Amazon Neputune が GraphRAG Toolkit をリリースしました

%reload_ext dotenv
%dotenv

import os

from graphrag_toolkit import LexicalGraphIndex, GraphRAGConfig
from graphrag_toolkit.storage import GraphStoreFactory
from graphrag_toolkit.storage import VectorStoreFactory

from llama_index.readers.web import SimpleWebPageReader

import nest_asyncio
nest_asyncio.apply()

+ GraphRAGConfig.extraction_llm = 'anthropic.claude-3-haiku-20240307-v1:0'
graph_store = GraphStoreFactory.for_graph_store(os.environ['GRAPH_STORE'])
vector_store = VectorStoreFactory.for_vector_store(os.environ['VECTOR_STORE'])

graph_index = LexicalGraphIndex(
    graph_store,
    vector_store
)

doc_urls = [
    'https://docs.aws.amazon.com/neptune/latest/userguide/intro.html',
    'https://docs.aws.amazon.com/neptune-analytics/latest/userguide/what-is-neptune-analytics.html',
    'https://docs.aws.amazon.com/neptune-analytics/latest/userguide/neptune-analytics-features.html',
    'https://docs.aws.amazon.com/neptune-analytics/latest/userguide/neptune-analytics-vs-neptune-database.html'
]

docs = SimpleWebPageReader(
    html_to_text=True,
    metadata_fn=lambda url:{'url': url}
).load_data(doc_urls)

graph_index.extract_and_build(docs, show_progress=True)

print('Complete')

### 質問文

What are the differences between Neptune Database and Neptune Analytics?

### 回答

Neptune Database and Neptune Analytics are complementary services with different purposes:

Neptune Database is a fully managed graph database service designed for high performance and availability. It is optimized for graph database workloads that require scaling to handle up to 100,000 queries per second. Neptune Database provides features like multi-AZ high availability, multi-region deployments, and is suitable for applications like fraud detection, social networking, and customer 360 views. [https://docs.aws.amazon.com/neptune-analytics/latest/userguide/neptune-analytics-vs-neptune-database.html, https://docs.aws.amazon.com/neptune/latest/userguide/intro.html]

On the other hand, Neptune Analytics is a memory-optimized graph analytics engine focused on quickly analyzing large graph datasets. It can load data from Neptune Database snapshots, Amazon S3, or other sources into memory for low-latency analytical queries and processing using graph algorithms. Neptune Analytics automatically provisions compute resources based on the graph size and is ideal for data science, investigative, and exploratory workloads that require fast iteration and insights on graph data. [https://docs.aws.amazon.com/neptune-analytics/latest/userguide/what-is-neptune-analytics.html, https://docs.aws.amazon.com/neptune-analytics/latest/userguide/neptune-analytics-vs-neptune-database.html]

In summary, while Neptune Database is an operational graph database, Neptune Analytics is an analytics engine optimized for deriving insights and trends from large graph datasets using in-memory processing and graph algorithms.

### 質問

What are the differences between Neptune Database and Neptune Analytics?

### 回答

Neptune Database and Neptune Analytics are two related but distinct services offered by AWS:

Neptune Database is a fully managed graph database service designed for high performance and scalability. It is optimized for handling highly connected datasets and transactional, operational workloads that require low latency queries and updates. [source_3]

On the other hand, Neptune Analytics is an analytics database engine focused on quickly analyzing large graph datasets stored in data lakes or existing graph databases like Neptune Database. It loads the graph data entirely into memory for low-latency analytic queries and processing using popular graph algorithms. Neptune Analytics is ideal for data science, investigative, and exploratory workloads that require fast iteration and insights over massive graph datasets. [source_1, source_2]

Some key differences are:

1. Neptune Database is a transactional database optimized for operational workloads, while Neptune Analytics is an analytics engine for processing large graphs in-memory.

2. Neptune Database provides multi-AZ high availability and multi-region deployments, whereas Neptune Analytics does not have those capabilities. [source_1]

3. Neptune Analytics complements Neptune Database by allowing you to run complex analytics on graph data without impacting the transactional database. [source_2]

4. Neptune Analytics can load data from Neptune Database snapshots or S3, analyze it quickly using algorithms, and discard the in-memory data when done. [source_2, source_4]

5. Neptune Analytics automatically provisions compute resources based on the graph size, while Neptune Database is provisioned separately. [source_4]

In summary, use Neptune Database for operational graph workloads requiring high availability and scalability, and Neptune Analytics for exploratory analytics over massive graphs using in-memory processing.