# AiiDA Team Meeting 2024-10-03

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

[TOC]

### Present

* Julian
* Ali
* Miki
* Edan
* Alex
* Kristjan
* Xing
* Jason (on holiday, please read my part, thanks!)
* Giovanni

### Catch-up round

*Max. 3 minutes each*

Jason (on holiday):

- The default set for the `verdi storage maintain` operation will **duplicate** loose to pack **without** compression. The most expected behavior will be compressed and delete the loose when it is sure that all the data are moved to pack. (partly overlap with [#6539](https://github.com/aiidateam/aiida-core/issues/6539))
- [AK] In summary:
- loose objects are not deleted after packing.
- packing objects are done without compression

Marnik:
* No time for coding :sob:
* Attended MSE conference in Darmstadt: gave talk on AiiDA.
* AiiDA coding days (18-22 Nov): Was looking for options in Engelberg, Kandersteg, ... but seems challenging. We got an offer from a place in Leysin where we have been before, which is reasonable. Travel from Brug is ~3h15m.
* Need help with further organisation (schedule, topics etc).

Miki
* Worked on `StructureData`:
* some examples on PwBaseWorkChain and EoS (via WorkGraph), [here](https://mikibonacci.github.io/aiida-atomistic/tutorial-qe)
* Hubbard property is implemented
* Testing Query of `StructureData`, default values, moving some properties in the repository (no query possible)

Xing:
* Focused on AiiDAlab in the last few weeks. Nothing to report in aiida-core.

Edan:
* AEP regarding naming conventions in AiiDA classes (coming soon)

Ali:
* Bug fix on `verdi process kill`, together with Alex:
* [PR aiida-core](https://github.com/aiidateam/plumpy/pull/288)
* [PR plumpy](https://github.com/aiidateam/plumpy/pull/288)
* Bug fix sqlite dependency:
* version check on sqlite C-language [PR aiida-core](https://github.com/aiidateam/aiida-core/pull/6567)
* Merged PRs:
* Add filters to verdi group delete [PR](https://github.com/aiidateam/aiida-core/pull/6556)
* progress bar added for verdi storage maintain [PR](https://github.com/aiidateam/aiida-core/pull/6562) and [PR](https://github.com/aiidateam/disk-objectstore/pull/171)

Kristjan:
* AiiDA docs search indexing (https://github.com/aiidateam/aiida-core/issues/6516):
* proposed solution: https://github.com/aiidateam/aiida-core/issues/6516#issuecomment-2367640925
* more standard way of setting up readthedocs, standard config will work (like robots.txt to hide old versions from search)
* But as a result, the docs will be hosted in a new location:
* from: https://aiida.readthedocs.io/projects/aiida-core/en/stable/
* to: https://aiida.readthedocs.io/en/stable/
* (or optionally:) https://aiida-core.readthedocs.io/en/stable/

Julian:
* GP & I yesterday talked with a student who applied for an AiiDA-focused student project yesterday [link to offer on SiROP](https://sirop.org/app/145359bd-c1d7-4b54-b57b-b70fc7d51d9f?_k=nHBXhZ73r8K0Xs41)
* Current projects we proposed to him via [this presentation](https://docs.google.com/presentation/d/12S1q1cO_xSXDtNDpSV_LaZfIHvCJHnAjANqpklcPyZg/edit?usp=sharing) (feel free to add more!!)
* 1. JS/TS: New AiiDA provenance explorer
* 2. ML: Integrated, local LLM for AiiDA components
* 3. SQL: Performance optimizations at the ORM level (e.g. bulk operations)
* 4. DevOps: AiiDA(lab) deployment via Microsoft Azure CycleCloud (rather than self-hosted )
* 5. Data management: Stashing features in AiiDA
* [[#6565] Refactor: Add the `_prepare_yml` method to `AbstractCode`](https://github.com/aiidateam/aiida-core/pull/6565): Tests were failing for `orm.PortableCode` due to missing `filepath_files` attribute. Now, we dump the node repository contents in a directory and monkeypatch the instance, adding the attribute on-the-fly (after discussion with SPH).
* GP: Add those as the additional files in the return dictionary
* Blog post ["Improvements in the ways to get your data out of AiiDA"](https://github.com/aiidateam/aiida-website/pull/72): `verdi process dump` and new QB syntax. Scheduled for 2024-10-04 (biweekly (at least for now) posts on Fridays) -> Should be ready for merge. Will publish tomorrow. [Document for blog post planning](https://hackmd.io/_Ic1IR6pQ6yZYnqZUflMjg)
* [[#6578] Add `--also-remote` option to `verdi process dump`](https://github.com/aiidateam/aiida-core/pull/6578): First implementation.
* [[#6570] DRAFT: Add `verdi storage mirror` and `raw/rich data dumping`](https://github.com/aiidateam/aiida-core/pull/6570): Dogfooding in next MSD GM.
* MBx: Check information on the total file size of the remote files that should be copied
* GP: First, check total size of files recursively for the whole WorkChain, to obtain the full amount of data
* One thing noted together with Ali: In the test for exporting Data nodes instantiates the various classes and returns the instances, rather than loading from the DB. Usually one would export nodes that are stored in the DB and then loaded?
* GP: Parametrize with instance and with loading, to test both cases

Alex:
* Benchmarking running aiida-workgraph (with the help Jason, Xing), serialization is slow but we can improve it
* Debug `verdi process kill`
* https://github.com/aiidateam/aiida-core/pull/6572
* https://github.com/aiidateam/aiida-core/pull/6575 (Ali is mainly on it)
* Debug process is running while executing already next step https://aiida.discourse.group/t/workchain-continues-before-finishing-the-pervious-step/472
* Created `pixi.toml` initializing aiida profile with isolated rabbitmq and psql

### Specific Agenda Items