6 14 2021 Tech Team Report - QualitativeDataRepository/TechnicalTeam GitHub Wiki

6-14-2021

Date	Task	Hours (Main)	Hours (EOLS)	Hours (PII)
7-Jun-2021	Deploy D8/Dataverse to prod, fix flyway script, update archivcalcopylocations, report, meeting	2
8-Jun-2021	Investigate guestbookresponses, change user …/traces api to allow element-level downloads in json/csv, investigate reason for missing datasetversion info in some responses.	3

Operations:

Deployed v5.5 and Drupal (sec update) to prod. In doing this roll-out, I realized our flyway sql script for anonymized access was not being triggered properly, so I made a fix and redeployed across dev, stage, and prod. I also went ahead and updated the existing archivalcopylocations for archived datasets on prod to point to the folder where the datacite.xml and zipped bag files exist.

Dataverse:

Looked into the options for exploring downloads per person in Dataverse. There is a general mechanism in the UI (Dataverse Edit/Data Project Guestbooks/Download All Responses) to get a report on all downloads from a Dataverse collection.
I also found a new /traces API intended to help with disabling/deleting users (discovering what they've done in Dataverse) that I extended to allow just retrieving their guestbook entries (or the other traces they have such as roles, etc.) This gives a per-person mechanism in the API to discover download activities.

Anno-Rep work
- Compare w/ Hypothesis selectors
Dataverse - in looking at guestbook responses, I noticed that some entries don't include the version used. I started looking into this, but I don't yet know if this is unique to QDR or a bug/feature of vanilla Dataverse. While knowing the dataset version isn't too relevant (given that the file doesn't change between versions) I'd like to figure this out. (My initial checking doesn't show any obvious difference where QDR's customizations could cause this.)

Still TBD: