redevops-rag

Hybrid RAG as a small, installable library + CLI — **DuckDB vector + BM25 fused via Reciprocal Rank Fusion, recency & keyword priors, and an optional cross-encoder rerank.**

It's the retrieval pipeline from redevops-io/rag-saas-platform carved out of its multi-tenant SaaS shell (no Auth0/Stripe/Kubernetes/workspace coupling) so you can drop the same retrieval over any folder — a docs tree, a repo, an Obsidian vault — in three lines.

Pipeline

query
  ├─ dense   : sentence-transformers embedding → DuckDB arraycosinesimilarity (threshold 0.4)
  └─ sparse  : DuckDB FTS BM25
        └─ Reciprocal Rank Fusion   score = Σ 1 / (k + rank),  k = 60
              └─ recency prior       0.5 ** (age_days / 90)
              └─ keyword prior       ×(1 + 0.05·term_hits), capped 1.5
                    └─ (optional) cross-encoder rerank  BAAI/bge-reranker-v2-m3  → top-k

Install

Not on PyPI yet — install from git:

pip install "git+https://github.com/redevops-io/redevops-rag.git"                          # core
pip install "redevops-rag[rerank] @ git+https://github.com/redevops-io/redevops-rag.git"   # + cross-encoder rerank
pip install "redevops-rag[llm]    @ git+https://github.com/redevops-io/redevops-rag.git"   # + answer synthesis

CLI

redevops-rag index ~/obsidian-vault              # chunk + embed + index into ./redevops_rag.duckdb
redevops-rag search "how do we rotate API keys"  # hybrid search, top chunks
redevops-rag --rerank search "..."               # add the cross-encoder rerank stage
redevops-rag ask "what's our incident process?"  # search + synthesized answer (needs an LLM, below)

Answer synthesis (ask) talks to any OpenAI-compatible endpoint — a local MLX/llama.cpp server, OpenAI, or Anthropic behind a gateway:

export REDEVOPSRAGLLMBASEURL=http://localhost:8080/v1   # e.g. a Mac running mlx_lm.server
export REDEVOPSRAGLLM_MODEL=DeepSeek-V4-Flash
export REDEVOPSRAGLLMAPIKEY=EMPTY                       # or sk-... for a cloud endpoint
redevops-rag ask "summarize our on-call runbook"

Library (the 3 lines)

from redevops_rag import RAG
rag = RAG(dbpath="vault.duckdb")          # add usereranker=True for the cross-encoder stage rag.index("~/obsidian-vault")              # chunk + embed + index (incremental: re-run anytime) hits = rag.search("zero-downtime deploys", k=8) 
answer = rag.ask("zero-downtime deploys")["answer"]   # if an LLM env is set

Why a folder, not a sync

Point it at one central copy of the knowledge base and query it — you don't copy 200k files to every machine, you index once and retrieve. For a team, run it on one box (or behind a thin service) so everyone hits a single index. Configuration / skills / CLAUDE.md are small — those belong in git, not in a RAG.

Configuration

| env | default | meaning | |-----|---------|---------| | REDEVOPSRAGEMBED_MODEL | BAAI/bge-small-en-v1.5 | sentence-transformers embedding model | | REDEVOPSRAGRERANK_MODEL | BAAI/bge-reranker-v2-m3 | cross-encoder for --rerank | | REDEVOPSRAGLLMBASEURL / MODEL / API_KEY | — | OpenAI-compatible endpoint for ask |

Status: v0.1, not yet large-corpus benchmarked. The retrieval logic is a faithful port

of the production pipeline; the packaging/CLI are new. Validate on your own data.

AGPL-3.0-or-later · redevops.io