Alex
2b6d4d509e
fix(graphrag): return a chunk once from entity_pages
...
The subject flag was in the GROUP BY, so a chunk two matching entities link --
one the page is about, one merely mentioned in it -- came back as two
identical pages and spent the caller's page budget twice on the same text.
It is aggregated with bool_or now, which is what the ordering wanted anyway.
Covered by a live test against a pgvector-shaped documents table: the graph
tables alone cannot answer this query, so nothing exercised it before.
2026-09-20 10:45:04 +01:00
..
2026-09-20 10:45:04 +01:00
2026-09-07 10:20:43 +01:00
2026-09-17 16:51:06 +01:00
2026-09-19 14:07:41 +01:00
2026-09-17 13:34:59 +01:00
2026-09-17 11:14:34 +01:00
2026-09-14 10:19:59 +01:00
2026-09-20 10:45:04 +01:00
2026-09-17 11:14:34 +01:00
2026-09-17 11:14:34 +01:00
2026-09-17 11:14:34 +01:00
2026-09-07 10:20:43 +01:00
2026-09-20 10:45:03 +01:00
2026-09-17 11:14:34 +01:00
2026-09-17 11:14:34 +01:00
2026-09-12 17:44:18 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-19 14:07:41 +01:00
2026-09-14 12:52:15 +01:00
2026-09-15 17:54:24 +01:00
2026-09-07 10:20:43 +01:00
2026-09-15 17:54:24 +01:00
2026-09-07 10:20:43 +01:00
2026-09-19 14:07:42 +01:00
2026-09-07 10:20:43 +01:00
2026-09-16 22:45:54 +01:00
2026-09-07 10:20:43 +01:00
2026-09-17 11:09:09 +01:00
2026-09-14 08:33:46 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-19 14:33:14 +01:00
2026-09-07 12:02:07 +01:00
2026-09-17 11:32:50 +01:00
2026-09-15 22:41:53 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-12 18:21:33 +01:00
2026-09-12 18:15:03 +01:00
2026-09-12 18:15:03 +01:00
2026-09-09 11:09:09 +01:00
2026-09-07 10:20:43 +01:00
2026-09-07 10:20:43 +01:00
2026-09-17 11:14:34 +01:00
2026-09-16 14:23:26 +01:00
2026-09-14 17:27:13 +01:00
2026-09-07 10:20:43 +01:00