Skip to content

External RAG#52

Open
cableman wants to merge 6 commits into
upstream/v0.9.6from
feature/external-ingest-rag
Open

External RAG#52
cableman wants to merge 6 commits into
upstream/v0.9.6from
feature/external-ingest-rag

Conversation

@cableman

Copy link
Copy Markdown
  • Moves ingestion with text extraction, embedding, chunking and vector store handling to external service.
  • Moves retrieval to external service

See https://github.com/AarhusAI/ingestion-service and https://github.com/AarhusAI/retrieval-agent.

@cableman cableman changed the base branch from upstream/v0.8.12 to upstream/v0.9.5 May 27, 2026 13:28
@cableman cableman force-pushed the feature/external-ingest-rag branch from 08fe31b to fa5a32d Compare May 28, 2026 10:44
cableman added 4 commits June 3, 2026 16:09
When EXTERNAL_INGESTION_ENGINE=external, process_file() delegates the
chunk + embed + vector-store step to an external HTTP service via
PUT {EXTERNAL_INGESTION_URL}/api/v1/ingest. Two transport modes:
S3-reference body when file.path starts with s3://, multipart fallback
otherwise. Default off; production unaffected until the env var is
flipped.

Restricted to the fresh-file path; pre-extracted content
(form_data.content) and knowledge-base re-add (form_data.collection_name)
keep the in-process pipeline as a known limitation.
@cableman cableman force-pushed the feature/external-ingest-rag branch from fa5a32d to 7275a9b Compare June 3, 2026 14:15
@cableman cableman changed the base branch from upstream/v0.9.5 to upstream/v0.9.6 June 3, 2026 14:38
@cableman cableman force-pushed the feature/external-ingest-rag branch from 7275a9b to 5b1e11d Compare June 4, 2026 07:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants