Add GraphRAG implementation (Issue #103)

Author: ausarkhan

GraphRAG/README.md

@@ -0,0 +1,26 @@
+# GraphRAG
+
+GraphRAG implementation using Llama-Index and NetworkX.
+
+## Features
+- Builds a knowledge graph from documents
+- Indexes data points in the graph
+- Retrieves and generates contextually relevant responses using an LLM
+- Visualizes the knowledge graph
+
+## Usage
+1. Install dependencies:
+   ```bash
+   pip install llama-index networkx
+   ```
+2. Run `graphrag.py`.
+3. Example query and knowledge graph visualization will be generated.
+
+## Requirements
+- Python 3.x
+- llama-index
+- networkx
+- OpenAI API key (for LLM and embeddings)
+
+## License
+MIT

GraphRAG/graphrag.py

@@ -0,0 +1,42 @@
+"""
+GraphRAG Implementation using Llama-Index and NetworkX
+Issue #103 for king04aman/All-In-One-Python-Projects
+"""
+import networkx as nx
+from llama_index.core import KnowledgeGraphIndex, SimpleNodeParser, QueryEngine
+from llama_index.llms.openai import OpenAI
+from llama_index.embeddings.openai import OpenAIEmbedding
+
+# Example documents
+DOCUMENTS = [
+    "Alice is a data scientist. She works at Acme Corp.",
+    "Bob is a software engineer. He collaborates with Alice on ML projects.",
+    "Acme Corp is a tech company based in New York."
+]
+
+# Step 1: Parse documents into nodes
+parser = SimpleNodeParser()
+nodes = parser.get_nodes_from_documents(DOCUMENTS)
+
+# Step 2: Build Knowledge Graph Index
+kg_index = KnowledgeGraphIndex(nodes)
+
+# Step 3: Create Query Engine (using OpenAI LLM and embeddings)
+llm = OpenAI(model="gpt-3.5-turbo")
+embed_model = OpenAIEmbedding(model="text-embedding-ada-002")
+query_engine = QueryEngine(kg_index, llm=llm, embed_model=embed_model)
+
+# Step 4: Example query
+query = "Who works at Acme Corp?"
+response = query_engine.query(query)
+print("Query:", query)
+print("Response:", response)
+
+# Step 5: Visualize Knowledge Graph
+G = nx.Graph()
+for node in nodes:
+    for rel in node.relationships:
+        G.add_edge(node.entity, rel.entity, label=rel.type)
+
+nx.write_gml(G, "knowledge_graph.gml")
+print("Knowledge graph saved as knowledge_graph.gml")

GraphRAG/requirements.txt

@@ -0,0 +1,2 @@
+llama-index
+networkx