ContextQMD
Back to Llama Index

Multi-Tenancy RAG with LlamaIndex

docs/examples/multi_tenancy/multi_tenancy_rag.ipynb

0.14.213.7 KB
Original Source

<a href="https://colab.research.google.com/github/run-llama/llama_index/blob/main/docs/examples/multi_tenancy/multi_tenancy_rag.ipynb" target="_parent"></a>

Multi-Tenancy RAG with LlamaIndex

In this notebook you will look into building Multi-Tenancy RAG System using LlamaIndex.

  1. Setup
  2. Download Data
  3. Load Data
  4. Create Index
  5. Create Ingestion Pipeline
  6. Update Metadata and Insert documents
  7. Define Query Engines for each user
  8. Querying

Setup

You should ensure you have llama-index and pypdf is installed.

python
!pip install llama-index pypdf

Set OpenAI Key

python
import os

os.environ["OPENAI_API_KEY"] = "YOUR OPENAI API KEY"
python
from llama_index.core import VectorStoreIndex
from llama_index.core.vector_stores import MetadataFilters, ExactMatchFilter
from llama_index.core import SimpleDirectoryReader
from llama_index.core.ingestion import IngestionPipeline
from llama_index.core.node_parser import SentenceSplitter

from IPython.display import HTML

Download Data

We will use An LLM Compiler for Parallel Function Calling and Dense X Retrieval: What Retrieval Granularity Should We Use? papers for the demonstartions.

python
!wget --user-agent "Mozilla" "https://arxiv.org/pdf/2312.04511.pdf" -O "llm_compiler.pdf"
!wget --user-agent "Mozilla" "https://arxiv.org/pdf/2312.06648.pdf" -O "dense_x_retrieval.pdf"

Load Data

python
reader = SimpleDirectoryReader(input_files=["dense_x_retrieval.pdf"])
documents_jerry = reader.load_data()

reader = SimpleDirectoryReader(input_files=["llm_compiler.pdf"])
documents_ravi = reader.load_data()

Create an Empty Index

python
index = VectorStoreIndex.from_documents(documents=[])

Create Ingestion Pipeline

python
pipeline = IngestionPipeline(
    transformations=[
        SentenceSplitter(chunk_size=512, chunk_overlap=20),
    ]
)

Update Metadata and Insert Documents

python
for document in documents_jerry:
    document.metadata["user"] = "Jerry"

nodes = pipeline.run(documents=documents_jerry)
# Insert nodes into the index
index.insert_nodes(nodes)
python
for document in documents_ravi:
    document.metadata["user"] = "Ravi"

nodes = pipeline.run(documents=documents_ravi)
# Insert nodes into the index
index.insert_nodes(nodes)

Define Query Engines

Define query engines for both the users with necessary filters.

python
# For Jerry
jerry_query_engine = index.as_query_engine(
    filters=MetadataFilters(
        filters=[
            ExactMatchFilter(
                key="user",
                value="Jerry",
            )
        ]
    ),
    similarity_top_k=3,
)

# For Ravi
ravi_query_engine = index.as_query_engine(
    filters=MetadataFilters(
        filters=[
            ExactMatchFilter(
                key="user",
                value="Ravi",
            )
        ]
    ),
    similarity_top_k=3,
)

Querying

python
# Jerry has Dense X Rerieval paper and should be able to answer following question.
response = jerry_query_engine.query(
    "what are propositions mentioned in the paper?"
)
# Print response
display(HTML(f'<p style="font-size:20px">{response.response}</p>'))
python
# Ravi has LLMCompiler paper
response = ravi_query_engine.query("what are steps involved in LLMCompiler?")

# Print response
display(HTML(f'<p style="font-size:20px">{response.response}</p>'))
python
# This should not be answered as Jerry does not have information about LLMCompiler
response = jerry_query_engine.query("what are steps involved in LLMCompiler?")

# Print response
display(HTML(f'<p style="font-size:20px">{response.response}</p>'))