December 2023 Summaries

Neo4j, a graph database provider, has released a newsletter summarizing its most popular blog posts from the previous year. The newsletter covers various topics including Graph Databases for Beginners, Knowledge Graphs and Large Language Models (LLMs), GenAI Stack Walkthrough, Vector Search, and more. Neo4j has also mentioned upcoming events such as NodeS 2023 live stream and GraphSummit Series in 2024. Additionally, the newsletter includes a tweet from @provenauthority that highlights the importance of interest graphs and interoperability between social spaces.

This week, Neo4j users and developers are gearing up for the holidays with a double dose of AdventOfCode puzzles, one solved using Neo4j and the other featuring 24 tips and tricks from Christophe Willemsen. The company is also launching its Neo4j Research panel, which allows users to share their experiences with researchers and influence future product development. Meanwhile, Chris Smith, Senior Software Engineer at Thomson Reuters Special Services, will be speaking at NODES 2023 about an end-to-end graph analytics pipeline identifying emerging communities of new technologies. Additionally, there are resources available on RAG, GenAI, Vector Indexes, and more, including a video on how to set up a Neo4j Knowledge Graph for RAG and tips from Christophe Willemsen on working with indexes, Cypher clauses, change data capture, composite databases, or vector indexes. Users can also follow along with the AdventOfNeo4j series, which shares one Neo4j 5 tip daily for 24 days in December 2023.

Neo4j has introduced a new vector index in version 5.11 to efficiently perform semantic search over unstructured text or other embedded data modalities, making it a great fit for most RAG applications that rely on both structured and unstructured data. The Neo4j Vector Index implementation in LangChain allows customization of the node label, text, and embedding property names. Users can customize the node label to store text chunks under the specified node label, where the info property is used to store text, and the vector property holds the text embedding representation. The index implementation also creates a unique node property constraint on the id property for faster imports and allows users to load additional documents into an instantiated vector index. Additionally, users can use custom retrieval queries to collect, transform, or calculate any additional graph information they want to return from the similarity search. The newly added vector index implementation in Neo4j makes it a perfect fit for highly complex and connected datasets.

This week in Neo4j, the GraphStuff.FM podcast is wrapping up 2023 with a recap episode featuring Large Language Models and Knowledge Graphs. The podcast also covers Cybersecurity for Kubernetes, a new Storage Engine for Neo4j that introduces Block format, and how to get more out of knowledge resources. Additionally, there's an opportunity to join the Neo4j Research panel, which allows researchers to share their experiences with product development teams and influence the future of Neo4j products. The week also highlights upcoming events like NODES 2023, where recordings are available for viewing, as well as a GraphSummit Series featuring Ming Yan Ti's session on Type-Safe Graphs With Neo4j & Prisma.

This article demonstrates how to integrate Neo4j Aura into GitHub Actions for automated testing. The author uses the Aura CLI, which is a command line interface for interacting with Neo4j Aura, to create and manage instances of the database. The author creates a new instance using the `aura instances create` command and verifies that it exists by running the `aura instances get` command. The author then runs tests in a Node.js environment using the credentials stored in the GitHub Action. Finally, the author deletes the instance using the `aura instances delete` command to shut down the database. The workflow is triggered every time code is pushed to the repository, and it consists of three jobs: one for creating the database, one for running the tests, and one for shutting down the database. The author provides instructions on how to set up the Aura CLI, create a new instance, run tests, and delete an instance using the `aura instances` commands.

The new Neo4j Driver for Javascript version 5.14 now supports querying Neo4j using Deno natively, thanks to a contribution from Braden MacDonald who created a tool to automatically generate Deno source code from the existing driver code. The generated code was adapted and submitted to the same acceptance test suite as the Node version of the driver, resulting in a new network layer that uses Deno native sockets for better performance and stability. To query Neo4j from Deno, users must first import the driver from the repository in their Typescript/Javascript file, and Deno will take care of downloading and installing the dependencies. The most straightforward way to run a query is using the Driver.executeQuery API, which provides a cluster-safe way to execute a single query transaction. Users can also use the Session.executeRead and Session.executeWrite APIs to manage transactions for different needs. A new scenario where data is sent from Neo4j to an external service to calculate credit scores has been demonstrated, showcasing how Deno's asynchronous iteration over records can be used to process data as it arrives. Additionally, support for explicit resource management was added to the driver in version 5.15.0, reducing boilerplate related to Driver and Session objects usage since the using keyword can replace close method calls. Users are encouraged to try out the new Deno driver and provide feedback on GitHub.

Neo4j is a graph database that has been featured in various news and updates, including working with RDF data and importing GTFS data into the platform. The company also explores the benefits of combining knowledge graphs with large language models (LLMs) for enhanced applications. Additionally, Neo4j has partnered with several organizations to deliver new courses on its integration with LLMs, such as "Neo4j & LLM Fundamentals" and "Build a Neo4j-backed Chatbot using Python". Furthermore, the company has introduced new tools like rdflib-neo4j for importing/exporting RDF data into Neo4j. The platform also features advanced graph databases like RAG with knowledge graphs, enabling users to represent data as interconnected entities and gain insights from the data.

The text discusses the integration of Large Language Models (LLMs) and Knowledge Graphs, two powerful technologies that are changing the AI landscape. LLMs, like ChatGPT, can understand and respond to human-like text, while knowledge graphs provide a structured representation of information, making it more intelligent and accurate. The combination of these two technologies is called retrieval augmented generation (RAG), which retrieves relevant information from knowledge graphs using vector and semantic search, then augments the response with contextual data. This process generates precise, accurate, and contextually relevant output. The text also highlights various use cases for this technology combination, including building apps, converting unstructured data to knowledge graphs, creating graph dashboards with natural language queries, generating cyberattack countermeasures, and fine-tuning an open-source LLM for text-to-Cypher translation. Additionally, it provides resources for getting started with Neo4j and LLMs, such as a free course on GraphAcademy and a guide on building a knowledge graph.

This week's focus is on RAG, Knowledge Graphs and LLMs, with articles exploring no-code dashboard building with Looker and learning how to build QGIS plugins with Python. The Neo4j Needle Graph App Dev Kit is also highlighted as a tool for accelerating graph app development. Additionally, upcoming events such as NODES 2023, API Days in Paris, and the Learning on Graphs Conference are announced, featuring exciting keynotes and tutorials. Alan Brown will be speaking at NODES about fighting financial crime using entity resolution and graph technology. The week also sees the release of a tweet from the Learning on Graphs Conference, highlighting its upcoming program with orals, tutorials, and poster sessions.