This tutorial contains code to set up a Knowledge Graph Index using NebulaGraph as the storage backend and OpenAI's Language Model (LLM) for query processing.
Setup
Environment Configuration
-
Load Environment Variables: Ensure you have a
.env
file set up with required environment variables. Usedotenv
to load them in your Python environment.
OPENAI_API_KEY=ak**
```python
import os
from dotenv import load_dotenv
load_dotenv()
```
-
Logging Configuration: Configure the logging settings.
import logging import sys logging.basicConfig( stream=sys.stdout, level=logging.INFO ) # logging.DEBUG for more verbose output
-
Import Dependencies: Import necessary libraries and modules.
# Import necessary modules and libraries from llama_index import ( KnowledgeGraphIndex, LLMPredictor, ServiceContext, SimpleDirectoryReader, ) from llama_index.storage.storage_context import StorageContext from llama_index.graph_stores import NebulaGraphStore from llama_index.llms import OpenAI from IPython.display import Markdown, display
NebulaGraph Setup
Ensure NebulaGraph (version 3.5.0 or newer) is set up with the following steps:
- Cluster Creation: Create a NebulaGraph cluster using one of the following options:
-
Docker Installation (Machines with Docker Installed):
- Option 0: Run the command:
curl -fsSL nebula-up.siwei.io/install.sh | bash
- Option 0: Run the command:
-
Desktop Installation:
- Option 1: Install NebulaGraph Docker Extension from Docker Hub.
Manual Setup via NebulaGraph Console: If the above options are not applicable, manually create the NebulaGraph space, tags, edges, and indexes using the NebulaGraph console. Refer to the provided commands in the code comments.
-
Environment Configuration: Set NebulaGraph environment variables.
os.environ["NEBULA_USER"] = "root" os.environ["NEBULA_PASSWORD"] = "nebula" # default is "nebula" os.environ["NEBULA_ADDRESS"] = "127.0.0.1:9669" # assumed NebulaGraph installed locally
Knowledge Graph Indexing
-
Define Graph Structure: Define the graph structure (space, tags, edges).
space_name = "llamaindex" edge_types, rel_prop_names = ["relationship"], ["relationship"] tags = ["entity"]
-
Initialize NebulaGraph Store and Storage Context:
graph_store = NebulaGraphStore( space_name=space_name, edge_types=edge_types, rel_prop_names=rel_prop_names, tags=tags, ) storage_context = StorageContext.from_defaults(graph_store=graph_store)
-
Load Data: Load documents/data to be indexed.
documents = SimpleDirectoryReader("data").load_data()
-
Create Knowledge Graph Index:
kg_index = KnowledgeGraphIndex.from_documents( documents, storage_context=storage_context, max_triplets_per_chunk=10, service_context=service_context, space_name=space_name, edge_types=edge_types, rel_prop_names=rel_prop_names, tags=tags, include_embeddings=True, )
Querying the Knowledge Graph
-
Initialize Query Engine:
from llama_index.query_engine import KnowledgeGraphQueryEngine query_engine = KnowledgeGraphQueryEngine( storage_context=storage_context, service_context=service_context, llm=llm, verbose=True, )
-
Querying:
response = query_engine.query( "your question on the data?", ) display(Markdown(f"<b>{response}</b>"))
Further Customization
- Adjust parameters and configurations in the code based on specific requirements or environment setup.
Top comments (0)