[Question]: NebulaGraph RAG

[Chengyang852]

Question Validation

• [X] I have searched both the documentation and discord for an answer.

Question

I want to use NebulaGraph to do RAG. Now that all my data has been stored in NG, what should I do next? Should I query all the data and then do RAG? Can you give a demo?

[dosubot[bot]]

To use NebulaGraph for Retrieval-Augmented Generation (RAG) with your existing data, follow these steps:

1. Install Required Packages:

   %pip install llama-index-llms-openai
   %pip install llama-index-embeddings-openai
   %pip install llama-index-graph-stores-nebula
   %pip install llama-index-llms-azure-openai
   %pip install nebula3-python
   %pip install ipython-ngql networkx pyvis

2. Set Up OpenAI or Azure OpenAI: For OpenAI:

   import os
   from llama_index.llms.openai import OpenAI
   from llama_index.core import Settings
   
   os.environ["OPENAI_API_KEY"] = "INSERT OPENAI KEY"
   
   llm = OpenAI(temperature=0, model="gpt-3.5-turbo")
   Settings.llm = llm
   Settings.chunk_size = 512

   For Azure OpenAI:

   import os
   import openai
   from llama_index.llms.azure_openai import AzureOpenAI
   from llama_index.embeddings.openai import OpenAIEmbedding
   from llama_index.core import Settings
   
   openai.api_type = "azure"
   openai.api_base = "https://<foo-bar>.openai.azure.com"
   openai.api_version = "2022-12-01"
   os.environ["OPENAI_API_KEY"] = "<your-openai-key>"
   openai.api_key = os.getenv("OPENAI_API_KEY")
   
   llm = AzureOpenAI(
      model="<foo-bar-model>",
      engine="<foo-bar-deployment>",
      temperature=0,
      api_key=openai.api.key,
      api_type=openai.api_type,
      api_base=openai.api_base,
      api_version=openai.api_version,
   )
   
   embedding_model = OpenAIEmbedding(
      model="text-embedding-ada-002",
      deployment_name="<foo-bar-deployment>",
      api_key=openai.api.key,
      api_base=openai.api.base,
      api_type=openai.api.type,
      api_version=openai.api.version,
   )
   
   Settings.llm = llm
   Settings.chunk_size = 512
   Settings.embed_model = embedding_model

3. Prepare NebulaGraph: Ensure you have a running NebulaGraph with the defined data schema:

   os.environ["NEBULA_USER"] = "root"
   os.environ["NEBULA_PASSWORD"] = "<password>"
   os.environ["NEBULA_ADDRESS"] = "127.0.0.1:9669"
   
   space_name = "paul_graham_essay"
   edge_types, rel_prop_names = ["relationship"], ["relationship"]
   tags = ["entity"]

4. Instantiate NebulaGraphStore:

   from llama_index.core import StorageContext
   from llama_index.graph_stores.nebula import NebulaGraphStore
   
   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)

5. Build the Knowledge Graph with LlamaIndex:

   from llama_index.core import KnowledgeGraphIndex, SimpleDirectoryReader
   
   documents = SimpleDirectoryReader(
      "../../../../examples/paul_graham_essay/data"
   ).load_data()
   
   index = KnowledgeGraphIndex.from_documents(
      documents,
      storage_context=storage_context,
      max_triplets_per_chunk=2,
      space_name=space_name,
      edge_types=edge_types,
      rel_prop_names=rel_prop_names,
      tags=tags,
   )

6. Perform Graph RAG Query:

   query_engine = index.as_query_engine()
   
   response = query_engine.query("Tell me more about Interleaf")
   from IPython.display import display, Markdown
   display(Markdown(f"<b>{response}</b>"))

7. Visualize the Graph RAG:

   %ngql --address 127.0.0.1 --port 9669 --user root --password <password>
   %%ngql
   USE paul_graham_essay;
   MATCH p=(n)-[*1..2]-()
    WHERE id(n) IN ['Interleaf', 'history', 'Software', 'Company'] 
   RETURN p LIMIT 100;
   %ng_draw

By following these steps, you can leverage NebulaGraph to perform Retrieval-Augmented Generation (RAG) with your existing data. This process involves loading your data, indexing it, storing the index, querying it, and evaluating the results to ensure accuracy and relevance. This allows you to enhance the capabilities of your LLM applications by incorporating your specific data into the responses generated by the LLM ^[1][2][3][4].

To continue talking to Dosu, mention @dosu.

[Chengyang852]

My data has been saved in NebulaGraph, not local data.

[logan-markewich]

Unless you constructed it with llama-index, the best you can do is text2cypher

[Chengyang852]

Unless you constructed it with llama-index, the best you can do is text2cypher

I would like to ask a question about llama-index. At present, I have only found the KnowledgeGraphIndex.from_documents function in llama-index to implement it, but this function seems to extract local data entity relationship and store it in NebulaGraph. Now that I have put the extracted entities and relationships into NebulaGraph, what should I do next? Extract the data from NG and then input it into this function?