fix: atomic ingestion persist with task reclaim and shared cleanup.

One transaction per task replaces prior artifact rows; workers skip the pipeline when content already exists, eval seeding reuses persist_artifacts, and deletes clear graph children via shared SQL.
This commit is contained in:
Per Stark
2026-06-12 16:27:07 +02:00
parent 4947d48ecf
commit 189c7d7c31
15 changed files with 794 additions and 487 deletions
+4
View File
@@ -1,7 +1,11 @@
# Changelog
## Unreleased
- Performance: ingestion skips per-task index rebuild; worker runs scheduled `REBUILD INDEX` (default every 24h via `index_rebuild_interval_secs`, `0` disables)
- Performance: ingestion persists all artifacts in a single SurrealDB transaction per task (atomic replace by task id)
- Fix: ingestion reclaims tasks after a successful persist without re-running the pipeline when `mark_succeeded` failed
- Fix: content deletion clears graph relationships via shared `TextContent::clear_ingested_children`
- Fix: regression re suggestion of relationships
- Internal: eval corpus DB seed uses `persist_artifacts` instead of a separate batched insert path
## 1.0.3 (2026-06-12)
- Search: filter results by type — knowledge entities, ingested content, or both