# AiiDA Team Meeting 2021-08-25

###### tags: `team meetings`
###### time: 10am CET

[TOC]

### Present

* Chris
* Simon
* Sebastiaan
* Marnik
* Sasha
* Giovanni

### Catch-up round

*Max. 3 minutes each*

Leopold:
* Archived reentry (see [archival notice](https://github.com/aiidateam/reentry#archival-notice))
* Pull requests: quick comment on the practive of assigning reviewers.
I've recently seen several cases where PR authors requested reviews from 3-4 reviewers (different authors, not pointing fingers here).
When my review is requested in this way, I'm always wondering what this means. Does the PR author really request that many people to review the PR or do they want one person to review?
If it is the latter, I would point out that reviewing a PR properly can take quite some time, and Github's interface does not notify you when someone else has started reviewing until they submit their review. I.e. several people may start doing this at the same time.
I would propose the following:
* If you want to have a review from *one* of a group, please say so in a comment and @mention them, asking them to self-assign the review if they take it on
* Only assign reviewers if you need/want a review from (all of) them specifically.
Even in that case, adding a friendly comment often helps moving things along faster (optional; just my 2 cents)
* Package by Paul Sax: track citations of codes used by workflow manager https://github.com/MolSSI/reference_handler
This uses a sqlite backend and is perhaps not that straightforward to integrate but the general idea makes a lot of sense to me
* [SEAMM](https://www.youtube.com/playlist?list=PLOTjazgLxvK2YEHDKL7mnf0FzMV3_Pjh1) workflow manager from MolSSI

Sebastiaan:
* **Releases**
* `aiida-core==1.6.5`: Backporting many fixes from `develop` on to the `v1.6.x` support branch.
* `aiida-pseudo==0.6.3`: Small fixes and updates the URLs for SSSP downloads to point to the new Materials Cloud Archive server, instead of the legacy URLs
* **Merged PRs**
* [CLI: add verbosity control](https://github.com/aiidateam/aiida-core/pull/5085)
* [CI: run certain workflows only on main repo, not on forks
](https://github.com/aiidateam/aiida-core/pull/5091)
* [Add the `JsonableData` data plugin](https://github.com/aiidateam/aiida-core/pull/5017)
* [Pre-commit: move yapf and mypy config to `pyproject.toml`](https://github.com/aiidateam/aiida-core/pull/4996)
* [`ProcessBuilder`: ensure instances do not interfere](https://github.com/aiidateam/aiida-core/pull/4984)
* **Open PRs**
* **[v2.0]** [Engine: implement functionality to immigrate completed `CalcJob`s ](https://github.com/aiidateam/aiida-core/pull/5086)
* **[v2.0]** [Entry points: add the `core.` prefix to all entry points ](https://github.com/aiidateam/aiida-core/pull/5073)
* [REST API: make the profile configurable as request parameter ](https://github.com/aiidateam/aiida-core/pull/5054)
* [Add the `ElectronicBandsData` data plugin](https://github.com/aiidateam/aiida-core/pull/5033)

Simon:

* [Updated the test-install GA workflow (#5082).](https://github.com/aiidateam/aiida-core/pull/5082)
* To check that all requirements/* files can be resolved from the Python Package Index.
* To check that we can install AiiDA with pip for all supported Python versions
* To check that we can install AiiDA with conda for all supported Python versions.
* Reminder: According to our [compatibility schedule](https://www.aiida.net/download/), we should drop support for NumPy 1.17 soon ([#5095](https://github.com/aiidateam/aiida-core/issues/5095)).
* Question: Who maintains the version history on https://www.aiida.net/download/ ? The list of releases is outdated. Should we maybe just link to https://github.com/aiidateam/aiida-core/releases/ instead?

Marnik:
* Nothing to report, captain!

Sasha:
* Was on holidays.

Chris:
* 🚀 Made v1.6.5 release (updated changelog on `develop`)
* ♻️ REFACTOR: Make `__all__` imports explicit (via pre-commit) (#5061)
* Move towards closing API clarification: https://github.com/aiidateam/aiida-core/issues/4558
* ♻️ Removed reentry and replaced with `importlib_metadata` (#5058)
* 👌 IMPROVE: CLI responsiveness (#5080)
* ♻️ Rewrote QueryBuilder
* 🔧 MAINTAIN: Make DjangoQueryBuilder a subclass of SqlaQueryBuilder (#5092)
* 👌 IMPROVE: QueryBuilder reprs (incl/ `queryhelp` -> `as_dict`) (#5081)
* ♻️ Fully abstracted backend [[#5093](https://github.com/aiidateam/aiida-core/pull/5093)]
* Dropped sqlalchemy-utils from requirements (still in `test` extras)
* 🔧 MAINTAIN: Update pylint version (& fix new issues) (#5087)
* 🔧 MAINTAIN: Add `verdi devel run-sql` (#5094)
* Open PRS:
* ✨ New query syntax (more object orientated) & programmatic retrieval of ORM class database fields:
* Sets fields on classes (`__fields__`) and methods (`@field_method`) (with inheritance)
* `QueryBuilder().append(Node, project=Node.fields.pk, filters=(Node.fields.pk <= 2) & (Node.fields.pk.in_([2, 3])))` [[#5088](https://github.com/aiidateam/aiida-core/pull/5088)]
* 👌 Add SQLA migration for parity with Django schema [#5097]
* ♻️ REFACTOR: Move to sqlachemy 1.4 (and v2 API) #5069

Giovanni:
* After discussion with Bonan, on the aiida-fireworks-scheduler, I proposed an simple implementation in [#5084](https://github.com/aiidateam/aiida-core/issues/5084) that might simplify the logic (remove the need to SSH to the computer just to fetch back the submission script). It's backward-compatible, and might help beyond the fireworks scheduler, with anything along the same lines (let some external engine re-arrange the submitted jobs and possibly batch them). Any feedback?

### Specific Agenda Items