feat: KGRag - Knowledge Graph-Enhanced RAG with Mellea

.gitignore

@@ -443,3 +443,11 @@ pyrightconfig.json
 .ionide
 
 # End of https://www.toptal.com/developers/gitignore/api/python,direnv,visualstudiocode,pycharm,macos,jetbrains
+
+# KG-RAG Example Data (Large Files - Keep Locally)
+docs/examples/kgrag/.env
+docs/examples/kgrag/dataset/*.jsonl.bz2
+docs/examples/kgrag/dataset/movie/*.json
+docs/examples/kgrag/dataset/movie/*.bz2
+docs/examples/kgrag/output/
+docs/examples/kgrag/data/

README.md

@@ -1,6 +1,5 @@
 <img src="https://github.com/generative-computing/mellea-contribs/raw/main/mellea-contribs.jpg" height=100>
 
-
 # Mellea Contribs
 
 The `mellea-contribs` repository is an incubation point for contributions to

docs/examples/kgrag/.env_template

@@ -0,0 +1,52 @@
+# Graph Database Configuration (Neo4j default; adjust for other graph DBs)
+NEO4J_URI=bolt://localhost:7687
+NEO4J_USER=neo4j
+NEO4J_PASSWORD=password
+
+# Data Directory
+KG_BASE_DIRECTORY=./dataset
+DATA_PATH=./data
+
+# ---------------------------------------------------------------------------
+# Primary LLM — any OpenAI-compatible endpoint
+# ---------------------------------------------------------------------------
+# Option A: OpenAI
+# API_KEY=sk-...
+# MODEL_NAME=gpt-4o-mini
+
+# Option B: Local Ollama
+# API_BASE=http://localhost:11434/v1
+# API_KEY=ollama
+# MODEL_NAME=llama3.2
+
+# Option C: vLLM / self-hosted
+# API_BASE=http://localhost:8000/v1
+# API_KEY=dummy
+# MODEL_NAME=meta-llama/Llama-3.1-8B-Instruct
+
+# Option D: Azure OpenAI
+# API_BASE=https://<your-resource>.openai.azure.com/openai/deployments/<deployment>/
+# API_KEY=<azure-key>
+# MODEL_NAME=gpt-4o-mini
+
+# ---------------------------------------------------------------------------
+# Optional: Separate evaluation model (defaults to primary LLM if unset)
+# ---------------------------------------------------------------------------
+# EVAL_API_BASE=...
+# EVAL_API_KEY=...
+# EVAL_MODEL_NAME=...
+
+# ---------------------------------------------------------------------------
+# Optional: Embedding model for vector entity alignment
+# ---------------------------------------------------------------------------
+# EMB_API_BASE=http://localhost:11434/v1
+# EMB_API_KEY=ollama
+# EMB_MODEL_NAME=nomic-embed-text
+# VECTOR_DIMENSIONS=768
+
+# Request Configuration
+MAX_RETRIES=3
+TIME_OUT=1800
+
+# OpenTelemetry — disable if you don't have a collector running
+OTEL_SDK_DISABLED=true