Handling in-press publications needs its own automated system

[krzywon]

Currently, there is no easy way to update publications when they go from 'In Press' to fully published. The current method is to add a 2nd copy of the same publication and delete the previous. The original is often not removed from publications.md. To remove the item from the list, the pages and json files associated with the page need to be deleted from the server performing the updates.

Instead, I propose an automated method to look for any updates to the list. This could also run as a cron on the server, but could be much less frequent (daily/weekly?) than the current process.

[krzywon]

As noted by @butlerpd in SasView ticket 617, this should extend beyond auto-updates and incorporate some way to handle duplicates as well.

[krzywon]

I have been working on this in branch auto-updater, and can delete duplicates, but I'm still working on updating publications that have changed (e.g. from in press to having a volume and page number).

[krzywon]

I created a test library in zotero that is a subset of the main sasview library. The latest commits to my branch are able to update entries with changes directly from the publisher and delete duplicates in that library with no apparent issues.

These changes will require all publications be within a group, not the top level library. This will also change how publications are added.

Implementation steps:

• Create a sasview-bot zotero account and get access to sasview
• Create an API key for remote editing of the sasview library
• Stop auto-update cron to prevent conflicts
• Restart auto-update cron after doing the following:
• • Merge branch into saspubs master
• • Backup sasview pubs in test library
• • Create sasview group in library and move all pubs into the group - this will add a second of each item in the library
• • Delete or merge duplicates in library
• • Modify config.py with new group
• • Pull suspubs updates on DANSE server
• • Test run scripts on DANSE server
• • Update cron jobs to find new entries every 30 minutes and find duplicates or changes to existing items once a day (using API key)
• Update publication editing site
• Email current library editors with changes

[krzywon]

I've tested the new scripts against a mirror database of the sasview publications and the updating worked well, but the one duplicate I introduced was not deleted. A little work is needed before I turn the crons back on.

[krzywon]

I found the issue removing duplicates and tested it on the server. I've started both recurring jobs pointing to the original sasview zotero database.