Record the preview backfill in the state file

This commit is contained in:
Fisto
2026-09-16 14:02:18 +02:00
parent 96fe17b4c9
commit c100bc269c
+6
View File
@@ -40,6 +40,12 @@ mobile draft at 390px.
- [x] C11 preview caching: Discord-hosted image/GIF previews (WebP 320px, hash-addressed, 5 MiB - [x] C11 preview caching: Discord-hosted image/GIF previews (WebP 320px, hash-addressed, 5 MiB
cap, one attempt) run after every walk in both modes. Live run `...-20260916090722`: cap, one attempt) run after every walk in both modes. Live run `...-20260916090722`:
`previewed: 2`; thumbs served over `/api/media/<scrape>/<item>/thumb` (200). `previewed: 2`; thumbs served over `/api/media/<scrape>/<item>/thumb` (200).
- [x] C26 YouTube preview root cause. Dedupe picked the newest row per URL; a paused scrape
(`gernal_chut`, stopped mid-walk) had no previews yet and hid an older scrape that did. Fixes:
dedupe prefers the richer row (downloaded > preview > newer), paused walks still run the
download/preview passes for collected pages, daemon start backfills previews for scrapes that
paused early, and preview gaps are now 200-600 ms. Live: backfill cached 24 previews in the
paused scrape within 45 s and a YouTube item then served `thumbUrl`.
- [x] C25 Live scrape progress. The media tab polls `/api/media/scrapes` every 4 s while open, - [x] C25 Live scrape progress. The media tab polls `/api/media/scrapes` every 4 s while open,
so the running row counts update; pages show `X/Yp (~Z%)` from `estimatedMessages` when the so the running row counts update; pages show `X/Yp (~Z%)` from `estimatedMessages` when the
search total is available (some channels deny the bot search with Missing Access, so it stays search total is available (some channels deny the bot search with Missing Access, so it stays