Here is an updated production-style README for your V6 FAISS-native RAG system.

OmniBioAI Dev Hub — RAG V6

Production-grade Retrieval-Augmented Generation (RAG) system powering the OmniBioAI ecosystem documentation, architecture search, workflow discovery, and developer assistant APIs.

---

Features

• FAISS-native vector search
• Incremental indexing
• Ollama local embeddings + local LLM inference
• FastAPI API server
• Streaming responses (SSE)
• Chunk-level document retrieval
• Repository-wide multi-project indexing
• Fully local execution
• No OpenAI dependency
• Production-safe embedding normalization
• Hybrid-ready architecture
• V6 dimension consistency enforcement

---

Architecture

Repositories
     ↓
Document Loader
     ↓
Chunker
     ↓
Ollama Embeddings (768-d)
     ↓
FAISS Vector Index
     ↓
RAG Engine
     ↓
FastAPI API
     ↓
LLM Answer Generation

---

V6 Major Improvements

FAISS Native Retrieval

Previous versions used brute-force cosine scanning across vectors.

V6 uses:

faiss.IndexFlatIP

Benefits:

• 10–50x faster retrieval
• scalable search
• lower latency
• future ANN support

---

Embedding Consistency Fix

A major issue in previous builds was embedding mismatch.

Old Problem

Stage	Model	Dimension
Indexing	all-MiniLM-L6-v2	384
Querying	nomic-embed-text	768

This caused FAISS assertion failures:

AssertionError: d == self.d

---

V6 Fix

Now BOTH ingestion and retrieval use:

nomic-embed-text

Dimension:

768

This guarantees:

• stable retrieval
• no dimension mismatch
• deterministic FAISS behavior

---

Repository Structure

omnibioai-dev-hub/
│
├── api/
│   ├── main.py
│   └── routes/
│
├── rag/
│   ├── engine.py
│   └── control_plane.py
│
├── index/
│   └── vector_store.py
│
├── embeddings/
│   └── embedder.py
│
├── ingestion/
│   └── doc_loader.py
│
├── processing/
│   └── chunker.py
│
├── scripts/
│   └── build_index.py
│
└── data/

---

Requirements

Python

Recommended:

Python 3.11

---

Ollama

Install:

curl -fsSL https://ollama.com/install.sh | sh

---

Required Ollama Models

Embedding Model

ollama pull nomic-embed-text

Generation Models

Recommended:

ollama pull mistral

Optional:

ollama pull llama3
ollama pull deepseek-coder
ollama pull deepseek-r1

---

Installation

Create Environment

conda create -n chemoinfo python=3.11 -y
conda activate chemoinfo

---

Install Dependencies

pip install fastapi uvicorn requests numpy faiss-cpu sentence-transformers

---

Build Index

Clean Existing Data

rm -rf data/*

---

Build V6 Index

python scripts/build_index.py

Expected output:

🚀 Incremental V6 Indexing Starting...
✅ V6 Index Complete

---

Run API Server

uvicorn api.main:app --host 0.0.0.0 --port 8082 --reload

---

Test Query API

curl -X POST http://localhost:8082/rag/query \
-H "Content-Type: application/json" \
-d '{"query":"What is workflow engine in OmniBioAI?"}'

Example response:

{
  "query": "What is workflow engine in OmniBioAI?",
  "answer": "According to the provided context...",
  "sources": [
    "../omnibioai-workflow-bundles/README.md"
  ],
  "context_used": 5,
  "version": "v6-faiss",
  "api_version": "v6"
}

---

Streaming API

Endpoint:

POST /rag/stream

Uses:

• Server-Sent Events (SSE)
• token streaming
• real-time generation

---

V6 Retrieval Pipeline

Step 1 — Chunking

Documents are split into semantic chunks.

---

Step 2 — Embedding

Each chunk is embedded using:

nomic-embed-text

Output dimension:

768

---

Step 3 — FAISS Indexing

Vectors are stored in:

faiss.IndexFlatIP

---

Step 4 — Query Embedding

User query is embedded using the SAME embedding model.

---

Step 5 — Vector Search

FAISS retrieves nearest chunks.

---

Step 6 — Prompt Assembly

Retrieved chunks become context.

---

Step 7 — LLM Generation

Prompt sent to local Ollama model.

---

Supported Repositories

Current indexing targets:

repos = [
    "../omnibioai",
    "../omnibioai-rag",
    "../omnibioai-toolserver",
    "../omnibioai-sdk",
    "../omnibioai-workflow-bundles",
    "../omnibioai-control-center",
    "../omnibioai-lims",
    "../omnibioai-model-registry",
    "../omnibioai-dev-docker"
]

---

Performance

Before V6

• brute-force cosine scan
• slow retrieval
• dimension mismatch bugs
• unstable indexing

---

After V6

• FAISS-native retrieval
• stable dimensions
• fast semantic search
• local-only execution
• scalable architecture

---

Troubleshooting

FAISS Dimension Mismatch

Error:

AssertionError: d == self.d

Cause:

Different embedding models used during indexing vs querying.

Fix:

Rebuild index using the SAME embedding model.

---

Ollama Timeout

Error:

Read timed out

Fix:

Use a smaller generation model:

model="mistral"

instead of:

deepseek-r1

---

Empty Retrieval Results

Check:

python -c "
from index.vector_store import VectorStore
import numpy as np

vs = VectorStore()
vs.add([np.random.rand(768)], [{'text':'test'}])

print(vs.index.ntotal)
"

Expected:

1

---

Future Roadmap

Planned V7 Features

• IVF indexes
• HNSW search
• metadata filtering
• hybrid BM25 + vector search
• reranking
• cross-encoder scoring
• persistent FAISS storage
• multi-user collections
• distributed indexing
• workflow-aware retrieval
• graph RAG
• plugin-aware retrieval

---

License

Internal OmniBioAI Development License.

---

OmniBioAI Ecosystem

RAG V6 powers:

• architecture discovery
• workflow documentation search
• plugin documentation retrieval
• developer assistant APIs
• AI infrastructure exploration
• cross-repository semantic search
• internal engineering copilots