A second bug I found shows old data on the bulk persons api.
3,020 Documents
but the bulk query returned
{‘id’: ‘p-027’, ‘slug’: ‘donald-trump’, ‘name’: ‘Donald Trump’, ‘category’: ‘politician’, ‘aliases’: , ‘shortBio’: “45th/47th President; flew on Epstein’s planes 7+ times, 16 black book numbers, signed EFTA into law”, ‘imageUrl’: ‘/images/persons/donald-trump.png’, ‘status’: None, ‘flightCount’: 3, ‘documentCount’: 12, ‘connectionCount’: 17, ‘emailCount’: 374}
This is related to the same underlying issue as your other report. The bulk /api/v1/persons endpoint caches document counts when the server first starts up, and those counts do not refresh until the next deployment. So if new documents get ingested between deploys, the bulk endpoint will show stale numbers.
The single-person endpoint (/api/v1/persons/:slug) has a separate but related bug where it caps at 200 documents because of a query limit.
Both are being fixed. The bulk endpoint will pull fresh counts from the database with proper cache invalidation, and the single endpoint will use a COUNT query instead of fetching documents and measuring the array length.
Appreciate the thorough testing. If you find anything else off with the API, keep posting.