# AiiDA Team Meeting 2023-04-05

###### tags: `team meetings`
###### time: 1400 CET

[TOC]

### Present

* Marnik
* Francisco
* Xing
* Jason
* Edan
* Giovanni
* Sebastiaan
* Kristjan

### Catch-up round

*Max. 3 minutes each*

Chris:
* I'm on holiday until 18th
* New release(s) of quantum-mobile (v23.04.03): https://quantum-mobile.readthedocs.io/en/latest/releases/index.html
* All codes and aiida installed in separate Conda environments
* Incorporates work progressing AiiDA plugins to aiida-core v2 (still some to finish): https://github.com/marvel-nccr/quantum-mobile/issues/211
* Suggest someone downloads "Developer preview for macOS / Arm64 (M1/M2)": <https://www.virtualbox.org/wiki/Downloads> (+vagrant), and tries running the ansible
* should/could maybe work up to: https://github.com/marvel-nccr/quantum-mobile/blob/f72194714e958e754f4c2da2313888edcad5db94/playbook-build.yml#L141, but then the conda stuff will be different

Marnik:
* Last two road map items:
* Future: make migration of data types defined by plugin packages possible [[#20](https://github.com/aiidateam/team-compass/issues/20)]
* Future: make aiida-core fully domain-agnostic [[#21]( https://github.com/aiidateam/team-compass/issues/21)]
* Start organisation of coding week (tentatively 10-14 July 2023):
* Francisco: do we have documents on the previous organisation somewhere?
* FFR: Yep should be in the same folder with the other coding weeks - https://drive.google.com/drive/folders/1HqntVMIMCech-YVXwidfswFGBGu7s_RJ
* JY: I create a new folder in https://drive.google.com/drive/folders/1CGshka-ktNgFXNSe9tKL3TgOKKIelo1n?usp=share_link
* Do we stick to Leysin? Or look into other options?
* How open do we make the invitation?
* GP: I actually need to look into the Pillar 3 budget... I already asked and should get an overview soon.

Giovanni:
* who has a Mac M1/M2 and can help with Quantum Mobile for M1?
* Mbx: I have one, but was considering if we should perhaps just get a mac mini? Nicola was mentioning we still had some funds for hardware?
* Marnik, did you have the time to work on reproducing the infinite recursion bug with Chris?

Jason:
* Coding week: the list of people who we want to invite (https://docs.google.com/spreadsheets/d/1TF8jVq7Eyf0wezFVATIdK75kTBMUjAjcGOAo0ao1IAQ/edit#gid=0).
* GSoC: the proposal submission phase terminated yesterday, I and Leopold decide to select Ahmed (https://github.com/AhmedBasem20) from Eygpt. He will working on "Ranking system for AiiDA plugin registry".
* There is a issue in AiiDAlab may caused from aiida-core, and probably the reason of memory leak of AiiDAlab. I and Sasha take sometime look it together but with no luck. It is again from we use threads in the widget to fetch process information from aiida engine. (https://github.com/aiidalab/aiidalab-widgets-base/pull/471#pullrequestreview-1357821001)

Francisco:
* Nothing to report on aiida-core.

Sebastiaan:

* Release 2.3.0: Prepared the [release branch](https://github.com/aiidateam/aiida-core/pull/5958). Please check the [`CHANGELOG`](https://github.com/aiidateam/aiida-core/blob/390c2bdf92d1eb9e91b1e4394d46e1c44780cf57/CHANGELOG.md) notes and see if you want to improve or add to the highlights. Ideally, we test the current `main` branch for a while without adding features, so we can see if there are any glaring bugs that were introduced, before releasing it. Who is running off of `main` regularly?
* [GP] Should we schedule releases a couple of weeks in advance and have a branch of AiiDAlab test the new release? This also means that AiiDAlab can quickly adopt the new version.
* Had a look at refactoring the `Scheduler` interface to make it compatible with FirecREST. In [this branch](https://github.com/sphuber/aiida-core/tree/fix/generalize-scheduler-interface), I generalized the `Scheduler` interface, making the main abstract methods: `submit_job`, `kill_job`, `get_jobs`. These methods are implemented by a new class, for now called `BashScheduler` which essentially reproduces the old behavior. Existing plugins can simply subclass from `BashScheduler` instead of `Scheduler` and they will continue to work without any other changes. The new generalized `Scheduler` interface should also make it possible to implement it for FirecREST. I made a [quick mock-up of the implementation](https://github.com/sphuber/aiida-core/blob/feature/firecrest/aiida/schedulers/plugins/firecrest.py), but cannot test it since I don't have access to a FirecREST instance. Then we just need a `FirecrestTransport` for file transfers, and then a user could configure a `Computer` using `firecrest` for the scheduler and transport type.

Xing:

* finalized the `XpsWorkChain` and merged to aiida-quantumespresso.
* [GP] plans for using the AiiDA-REST API for AiiDA Quantum ESPRESSO (e.g. in AiiDAlab)?
* [XW] I am working on this. I will show a demo in next AiiDAlab meeting.
* [GP] OK! Good to make sure also Sebastiaan and Chris are aware of the tests.
* [XW] Yes, I am also preparing a feature list that need to be added to aiida-restapi.

Kristjan:

* Nothing aiida-related to report.

### Specific Agenda Items