V 5

[ricardonpa]

Hi @amkrajewski ,

Please check the new modifications to ULTERA-contribute:

• Added new ULTERA manual (Contributor's Guide)
• Simplified README by moving excess content to the manual
• Added csv2README to automatically update the README if csv files exist
• Added new action to notifying new contributions*

• This last modification was implemented in replacement of automatically notifying the ULTERA-maintainers team/ opening a PR to merge new contributions committed to fork repositories. Unfortunately I couldn't make any of these two approaches to work entirely because of permission issues. This way the user is instructed to notify about new contributions by creating a new issue here at ULTERA-contribute, which conversely notify the ULTERA-maintainers team to run the ingestion pipeline.

[amkrajewski]

Hi @ricardonpa, great work! I have a bunch of comments to discuss and modify some things before we can merge, considering some indirect implications like backward compatibility, but it, again, looks great :)

Some quick notes for future direction:

• We probably want to have an action that will generate the manual based on the ULTERA repo once we make that one public. Some tables and text should be live updated based on the database content and rulesets.

• Template V5 may break the upload scheme for collections assigned "parent database", which expects the "Parent Database" name in place of the "Direct Fetch" marker (automatically set to False), which seems to be removed.

• The automatic PR feature, as you noted, won't work because of the permissions. What we want is either (a) bot looking for mentions (akin to JOSS) with write to ULTERA-push, or (b) a CRON job Action within ULTERA-push checking linked submodules for updates once a day and opening PRs to be approved (akin to arXiv). I'm leaning towards the latter as it is still elegant but has fewer moving parts to break.

[amkrajewski]

@ricardonpa I just wanted to bump this issue, so we keep it on the radar. Thanks again for these awesome revisions - we should integrate them fairly soon (after crystALL).

A couple of quick tasks I see in regard to template V5, to throw around before I forget:

• Rename it to not include the version number, since we now have backwards compatibility.
• Synchronize the known phases,
• approved property names (with minimum/maximum/estimated qualifiers versions) from YAML in manual,
• units (also in the properties table YAML in manual repo,
• processing techniques; so that users can see everything recognized in the database.
• Then, include the examples in auto-suggestions for the above fields.
• Adjust the dragging behavior, so that (1) nicknames do increase as you drag, (2) compositions do not, (3) structures do not (e.g. B2 now becomes B3), (4) comments do not, [property fields are already fine!], (5) pointers do not, and (6) DOI do not.
• Add suggestion to delimit parts of comment with ; and avoid , as best practice. Also, add key information for the processes, like annealed at 1273K for 24h in Ar; water quenched, HIPed at 1473K under 150MPa for 2h then annealed at 873K for 24h, and after VAM cold-rolled with 80% reduction then recrystallized at 1000K for 5min and quenched

[amkrajewski]

We also need to sit down for a moment and decide how to merge the conflicts with main (not too many)