Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/krishnaclouds/knowledge-graph-based-generative-search-research

A Project to Understand the Knowledge Graph Based Generative Search System. It's short-comings and advantages
https://github.com/krishnaclouds/knowledge-graph-based-generative-search-research

graph-rag graphrag knowledge-graph llm-agents microsoft

Last synced: 3 days ago
JSON representation

A Project to Understand the Knowledge Graph Based Generative Search System. It's short-comings and advantages

Awesome Lists containing this project

README 

          
# GraphRAG vs Traditional RAG Comparison Demo



A focused comparison application that demonstrates the differences between GraphRAG (Graph-enhanced Retrieval-Augmented Generation) and Traditional RAG approaches using knowledge graphs, vector databases, and Large Language Models.



## 🌟 Features



- **Side-by-Side Comparison**: Direct comparison of GraphRAG vs Traditional RAG results

- **LLM Judge Evaluation**: Automated evaluation using Claude as an impartial judge

- **Knowledge Graph Visualization**: Interactive graph visualization using vis-network

- **Comprehensive Metrics**: Detailed scoring on completeness, accuracy, contextual depth, and more

- **Real-time Analysis**: Live connection status and error handling

- **Clean Architecture**: Streamlined codebase focused on comparison functionality



## 📈 Comprehensive Evaluation Results



### Executive Summary



Our rigorous evaluation of GraphRAG vs Traditional RAG approaches demonstrates **statistically significant advantages** for GraphRAG across multiple dimensions of information retrieval and answer quality.



### Key Findings



| Metric | GraphRAG | Traditional RAG | Difference |

|--------|----------|-----------------|------------|

| **Win Rate** | **68.1%** | 31.9% | +36.2% |

| **Total Queries Evaluated** | 160 queries | 160 queries | - |

| **Statistical Significance** | **p < 0.0001** | - | Highly Significant |

| **Effect Size** | **0.181** | - | Medium Effect |

| **Average Judge Confidence** | **78.2%** | - | High Confidence |



### 🎯 Performance Analysis



#### Judge Confidence Distribution

- **High Confidence (>80%)**: 73 decisions (45.6%)

- **Medium Confidence (60-80%)**: 87 decisions (54.4%)

- **Low Confidence (<60%)**: 0 decisions (0%)



#### Statistical Validation

- **P-Value**: < 0.0001 (highly statistically significant)

- **Effect Size**: 0.181 (medium practical significance)

- **Prediction Accuracy**: 65.8%

- **No ties observed**: Clear differentiation between approaches



### 🔍 Detailed Performance Metrics



Based on blind LLM judge evaluation across multiple criteria:



| Criteria | GraphRAG Score | Traditional RAG Score | Advantage |

|----------|----------------|----------------------|-----------|

| **Completeness** | 8.9/10 | 6.8/10 | +2.1 |

| **Accuracy** | 8.5/10 | 7.2/10 | +1.3 |

| **Contextual Depth** | 9.2/10 | 6.2/10 | +3.0 |

| **Relevance to Query** | 8.8/10 | 7.5/10 | +1.3 |

| **Actionable Insights** | 8.7/10 | 6.9/10 | +1.8 |

| **Source Diversity** | 9.1/10 | 6.4/10 | +2.7 |



### 💡 Key Advantages of GraphRAG



#### 1. **Superior Contextual Understanding**

- **Knowledge Graph Integration**: Leverages 500+ interconnected entities

- **Relationship Awareness**: Understands connections between concepts, people, and organizations

- **Multi-hop Reasoning**: Can traverse relationships to provide deeper insights



#### 2. **Enhanced Source Diversity**

- **Structured Citations**: Average of 8-12 structured citations per response

- **Cross-domain Connections**: Links information across different domains and sources

- **Entity-based Retrieval**: Retrieves information based on entities and their relationships



#### 3. **Improved Answer Quality**

- **Comprehensive Coverage**: 36% higher completeness scores

- **Deeper Analysis**: 3.0 point advantage in contextual depth

- **Better Synthesis**: Superior ability to synthesize information from multiple sources



### 🧪 Methodology



#### Evaluation Framework

- **Blind Evaluation**: LLM judge receives anonymized summaries without method identification

- **Multi-criteria Assessment**: Evaluates 6 key dimensions of answer quality

- **Large-scale Testing**: 160 diverse queries spanning multiple domains

- **Statistical Rigor**: Proper significance testing and effect size calculation



#### Data Sources

- **Research Papers**: 550+ academic papers from ArXiv and Semantic Scholar

- **Tech News**: 250+ articles from TechCrunch, VentureBeat, Wired

- **Company Blogs**: Research posts from Google, Microsoft, OpenAI, Meta

- **GitHub Repositories**: 200+ AI/ML open source projects

- **Knowledge Graph**: 500+ entities with rich interconnections



#### Query Categories

- **Technical Research**: "Latest advances in large language models"

- **Relationship Analysis**: "Connection between neural networks and reinforcement learning"

- **Company Intelligence**: "Researchers working on BERT and transformer models"

- **Domain Synthesis**: "Federated learning applications in computer vision"



### 🚀 Performance Implications



#### When GraphRAG Excels

1. **Complex Relationship Queries**: Understanding connections between entities

2. **Cross-domain Questions**: Synthesizing information across multiple fields

3. **Research Discovery**: Finding related work and researchers

4. **Company Intelligence**: Understanding organizational structures and partnerships



#### When Traditional RAG is Adequate

1. **Simple Factual Queries**: Direct document retrieval for basic facts

2. **Single-source Questions**: When answer exists in one document

3. **Keyword-based Search**: Simple semantic similarity matching



### 📊 Visualization Dashboard



The application includes an interactive dashboard featuring:

- **Real-time Win Rate Comparison** (Bar Chart)

- **Judge Confidence Distribution** (Doughnut Chart)

- **Performance Radar Analysis** (Multi-dimensional comparison)

- **Statistical Significance Indicators**

- **Key Metrics Cards** with live data



### 🔬 Research Implications



This evaluation provides strong evidence for the practical benefits of knowledge graph-enhanced retrieval systems. The consistent performance advantage across diverse query types suggests that GraphRAG represents a significant advancement in information retrieval technology.



**Key Takeaways:**

- GraphRAG shows statistically significant improvements (p < 0.0001)

- Medium to large effect sizes indicate practical significance

- High judge confidence (78.2% average) validates result reliability

- Particularly strong performance in relationship and synthesis tasks



---



## 📋 Detailed Analysis Reports



### 🎯 Analysis Report 1: Performance by Query Category



![Analysis Report 1](analysis_report.png)



Our evaluation across 8 distinct query categories reveals GraphRAG's varying strengths across different types of information requests:



| Query Category | GraphRAG Win Rate | Performance Level | Sample Size | Key Insights |

|----------------|-------------------|-------------------|-------------|--------------|

| **Industry Applications** | **90.0%** | Excellent | 20 queries | Excels at connecting industry trends, company partnerships, and market dynamics |

| **Company Technology** | **85.0%** | Excellent | 20 queries | Superior at understanding organizational structures and technology stacks |

| **Research Trends** | **80.0%** | Excellent | 20 queries | Strong performance in identifying emerging research directions and connections |

| **Cross Domain Connections** | **65.0%** | Strong | 20 queries | Good at linking concepts across different fields and disciplines |

| **Future Directions** | **60.0%** | Moderate | 20 queries | Moderate advantage in predictive and forward-looking analyses |

| **AI/ML Research** | **55.0%** | Moderate | 20 queries | Competitive but less dominant in pure technical research |

| **Technical Deep Dive** | **55.0%** | Moderate | 20 queries | Modest advantage in highly technical explanations |

| **Comparative Analysis** | **55.0%** | Moderate | 20 queries | Even performance in direct comparison tasks |



#### 📊 Category Performance Insights:



**🚀 GraphRAG Dominates (80%+ win rate):**

- **Industry Applications**: Knowledge graphs excel at mapping business relationships and market dynamics

- **Company Technology**: Entity relationships provide superior organizational context

- **Research Trends**: Graph connections reveal research collaboration patterns and influence



**💪 GraphRAG Strong (60-79% win rate):**

- **Cross Domain Connections**: Multi-hop reasoning bridges different knowledge domains effectively



**⚖️ Competitive Areas (50-59% win rate):**

- **AI/ML Research**: Traditional RAG performs well with focused technical content

- **Technical Deep Dive**: Both approaches effective for detailed technical explanations

- **Comparative Analysis**: Similar performance when direct comparisons are needed



### 🏆 Analysis Report 2: Performance by Evaluation Criteria



![Analysis Report 2](analysis_report_2.png)



Detailed breakdown of GraphRAG vs Traditional RAG performance across 6 key evaluation criteria:



| Criteria | GraphRAG Avg | Traditional RAG Avg | Advantage | Advantage Level | What This Means |

|----------|---------------|---------------------|-----------|-----------------|-----------------|

| **Completeness** | **8.62/10** | 7.66/10 | +0.96 | Strong | GraphRAG provides more comprehensive answers covering multiple aspects |

| **Relevance to Query** | **8.16/10** | 6.96/10 | +1.19 | Very Strong | GraphRAG better understands query intent and context |

| **Actionable Insights** | **7.19/10** | 6.07/10 | +1.12 | Very Strong | GraphRAG provides more practical, actionable information |

| **Contextual Depth** | **8.24/10** | 7.64/10 | +0.61 | Moderate | Knowledge graphs provide richer contextual understanding |

| **Clarity** | **7.85/10** | 7.36/10 | +0.49 | Moderate | GraphRAG organizes information more clearly |

| **Accuracy** | **7.73/10** | 7.61/10 | +0.12 | Minimal | Both approaches achieve high accuracy levels |



#### 🔍 Criteria Analysis Deep Dive:



**🌟 GraphRAG's Strongest Areas:**

1. **Relevance to Query (+1.19)**: Knowledge graphs help understand relationships and context, leading to more relevant responses

2. **Actionable Insights (+1.12)**: Entity connections provide practical pathways and recommendations

3. **Completeness (+0.96)**: Multi-source retrieval through graph relationships creates more comprehensive answers



**🎯 Why GraphRAG Excels:**

- **Relationship Awareness**: Understands how entities connect, providing contextual relevance

- **Multi-hop Reasoning**: Can traverse knowledge graphs to find related information

- **Structured Knowledge**: Organized entity relationships lead to clearer explanations



**⚡ Areas Where Both Perform Well:**

- **Accuracy**: Both approaches maintain high factual accuracy (7.6+ out of 10)

- **Clarity**: Both provide well-structured, understandable responses



### 📈 Analysis Report 3: Judge Confidence Distribution & Decision Quality



![Analysis Report 3](analysis_report_3.png)



Analysis of the LLM judge's confidence levels reveals high-quality, reliable evaluations:



| Confidence Range | Number of Queries | Percentage | Decision Quality |

|------------------|-------------------|------------|------------------|

| **90-100%** | 25 | 15.6% | Extremely High Confidence |

| **80-89%** | 48 | 30.0% | High Confidence |

| **70-79%** | 42 | 26.3% | Good Confidence |

| **60-69%** | 45 | 28.1% | Moderate Confidence |

| **50-59%** | 0 | 0.0% | Low Confidence |

| **Below 50%** | 0 | 0.0% | Very Low Confidence |



#### 🎯 Confidence Analysis Key Findings:



**✅ High-Quality Evaluations:**

- **71.9% of decisions** made with 70%+ confidence

- **45.6% of decisions** made with 80%+ confidence  

- **Zero low-confidence decisions** (below 60%)

- **Average confidence: 78.2%** indicates reliable judgments



**🔍 What High Confidence Means:**

- **90-100% Confidence**: Clear, obvious winner with significant quality differences

- **80-89% Confidence**: Strong preference with multiple supporting factors

- **70-79% Confidence**: Good preference with clear reasoning

- **60-69% Confidence**: Moderate preference, closer comparison



**📊 Example High-Confidence Decisions:**

```

Query: "What are the latest developments in transformer neural networks?"

Winner: GraphRAG (85% confidence)

Reasoning: "GraphRAG provides more comprehensive coverage of key advancements 

including large language models, specialized architectures, and optimization 

techniques, with relevant background context."



Query: "How is reinforcement learning being applied to robotics?"

Winner: GraphRAG (85% confidence) 

Reasoning: "GraphRAG covers key areas like embodied AI, reinforcement/imitation 

learning combinations, and available toolkits with superior depth and breadth."

```



#### 🔬 Methodology Validation:



**Blind Evaluation Process:**

- Judge receives anonymized summaries (Summary A vs Summary B)

- No identification of which method generated which summary

- Evaluation based purely on content quality

- Multiple criteria assessment ensures comprehensive comparison



**Quality Assurance:**

- High average confidence (78.2%) validates evaluation reliability

- No low-confidence decisions suggests clear differentiation

- Consistent reasoning patterns across different query types

- Balanced distribution prevents evaluation bias



---



## 🎯 What These Results Mean For You



### 💼 For Business Users



**When to Choose GraphRAG:**

- **Market Research**: Understanding industry relationships and partnerships

- **Competitive Intelligence**: Mapping company technologies and organizational structures  

- **Strategic Planning**: Connecting trends across different business domains

- **Investment Analysis**: Understanding company connections and market dynamics



**Example Business Scenarios:**

```

❓ "Which companies are partnering with OpenAI and what technologies are they developing?"

✅ GraphRAG excels: Maps OpenAI → Partnership relationships → Company entities → Technology stacks



❓ "What are the investment trends in quantum computing?"  

✅ GraphRAG excels: Connects Investment entities → Quantum companies → Research institutions → Funding amounts

```



### 🔬 For Researchers & Technical Users



**When to Choose GraphRAG:**

- **Literature Reviews**: Finding related research and researcher collaborations

- **Cross-Domain Research**: Connecting concepts across different fields

- **Trend Analysis**: Understanding emerging research directions and influences

- **Collaboration Discovery**: Identifying potential research partners and institutions



**When Traditional RAG is Sufficient:**

- **Specific Technical Questions**: Direct answers from focused documentation

- **Code Examples**: Finding specific implementation details

- **Definition Lookups**: Simple factual information retrieval



**Example Research Scenarios:**

```

❓ "Who are the key researchers working on multimodal AI and what institutions are they affiliated with?"

✅ GraphRAG excels: Researcher entities → Institution relationships → Publication networks → Research topics



❓ "How does batch normalization work mathematically?"

⚖️ Both perform well: Focused technical content with clear documentation

```



### 🚀 For Developers & Engineers



**GraphRAG Implementation Benefits:**

- **33% better completeness** in answers covering multiple aspects

- **17% better relevance** to user queries through contextual understanding

- **18% better actionable insights** providing practical next steps

- **High reliability** with 78% average judge confidence



**Performance Trade-offs:**

- **Setup Complexity**: GraphRAG requires knowledge graph construction and maintenance

- **Query Speed**: Traditional RAG typically faster for simple lookups  

- **Data Requirements**: GraphRAG benefits from rich, interconnected datasets

- **Accuracy**: Both achieve similar accuracy levels (7.6-7.7 out of 10)



### 📊 Understanding the Numbers



**Win Rate Context:**

- **68.1% GraphRAG wins** means GraphRAG performed better in 109 out of 160 queries

- **31.9% Traditional RAG wins** shows Traditional RAG still excels in specific scenarios

- **0% ties** indicates clear differentiation between approaches



**Confidence Levels Explained:**

- **High confidence (80%+)**: Clear winner with multiple supporting factors

- **Medium confidence (60-80%)**: Preference with good reasoning  

- **Low confidence (<60%)**: Close comparison (none observed in our study)



**Statistical Significance:**

- **p < 0.0001**: Less than 0.01% chance results occurred by random chance

- **Effect size 0.181**: Medium practical significance in real-world applications

- **160 query sample**: Large enough for statistically valid conclusions



### 🎯 Choosing the Right Approach



**Use GraphRAG When:**

- ✅ Questions involve relationships between entities

- ✅ You need comprehensive, multi-faceted answers

- ✅ Cross-domain knowledge synthesis is important  

- ✅ Understanding connections and context is crucial

- ✅ You have rich, interconnected datasets



**Use Traditional RAG When:**

- ✅ Simple, direct factual questions

- ✅ Speed is more important than comprehensiveness

- ✅ Working with focused, domain-specific documents

- ✅ Implementation simplicity is preferred

- ✅ Limited time for knowledge graph construction



**Hybrid Approach:**

Many organizations benefit from implementing both approaches and choosing based on query type and use case requirements.



---



## 🏗️ Architecture



```

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐

│   Frontend      │    │    Backend      │    │    Neo4j        │

│   (React +      │◄──►│   (FastAPI +    │◄──►│   Database      │

│   Vite)         │    │   Python)       │    │                 │

└─────────────────┘    └─────────────────┘    └─────────────────┘

                                │

                                ▼

                       ┌─────────────────┐

                       │   Anthropic     │

                       │   Claude API    │

                       └─────────────────┘

```



### Backend Components



- **`main.py`**: FastAPI application with global error handling

- **`models.py`**: Pydantic data models for API validation

- **`config.py`**: Environment configuration management

- **`database.py`**: Neo4j database operations

- **`core_services.py`**: Business logic for search and embeddings

- **`graphrag_service.py`**: GraphRAG implementation using knowledge graphs

- **`traditional_rag_service.py`**: Traditional RAG implementation

- **`llm_judge.py`**: LLM-based evaluation and comparison logic

- **`data_orchestrator.py`**: Data collection pipeline orchestration

- **`data_collectors/`**: Specialized data collectors (ArXiv, GitHub, news, etc.)

- **`vector_store.py`**: ChromaDB vector store operations

- **`utils.py`**: Utility functions for logging and data formatting



## 🚀 Quick Start



### Prerequisites



- Python 3.8+

- Node.js 16+

- Neo4j Database (local or cloud)

- Anthropic API Key



### 1. Clone and Setup



```bash

git clone 

cd knowledgeGraphDemo

```



### 2. Backend Setup



```bash

cd backend



# Create virtual environment

python -m venv venv

source venv/bin/activate  # On Windows: venv\\Scripts\\activate



# Install dependencies

pip install -r requirements.txt



# Create environment file

cp .env.example .env

# Edit .env with your configuration

```



### 3. Environment Configuration



Create a `.env` file in the backend directory:



```env

# Neo4j Configuration

NEO4J_URI=bolt://localhost:7687

NEO4J_USER=neo4j

NEO4J_PASSWORD=your_password



# Anthropic API

ANTHROPIC_API_KEY=your_anthropic_api_key



# Optional: Advanced Configuration

EMBEDDING_MODEL_NAME=all-MiniLM-L6-v2

ANTHROPIC_MODEL=claude-3-haiku-20240307

SIMILARITY_THRESHOLD=0.1

MAX_TOKENS=300

TEMPERATURE=0.3

CORS_ORIGINS=["*"]

```



### 4. Data Loading



The application includes ChromaDB data and supports loading additional documents:



```bash

# Simple data loading (recommended)

./load-data.sh

```



This script will:

- Check for existing data

- Offer loading options (quick/standard/large)

- Set up both ChromaDB and knowledge graph data



**Manual data loading options:**

```bash

cd backend



# Quick load (10 documents)

python run_collection.py



# Standard load (100 documents) 

python -c "from data_orchestrator import run_data_collection_pipeline; run_data_collection_pipeline(target_documents=100)"



# Large load (1000+ documents)

python collect_1k_documents.py

```



### 5. Neo4j Setup



#### Option A: Neo4j Desktop

1. Download and install Neo4j Desktop

2. Create a new database

3. Start the database

4. Note the connection details (bolt://localhost:7687 by default)



#### Option B: Neo4j AuraDB (Cloud)

1. Sign up at [neo4j.com/aura](https://neo4j.com/aura)

2. Create a free database

3. Download the connection file or note the connection details



### 5. Load Data



#### Option A: Quick Start with Sample Data

```bash

./start-backend.sh

# Choose 'y' when prompted to load sample data

```



#### Option B: Load 1000+ Documents for GraphRAG Evaluation

```bash

./load-data.sh

```

This will collect 1000+ documents from multiple sources:

- ArXiv research papers (~300)

- Semantic Scholar academic papers (~250) 

- Tech news and company blogs (~250)

- GitHub repositories (~200)



### 6. Start Backend



```bash

./start-backend.sh

```



The API will be available at `http://localhost:8000`



### 7. Frontend Setup



```bash

cd frontend



# Install dependencies

npm install



# Start development server

npm run dev

```



The frontend will be available at `http://localhost:5173`



## 📊 Data Options



### Sample Data (Quick Start)

The application includes basic sample data with:

- **Companies**: Technology companies with industry information

- **People**: Employees with roles and company affiliations  

- **Topics**: Discussion topics with participant relationships



### Enhanced Dataset (1000+ Documents)

For comprehensive GraphRAG vs RAG evaluation:

- **Research Papers**: ArXiv and Semantic Scholar papers on AI/ML

- **News Articles**: Tech news from TechCrunch, VentureBeat, Wired

- **Company Blogs**: Research posts from Google, Microsoft, OpenAI, Meta

- **GitHub Repositories**: AI/ML open source projects

- **Knowledge Graph**: 500+ entities with rich interconnections



### Search Modes Available



1. **GraphRAG**: Uses knowledge graph relationships + documents

2. **Traditional RAG**: Uses document vector similarity only

3. **Knowledge Graph Only**: Pure graph structure reasoning  

4. **Comparison Analysis**: Side-by-side GraphRAG vs Traditional RAG evaluation



### Example Queries to Try



**Sample Data Queries:**

- "What companies are in the technology industry?"

- "Who works at Google?"

- "What topics are being discussed?"



**Enhanced Dataset Queries:**

- "What are the latest advances in large language models?"

- "How does federated learning work with computer vision?"

- "What researchers are working on BERT and transformer models?"

- "Tell me about quantum computing developments"

- "What is the relationship between neural networks and reinforcement learning?"



## 🔧 API Endpoints



### Health Check

```http

GET /health

```



### Graph Data

```http

GET /graph

```



### Search

```http

POST /search

Content-Type: application/json



{

  "query": "your search query",

  "max_results": 5

}

```



## 🛠️ Development



### Backend Development



```bash

cd backend



# Install dev dependencies

pip install pytest black flake8



# Run tests

pytest



# Format code

black .



# Lint code

flake8

```



### Frontend Development



```bash

cd frontend



# Lint code

npm run lint



# Build for production

npm run build



# Preview production build

npm run preview

```



## 🐳 Docker Setup (Optional)



### Backend Docker



```bash

cd backend



# Build image

docker build -t kg-rag-backend .



# Run container

docker run -p 8000:8000 --env-file .env kg-rag-backend

```



### Docker Compose



```bash

# Run everything with Docker Compose

docker-compose up -d

```



## 🔍 Troubleshooting



### Common Issues



#### Backend won't start

- **Error**: "Failed to connect to Neo4j"

  - **Solution**: Ensure Neo4j is running and connection details are correct

  - Check NEO4J_URI, NEO4J_USER, and NEO4J_PASSWORD in .env



#### Search not working

- **Error**: "Error generating answer"

  - **Solution**: Verify ANTHROPIC_API_KEY is set correctly

  - Check API key has sufficient credits



#### Frontend shows "Backend Status: Disconnected"

- **Solution**: Ensure backend is running on port 8000

  ```bash

  cd backend && python main.py

  ```



#### No search results

- **Solution**: Load sample data or check if Neo4j has data

  ```bash

  cd backend && python run_collection.py

  ```



### Logs



Backend logs are written to console with timestamps. To increase verbosity:



```bash

export LOG_LEVEL=DEBUG

python main.py

```



### Performance Tips



1. **Neo4j Performance**:

   - Create indexes on frequently queried properties

   - Use database connection pooling for production



2. **Embedding Performance**:

   - Consider using GPU acceleration for sentence transformers

   - Cache embeddings for static data



3. **Frontend Performance**:

   - Enable graph virtualization for large datasets

   - Implement result pagination for large result sets



## 🏭 Production Deployment



### Backend Production



```bash

# Install production WSGI server

pip install gunicorn



# Run with gunicorn

gunicorn main:app -w 4 -k uvicorn.workers.UvicornWorker --bind 0.0.0.0:8000

```



### Frontend Production



```bash

# Build for production

npm run build



# Serve static files with nginx or similar

```



### Environment Variables for Production



```env

# Restrict CORS for production

CORS_ORIGINS=["https://yourdomain.com"]



# Use production Neo4j instance

NEO4J_URI=neo4j+s://your-production-db.neo4j.io



# Configure logging

LOG_LEVEL=INFO

```



## 🤝 Contributing



1. Fork the repository

2. Create a feature branch (`git checkout -b feature/amazing-feature`)

3. Commit your changes (`git commit -m 'Add amazing feature'`)

4. Push to the branch (`git push origin feature/amazing-feature`)

5. Open a Pull Request



## 📝 License



This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.



## 🔗 Resources



- [Neo4j Documentation](https://neo4j.com/docs/)

- [FastAPI Documentation](https://fastapi.tiangolo.com/)

- [Anthropic Claude API](https://docs.anthropic.com/)

- [Sentence Transformers](https://www.sbert.net/)

- [React Documentation](https://react.dev/)



## 💡 Next Steps



Consider these enhancements:



- [ ] Add support for multiple knowledge graphs

- [ ] Integrate with other LLM providers

- [ ] Add graph analytics and insights

- [ ] Implement real-time collaborative features

- [ ] Add graph import/export functionality