5.4 KiB
TrustGraph
🚀 Get Started 💬 Join the Discord 📖 Read the Blog 📺 YouTube
Introduction
TrustGraph deploys a full E2E (end-to-end) AI solution with native GraphRAG in minutes. TrustGraph is designed for maximum flexibility and modularity whether it's calling Cloud LLMs or deploying SLMs On-Device. TrustGraph ingests data to build a RDF style knowledge graph to enable accurate and private RAG responses using only the knowledge you want, when you want.
The pipeline processing components are interconnected with a pub/sub engine to maximize modularity for agent integration. The core processing components decode documents, chunk text, create mapped embeddings, generate a RDF knowledge graph, generate AI predictions from either a Cloud LLM or On-Device SLM.
The processing showcases the reliability and efficiences of GraphRAG algorithms which can capture contextual language flags that are missed in conventional RAG approaches. Graph querying algorithms enable retrieving not just relevant knowledge but language cues essential to understanding semantic uses unique to a text corpus.
Deploy in Minutes
TrustGraph is fully containerized and is launched with a Docker Compose file. These files have already been prebuilt. Simply select a file that matches your desired model deployment and graph store configuration.
| Model Deployment | Graph Store | Launch File |
|---|---|---|
| AWS Bedrock | Cassandra | tg-launch-bedrock-cassandra.yaml |
| AWS Bedrock | Neo4j | tg-launch-bedrock-neo4j.yaml |
| AzureAI Serverless Endpoint | Cassandra | tg-launch-azure-cassandra.yaml |
| AzureAI Serverless Endpoint | Neo4j | tg-launch-azure-neo4j.yaml |
| Anthropic API | Cassandra | tg-launch-claude-cassandra.yaml |
| Anthropic API | Neo4j | tg-launch-claude-neo4j.yaml |
| Cohere API | Cassandra | tg-launch-cohere-cassandra.yaml |
| Cohere API | Neo4j | tg-launch-cohere-neo4j.yaml |
| Llamafile | Cassandra | tg-launch-llamafile-cassandra.yaml |
| Llamafile | Neo4j | tg-launch-llamafile-neo4j.yaml |
| Mixed Depoloyment | Cassandra | tg-launch-mix-cassandra.yaml |
| Mixed Depoloyment | Neo4j | tg-launch-mix-neo4j.yaml |
| Ollama | Cassandra | tg-launch-ollama-cassandra.yaml |
| Ollama | Neo4j | tg-launch-ollama-neo4j.yaml |
| OpenAI | Cassandra | tg-launch-openai-cassandra.yaml |
| OpenAI | Neo4j | tg-launch-openai-neo4j.yaml |
| VertexAI | Cassandra | tg-launch-vertexai-cassandra.yaml |
| VertexAI | Neo4j | tg-launch-vertexai-neo4j.yaml |
Launching TrustGraph is as simple as running one line:
docker compose -f <launch-file> up -d
Core TrustGraph Features
- PDF decoding
- Text chunking
- On-Device SLM inference with Ollama or Llamafile
- Cloud LLM infernece:
AWS Bedrock,AzureAI,Anthropic,Cohere,OpenAI, andVertexAI - Chunk-mapped vector embeddings with HuggingFace models
- RDF style Knowledge Graph extraction
- Apache Cassandra or Neo4j as the graph store
- Qdrant as the VectorDB
- Build and load Knowledge Cores
- GraphRAG query service
- Grafana telemetry dashboard
- Module integration with Apache Pulsar
- Container orchestration with
Dockeror Podman
Architecture
TrustGraph is designed to be modular to support as many Language Models and environments as possible. A natural fit for a modular architecture is to decompose functions into a set of modules connected through a pub/sub backbone. Apache Pulsar serves as this pub/sub backbone. Pulsar acts as the data broker managing data processing queues connected to procesing modules.
Pulsar Workflows
- For processing flows, Pulsar accepts the output of a processing module and queues it for input to the next subscribed module.
- For services such as LLMs and embeddings, Pulsar provides a client/server model. A Pulsar queue is used as the input to the service. When processed, the output is then delivered to a separate queue where a client subscriber can request that output.
Naive Knowledge Extraction
TrustGraph extracts knowledge from a text corpus (PDF or text) to a knowledge graph using 3 parallel processes. These processes focus on individual elements needed to build a rich RDF knowledge graph. The extraction focuses on:
- Topics
- Entities
- Semantic Relationships
The extraction prompts are built through templates, enabling customized extraction processes for a specific use case.
GraphRAG Queries
Once the knowledge graph has been built or a knowledge core has been loaded, GraphRAG queries are launched with a single line:
scripts/query-graph-rag -q "Write a blog post about the 5 key takeaways from SB1047 and how they will impact AI development."