tree/main/eos.
* Very curious about no-comms broker: quick clarification: I suppose this only allows for _running_ in a sustained Python instance?
* Next AiiDA-core days: 11-12 April?

Julian:
* Finished updating Renkulab setup and created [PR](https://github.com/SwissDataScienceCenter/contributed-project-templates/pull/79) on the `contributed-project-templates` repo
* Some updates and refactoring to `workchain dump` → Also, change to `verdi (calcjob|process) dump?`
* [Small fix](https://github.com/aiidateam/aiida-core/pull/6326) to `verdi quicksetup` which was breaking when there was a profile with an sqlite-backend in the config
* [Some testing](https://github.com/aiidateam/aiida-core/pull/6322#discussion_r1530627156) on using relative `AIIDA_PATH` → Should add validation
* SwissTwins: Meeting with Rico Haeuselmann on his [proof-of-concept AiiDA plugin](https://github.com/DropD/aiida-icon-clm) to drive a W&C simulation code

Ali
* Done:
* Deprecated --reset flag in `verdi daemon restart` [#6317](https://github.com/aiidateam/aiida-core/pull/6317)
* Manual testing of FirecREST, and sent a report with requested features
* `aiida-registry` fixed plugins info from `setup.json` to `pyproject.toml` [#309](https://github.com/aiidateam/aiida-registry/pull/309#issuecomment-2009160806)

* To do:
* `verdi computer delete` to allow users to delete computers with associated nodes issue [#5395](https://github.com/aiidateam/aiida-core/issues/5395)
* `aiida-registry` add a GHA to email respective maintaniers on errors. The goal is to make everthing apear less buggy.
* Find a way to integrate async in aiida transport /& monitoring of transport

Sebastiaan:
* **Merged PRs**:
* [Caching: `NodeCaching._get_objects_to_hash` return type to dict](https://github.com/aiidateam/aiida-core/pull/6323): Makes it easier to debug caching
* [Caching: Include the node's class in objects to hash](https://github.com/aiidateam/aiida-core/pull/6321): Ensures that `Data` subclasses are distinguished in hashing
* [Config: Check current working directory for existing config dir](https://github.com/aiidateam/aiida-core/pull/6322): Behavior for finding `.aiida` directory now matches `git`. The current working directory (and all its parent folders) are checked iff `AIIDA_PATH` is not set. This was a prerequisite change for `verdi init`. (on a side note we realized that `AIIDA_PATH` is not validated at all and when it is used to specify relative paths, all hell breaks loose. Simple solution would be to add validation and require absolute paths, but technically this is breaking. Is it likely anyone is relying on this?)
* [Config: Set existing user as default for read-only storages](https://github.com/aiidateam/aiida-core/pull/6318): Fixes bug when creating profile from archive, where many commands would error because no default user would be defined (assumed practically throughout AiiDA's API)
* [Refactor: Abstract message broker functionality](https://github.com/aiidateam/aiida-core/pull/6303): Preparatory to allow running AiiDA in "light" mode, without RabbitMQ required as a service.
* [ORM: Switch to pydantic for code schema definition](https://github.com/aiidateam/aiida-core/pull/6190): Follow up of adopting pydantic

* **Open PRs**:
* [CLI: Add the `verdi init` command](https://github.com/aiidateam/aiida-core/pull/6315): All prerequisites are merged, this is ready for review. Need to decide on the desired behavior.
* [Caching: Add `Process.CACHE_VERSION` class attribute to control caching](https://github.com/aiidateam/aiida-core/pull/6328): Complementary to [this PR](https://github.com/aiidateam/aiida-core/pull/6215) that intends to remove all version information (core and plugin) from hash computation. This would lose control for plugins to "reset" cache if the implementation of (especially `CalcJob`) plugins changes significantly. This PR introduces a method for them to reset the cache when needed.

Edan:
* QB PR complete [#6245](https://github.com/aiidateam/aiida-core/pull/6245)
* Issue [#6327](https://github.com/aiidateam/aiida-core/issues/6327) - proposing to switch references to backend keys in code and docs, e.g. no references to `id`, as QB PR now allows the use of `pk` (backend mapping)
* See [#6311](https://github.com/aiidateam/aiida-core/pull/6311) for proposed guidelines for using the `versionadded` directive in documentation
* For Docker users, please give the new `aiida-core-dev` image [#6298](https://github.com/aiidateam/aiida-core/pull/6298)

Xing:
* **Open PRs**:
* [add should_run_scf for PwBandsWorkChain](https://github.com/aiidateam/aiida-quantumespresso/pull/1016): Set the `scf` step as an option in the PwBandsWorkChain. The goal is share one `scf` step among other workchains (e.g., bands, pdos). Anyone againt this?

* **Merged PRs**
* [XPS: Fix inputs for molecules](https://github.com/aiidateam/aiida-quantumespresso/pull/1013) together with Michael.

Jason:
* Nothing very specific to aiida-core.

Kristjan:
* Backup pr: https://github.com/aiidateam/aiida-core/pull/6069
* all discussed changes are done
* CI passes (except for the docker action)
* Ready for some final tests and hopefully we'll get it merged soon.
* Google summer of code:
* new aiida-explorer, general issue: https://github.com/materialscloud-org/issues/issues/33
* hierarchical/foldable workflow nodes
* Does the new rest api need work before the graph browser can be implemented?
* timeline:
* 2nd of April: student proposals
* 1st of May: decision on who we pick
* 1st of June: work starts

### Specific Agenda Items