# 2025-09-09 Simplify real-world workflows https://github.com/aiidateam/team-compass/issues/23 [toc] ## Common issues for all 3 workflows ### Computer/Code setup * [[# 6925] Add `verdi computer/code search` to query registries](https://github.com/aiidateam/aiida-core/pull/6925) * [mm] Improve wording for `verdi computer setup` * Default for options during `verdi computer setup` * Still prompt for it, but show default in square brackets, star, etc., with direct enter default is selected * Set `async_ssh` as default, due to automatic configuration? * [x] [[1037] `Feature`: Add automatic setup of QE binaries](https://github.com/aiidateam/aiida-quantumespresso/issues/1037) -> Timo PR draft * [ ] ??? Feature to `dump` computers & codes, and then re-import them later on -> Timo makes issue -> Maybe easier to "duplicate" | `verdi profile grab-resources <other-profile>` ### Builder interaction * [[25] Usability: Improve interactions with `Builder` instances](https://github.com/aiidateam/team-compass/issues/25) * custom_scheduler_commands, mpirun_extra_params, prepend_text * Auto-complete works up to `parameters` -> Check again * Schemas for builder? * `PwCalculation.get_builder()` -> Doesn't show anything, basically * Print full data? * `PwCalculation.get_builder_from_protocol()` -> Shows more, but only populated parts How to obtain a sort of schema (basically a dictionary with all (nested) namespaces): ```python from aiida_quantumespresso.workflows.pw.base import PwBaseWorkChain b = PwBaseWorkChain.get_builder() b._port_namespace.get_description() # you can see that we have the nested as in the workchain->...->calcjob. ``` ### Docs * [📚 Discussion regarding documentation.](https://github.com/aiidateam/aiida-quantumespresso/issues/1107) * Who takes care? * Review PRs by MBx? * [[# 1074] 📚 Add basic `PwBaseWorkChain` documentation](https://github.com/aiidateam/aiida-quantumespresso/pull/1074) (rendered [here](https://aiida-quantumespresso--1074.org.readthedocs.build/en/1074/howto/workflows/pw/base.html)) * [[# 934] 📚 Docs: Add first version of `PwRelaxWorkChain` how to](https://github.com/aiidateam/aiida-quantumespresso/pull/934) (rendered [here](https://aiida-quantumespresso--934.org.readthedocs.build/en/934/howto/workflows/pw/relax.html)) * `run` vs `submit` * First, quick way how to run something. Then, longer way / advanced section for (what do you need to do to run real stuff: remote computer, submit...). One central place for this. * Quick overview for setting up Computer / Code early on in the documentation. Three schemas, for a Computer setup, config, and code setup. Highlight YAML way as default over interactive? * [mb] User should not use `PwCalculation` directly, no `builder` protocols available, at least in get started. Then, How-To, shows to set it up a calculation from a CalcJob * Example: https://atomistic-cookbook.org/ (recipe of the day) * Blog posts? * Everything under one domain: * Know-how goes hand-to-hand, better to have it written also real-world stories * aiida-qe coding week on docs? ### Other - [ ] ??? aiida-pseudo simplify how to use custom pseudos, with custom cutoffs - [ ] (let's not do that for now) download missing pseudos on-the-fly - Things that can be achieved via QE, but could be simplified - [ ] Get pseudo family from UpfData (attribute/extras of UpfData?) - [ ] Get the group a node belongs to from the node ## Relaxation ## Wannier ## Phonon - [x] Where to set `queue_name`? - [ ] Setup `metadata` on WC -> All `CalcJob`s inherit from this. - [ ] `options` on top-level already can be specified and is set through to all `CalcJob`s *** - [ ] mcp done by Ali here: https://github.com/khsrali/aiida-mcp - `aiida-pseudo install -a` --- Blog posts until then?