Complete outstanding ingestion and document UX work
This commit is contained in:
@@ -1,5 +1,10 @@
|
||||
# kb — Next Steps
|
||||
|
||||
> Implementation status (2026-08-22): `document_id`, metadata-only info,
|
||||
> title/filename filters, document-level `find`, and short Docling chunk
|
||||
> filtering are implemented on `feature/tasks-5-15-completion`. OCR is already
|
||||
> enabled through RapidOCR; further OCR tuning remains measurement-gated.
|
||||
|
||||
UX improvements to make documents easier to find and inspect, prompted by a session where searching for an uploaded PDF (`M38T_PHEV_RHD_OM_EN_UK_20251209.pdf`, doc id 2077, 1801 chunks) surfaced lots of chunk hits but no obvious path back to the original document.
|
||||
|
||||
## Problems observed
|
||||
|
||||
Reference in New Issue
Block a user