Our GraphRAG-SDK version 0.2.0 now supports Ollama and Azure OpenAI, expanding deployment options for graph-based RAG applications.
This integration enables developers to leverage local LLM deployments through Ollama or enterprise-grade Azure OpenAI services while maintaining FalkorDB’s efficient graph storage and retrieval capabilities.
TL;DR
Local LLM Support
- Run GraphRAG workloads entirely on-premises using Ollama
- Reduce latency and maintain data privacy
```python
from graphrag_sdk.models.ollama import OllamaGenerativeModel
# Initialize your Ollama preferred model
model = OllamaGenerativeModel(model_name="preferred_model")
```
Enterprise Integration
- Seamless connection to Azure OpenAI services
- Production-ready deployment options for enterprise environments
```python
from graphrag_sdk.models.azure_openai import OpenAiGenerativeModel
# Initialize your Azure OpenAI preferred model
model = AzureOpenAiGenerativeModel(model_name="preferred_model")
```
The update streamlines the implementation of graph-based retrieval augmented generation (RAG) systems, particularly beneficial for AI developers working with large document collections requiring sophisticated knowledge extraction and querying capabilities.
How to use our GraphRAG-SDK
For Knowledge Management Systems
Document Processing
- Convert enterprise documentation into queryable knowledge graphs
- Enable complex queries across multiple documents while maintaining relationships between concepts
- Reduce token usage by 26-97% compared to traditional RAG approaches
For Data Integration Projects
Flexible Source Processing
- Ingest data from URLs, CSV files, and JSON sources
- Automatically generate ontologies for different domains
- Connect multiple knowledge graphs for comprehensive analysis
For Enhanced RAG Applications
Improved Query Understanding
- Capture complex relationships between entities
- Provide explainable and verifiable responses
- Enable visual debugging of query results
The SDK is particularly suitable in scenarios requiring complex reasoning across large datasets while maintaining high performance and accuracy. Its multi-model support and flexible deployment options make it suitable for both cloud and on-premises implementations.
Check it out here, happy coding: https://github.com/FalkorDB/GraphRAG-SDK/releases