RAG Templates - Enterprise RAG Framework

Production-ready RAG applications with InterSystems IRIS. Zero-configuration APIs, enterprise-grade architecture, and seamless LangChain integration.

🚀 Quick Start

Python - Zero Configuration

from rag_templates import RAG

# Works immediately - no setup required
rag = RAG()
rag.add_documents(["Your documents here"])
answer = rag.query("What is machine learning?")
print(answer)

JavaScript - Zero Configuration

import { RAG } from '@rag-templates/core';

const rag = new RAG();
await rag.addDocuments(["Your documents here"]);
const answer = await rag.query("What is machine learning?");
console.log(answer);

ObjectScript Integration

// Direct IRIS integration via Embedded Python
Set bridge = ##class(RAG.PythonBridge).%New()
Set result = bridge.Query("What is machine learning?", "basic")
Write result.answer

🏗️ Core Architecture

Schema Manager

Centralized schema management with automatic migration support:

• Universal dimension authority for all vector tables
• Automatic schema detection and migration
• Customizable table names and field configurations
• Version tracking and rollback capabilities

IRISVectorStore Interface

LangChain-compatible vector store with enterprise features:

from rag_templates.storage import IRISVectorStore

# Drop-in LangChain replacement
vector_store = IRISVectorStore(connection_manager, config_manager)
retriever = vector_store.as_retriever(search_kwargs={"k": 5})

Configuration System

Environment-aware configuration with validation:

from rag_templates.config import ConfigurationManager

config = ConfigurationManager()
# Supports RAG_DATABASE__IRIS__HOST env vars
# Automatic YAML loading with schema validation

🛠️ Available RAG Techniques

Technique	Description	Best For	Status
basic	Standard vector similarity	General purpose, fast queries	✅ Production
colbert	Token-level embeddings with MaxSim	High precision retrieval	✅ Production*
crag	Corrective RAG with self-correction	Accuracy-critical applications	✅ Production
hyde	Hypothetical Document Embeddings	Complex reasoning tasks	✅ Production
graphrag	Graph-based knowledge retrieval	Structured knowledge bases	✅ Production
hybrid_ifind	Multi-modal search combination	Enterprise search	✅ Production
noderag	Node-based structured retrieval	Hierarchical data	✅ Production

*ColBERT: Includes experimental Pylate integration with pluggable backend support (native/pylate).

🎯 Developer Experience

Three-Tier API Design

Simple API - Zero configuration for prototypes:

rag = RAG()  # Works immediately

Standard API - Production configuration:

rag = ConfigurableRAG({
    'technique': 'colbert',
    'llm_provider': 'openai'
})

Enterprise API - Full control:

config = ConfigManager.from_file('enterprise-config.yaml')
rag = ConfigurableRAG(config)

Key Configuration Options

Variable	Description	Default
RAG_DATABASE__IRIS__HOST	IRIS database host	localhost
RAG_EMBEDDING__MODEL	Embedding model name	all-MiniLM-L6-v2
RAG_LLM__PROVIDER	LLM provider	openai
RAG_TECHNIQUE	RAG technique to use	basic

🔗 MCP Integration

The Model Context Protocol (MCP) integration allows you to easily deploy and manage RAG services as "microservices". This design enables flexible deployment across various environments and seamless integration with existing enterprise systems.

Creating MCP Servers

Create MCP servers in minutes:

import { createMCPServer } from '@rag-templates/mcp';

const server = createMCPServer({
    name: "knowledge-server",
    description: "Company knowledge base",
    ragConfig: { technique: 'graphrag' }
});

server.start();

IRIS SQL Tool Integration

The IRIS SQL tool provides direct SQL access and advanced vector search capabilities within your MCP-deployed RAG services. It allows for efficient data manipulation and retrieval directly from InterSystems IRIS databases, leveraging its powerful SQL and vector functionalities. This integration streamlines data management for RAG applications, enabling complex queries and high-performance vector lookups.

For detailed setup and usage, refer to the MCP Integration Guide.

📚 Documentation

Guide	Description
📖 User Guide	Complete usage guide and best practices
🔗 MCP Integration Guide	Multi-Cloud Platform integration, MCP server creation, and IRIS SQL tool usage
📋 Documentation	Additional documentation and guides

✅ Verification

# Quick setup and validation
make setup-env && make install
make validate-iris-rag && make test-unit

# Full end-to-end testing with 1000+ documents
make load-1000 && make test-1000

# Performance benchmarking
make test-ragas-1000-enhanced

🌟 Key Features

• Zero Configuration: Production-ready defaults, works immediately
• Enterprise Architecture: Schema management, migrations, monitoring
• LangChain Compatible: Drop-in replacement for existing workflows
• Multi-Language: Python, JavaScript, and ObjectScript support
• MCP-First Design: Trivial MCP server creation
• Advanced RAG: 7+ sophisticated retrieval techniques
• Performance Optimized: Built-in caching and optimization

📚 References & Research

RAG Technique Papers & Implementations

Technique	Original Paper	Key Repository	Additional Resources
Basic RAG	Retrieval-Augmented Generation	Facebook Research	LangChain RAG
ColBERT	ColBERT: Efficient and Effective Passage Retrieval	Stanford ColBERT	Pylate Integration
CRAG	Corrective Retrieval Augmented Generation	CRAG Implementation	LangGraph CRAG
HyDE	Precise Zero-Shot Dense Retrieval	HyDE Official	LangChain HyDE
GraphRAG	From Local to Global: A Graph RAG Approach	Microsoft GraphRAG	Neo4j GraphRAG
NodeRAG	Hierarchical Text Retrieval	NodeRAG Implementation	Hierarchical Retrieval

Core Technologies

• Vector Databases: InterSystems IRIS Vector Search
• Embeddings: Sentence Transformers, OpenAI Embeddings
• LLM Integration: LangChain, OpenAI API
• Evaluation: RAGAS Framework

🤝 Contributing

We welcome contributions! See our Contributing Guide for details.

📄 License

MIT License - see LICENSE file for details.

---

Transform your RAG development from complex to enterprise-ready. Start building in minutes, scale to production.