Commit graph

8 commits

Author SHA1 Message Date
mateo-berri
c15f4e066f fix(rag): let the managed store's params win over caller kwargs on the search call 2026-09-02 18:59:28 -07:00
mateo-berri
91061675ae test(rag): wrap the store-forwarding patches so every new line fits in 120 chars 2026-09-02 18:04:52 -07:00
mateo-berri
8ed1da40de fix(rag): forward the managed vector store's params to the search call
/v1/rag/query folded the registry store's litellm_params into retrieval_config,
where the caller allowlist dropped api_key, api_base, and provider extras such
as Milvus outputFields and milvus_text_field, so a managed Milvus store 500'd
with MILVUS_API_KEY is not set while the direct search endpoint worked. The
store's params now travel as a trusted vector_store_params argument straight to
the search call, never through the completion kwargs, and the caller allowlist
stays in place.
2026-09-02 17:55:02 -07:00
mateo-berri
babe7816ad fix(rag): store-wins merge, single lookup, allowlisted search params
rag_query reuses the store resolved during authorization instead of a
second registry lookup, merges registry data store-wins so callers
cannot override a managed store's provider or credentials, and logs ids
instead of the merged config, which can carry resolved credentials.
aquery forwards only allowlisted retrieval_config keys to vector store
search, keeping caller-supplied connection overrides like api_base and
api_key away from the search call
2026-09-01 12:52:45 -07:00
mateo-berri
753f3705d0 Merge litellm_internal_staging (4c3ef9ae0a) into litellm_fix_s3_vectors_search 2026-09-01 12:27:23 -07:00
Devin AI
cb70941660 fix(tests): drain the global logging worker in RAG aquery billing tests instead of polling
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-08-28 09:58:22 +00:00
michelligabriele
a1514efa21
fix(vector_stores): S3 Vectors search router bypass + rag query config drop + UI error swallow 2026-07-27 17:07:58 +02:00
Yassin Kortam
215ce9f7c1
fix(rag): track LLM completion usage and spend for /v1/rag/query (#32438) 2026-07-17 17:45:27 +00:00