MAJW work 2023-08-14 to 2023-08-18
to : Martin Holmes
Minutes: 320
On Monday, started work on figuring out how best to rationalize the duplication of information across person records and event records. After some analysis, and the addition of a diagnostic which catches 58 cases where relation elements in the prosopography do not have analogues in the event records, I began working on a preprocessing step which will reconstruct at build time all the relation elements in person records, so that we don’t have to maintain them, but they’re available when we build the site.
On Tuesday I started looking at these cases, and did a deep dive on two, determining that their source is the original database itself. Created a brief report on these two and passed that to the team to investigate further. Once we know whether these are errors or simply cases where the same mention/event is used for multiple relationships, we can figure out what to do.
On Wednesday, got the feedback on those two cases, which are both errors, but of different types; fixed them in the source per AB’s instructions, and then enhanced the diagnostic output to make it easier to do the sort of deep dive that these cases will need, as well as updating the way our collection statistics are calculated.
On Friday, I added a new requested diagnostic to find cases where events have businesses specified but their primaries are only associated relationally; this works well (and I enhanced other diagnostics too, to add helpful links), but I don’t think it’s finding an actual problem. The problem is in the processing which harvests those businesses and presents them on a person page, so I’ve suppressed the display of businesses in those contexts.