May 2024 Summaries

This code creates a server with a single REST endpoint that generates GenAI answers backed by data stored in a Neo4j Graph Database. The Neo4j LangChain Starter Kit for Python developers showcases how to combine Neo4j, FastAPI, OpenAI, and LangChain technologies. It provides a generic framework for generating GenAI answers using a Neo4j database as the knowledge graph. The kit is designed to be simple and portable, allowing users to easily swap out different databases and models. To run the starter kit, users need to have credentials to an actively running and populated Neo4j database and an active OpenAI API key. The kit includes examples of how to use vector similarity searches with Neo4jVector and a prompt plus the answers from both chains to generate a composite answer.

The GQL ISO standard has been officially published as the new international standard for graph query languages, marking a significant milestone in the evolution of graph database languages. The development of GQL is a result of efforts by Neo4j and other vendors to create a standardized language that can unify the various query languages used in the industry. The openCypher project, which was initially launched as an open initiative to make Cypher available to the ecosystem, has been rebranded as a means to help Cypher database and tooling vendors transition to GQL. The new standard promises to bring about convergence in the graph query language scene, making it easier for users and developers to find common ground and work together.

Got your eye on NODES 2024? The deadline to submit your talk is fast approaching, and developers and data scientists from all over the world are getting ready to share their latest graph-powered projects at this year’s free online conference. To craft a compelling presentation, consider choosing from three session formats: a 30-minute talk with Q&A, a 10-minute lightning talk, or a two-hour hands-on workshop, and selecting one of four tracks. Think about your audience, outline your main points, start strong with an attention-grabbing opening line, sell your session by explaining what attendees will gain from it, add a personal touch, nail your title, get feedback from peers, and consider starting out with a lightning talk if you're new to speaking. The deadline for submitting your graph story is June 15th.

Neo4j has achieved full compatibility with AWS PrivateLink, earning an AWS PrivateLink Service Ready status. This means that customers can securely access Neo4j AuraDS Enterprise on AWS using the integration, which is only available on the Enterprise tier of Neo4j Aura. The setup involves six steps: setting up a Neo4j Aura instance, configuring network access and AWS endpoint, testing the network configuration, connecting to Neo4j Aura using the Python API, connecting to Neo4j Browser, and finally connecting to Neo4j Bloom. Once set up, the deployment is entirely private, with neither the database, browser, nor Bloom being publicly available. The setup process was verified to ensure that it works properly, and customers can reach out to Neo4j's support team for any questions or concerns.

The Neo4j community is actively addressing climate change through technology. The GraphStuff.FM podcast features guests discussing the topic, while a GraphRAG article explores the benefits of GraphQL Aura API and chatbots in enhancing conference experiences. The NODES 2024 Call for Papers is now open, inviting graph story submissions. For beginners, a Cypher Aggregations course is available, offering more advanced content after following previous segments. Upcoming events include Neo4j Live, AWS Summit, Google Cloud Summit, jPrime 2024, and GraphSummit Series. The community is also exploring getting started with graphs through various resources, including the GraphAcademy Cypher Aggregations course, a livestream on Neo4j Live, and a free trial of Neo4j AuraDB. Maria Di Maro, a Postdoctoral Researcher at the University of Naples Federico II, showcased the potential of graphs in extracting information more intricately than traditional methods. The ExpectAI podcast discusses reducing carbon emissions through technology, while Fanghua Yu demonstrates integrating LlamaParse with Neo4j to create knowledge graphs. Jonathan Giffard provides an overview of the forthcoming Aura GraphQL API, and Peter Hanssens shares the development process of a chatbot for DataEngBytes attendees.

Neo4j has joined the Connect with Confluent technology partner program, expanding its data streaming capabilities through a fully managed integration with Confluent Cloud. This partnership enables organizations to stream any data directly into Neo4j, unlocking real-time insights and analytics for applications such as real-time recommendations, fraud detection, and network analysis. By integrating Neo4j's graph algorithms with Confluent's scalable infrastructure, companies can achieve unparalleled performance and scalability, driving innovation and competitive advantage in the data-driven landscape. The partnership also provides access to Confluent engineering, sales, and marketing resources, ensuring customer success at every stage of the onboarding process.

The text discusses a project that aims to bring the epic Mahabharata to life using a Neo4j-powered chatbot with Google Gemini. The chatbot is designed to translate natural language queries into efficient Cypher statements, allowing non-technical users to explore the intricate web of relationships within the epic. The project leverages Gradio, a user-friendly Python web framework, and Gemini API to create an intuitive platform for user interaction. While Gemini is a powerful tool, its accuracy depends on various factors such as query complexity, training data, and refinement needs. The chatbot's potential extends beyond just relationship analysis, with features like sentiment analysis, creative text formats, and knowledge graphs that could provide a more immersive experience. Ultimately, the project represents a fascinating journey towards interactive storytelling, where users can engage with narratives like the Mahabharata through an interactive conversation.

Neo4j has released Customer Managed Keys (CMKs) for its fully managed graph database as a service, Neo4j Aura. This feature allows organizations to encrypt their Neo4j-managed graph databases with their own cloud-based keys, providing increased autonomy and security. CMKs enable users to define access permissions and usage policies according to their specific security requirements, ensuring robust encryption and access controls. By using CMKs, organizations can comply with strict data protection and privacy regulations such as GDPR, HIPAA, and PCI DSS. The feature is currently available for Amazon Web Services keys, with plans to support Microsoft Azure and Google Cloud Platform keys in the future. To get started, users can log in to the Aura Console or visit Neo4j Support.

The NODES 2024 conference invites developers and data scientists to share their knowledge on graph-driven innovation through presentations, workshops, and panel discussions. The conference features various talk tracks covering applications, AI, data science, graphs, and more. Notable talks from the previous year's conference provide inspiration for attendees. Proposals must be submitted by June 15th, and the deadline is approaching soon.

The Going Meta series explored various aspects of semantics, ontologies, and knowledge graphs. The first season covered foundational principles of graph databases, data engineering, knowledge management, developer/data integration, and advanced semantics, with a focus on GraphRAG and knowledge graphs in Action. Episodes 21-27 focused on GenAI, RAG, and knowledge graphs, exploring how these concepts can be used to improve AI results, create reflection agents, and build domain-specific information systems. The series concluded with a recap of the first season's episodes and announced the start of Season 2, which will delve deeper into the topics covered in previous seasons.

The author of the blog post is exploring the Mahabharata, an ancient Indian epic, using Neo4j graph database technology. The post analyzes the importance of relationships in the epic and how they form the backbone of the narrative, influencing characters' decisions and shaping the plot. The author believes that graph databases are well-suited to model and analyze these complex relationships, as they treat relationships as first-class entities. By constructing a comprehensive graph of characters and their relationships using Neo4j, the author aims to uncover hidden connections and gain a deeper understanding of the epic. The post also touches on the potential for a Neo4j-powered chatbot to make the epic's complex relationships accessible to everyone, transcending traditional query methods.

The Institute for Democracy, Journalism, and Citizenship at Syracuse University is using Neo4j graph database technology to analyze complex data from political ads, funders, and candidates, uncovering hidden patterns and relationships that reveal insights into the flow of money and influence in the digital political landscape. By modeling the network of political ads, funders, and candidates as nodes connected by relationships, researchers can follow these connections to identify nuanced patterns and behaviors, such as funder content strategies and similarities between funders' approaches. This technology has far-reaching applications beyond politics, including investigating complex financial crimes like money laundering and fraud, and uncovering new insights in healthcare and scientific research. The IDJC's pioneering research powered by Neo4j graph technology has the potential to profoundly impact the integrity and transparency of democratic processes, empowering voters to critically evaluate political messages and resist attempts to influence their opinions, while providing journalists and watchdog groups with powerful new means to investigate the sources and spread of misinformation and disinformation in political campaigns.

The world of graph databases is abuzz with exciting developments, including the recent ISO GQL publication and a discussion panel on May 13. The Neo4j community has also been busy creating practical guides for GraphRAG, Testcontainers, and a metadata management tool. Meanwhile, the NODES 2024 Call for Papers is now open, inviting graph enthusiasts to share their stories. For those new to graph databases, an Introduction to Neo4j Graph Data Science course was highlighted as a great resource. The coming week promises more exciting events, including livestreams, conferences, meetups, and webinars featuring the Neo4j team and community members like Will Lyon, who is helping developers solve spatial data problems with graphs.

Neo4j Graph Data Science (GDS) was used to extract topics from documents in a vector store, allowing for semantic search capabilities. The GDS toolset enabled the creation of a knowledge graph representing documents and related topics. The graph's vector search capability facilitated searches over vector representations of topics and documents. By merging duplicated or closely related themes, the algorithm improved the efficiency of semantic searches. The use of stem words to identify common root words helped in identifying synonyms, while other techniques like Leiden community detection were used to group similar themes together. The long summary theme group index outperformed other indexing strategies, finding 27% more relevant movies than the movie index. The technique provided a structured approach to topic modeling and knowledge graph creation, allowing for better semantic search capabilities in RAG applications.

LlamaParse is a proprietary parsing tool for complex documents with embedded objects like tables and figures, which integrates seamlessly with LlamaIndex ingestion and retrieval. It enables the building of retrieval systems over complex, semi-structured documents, facilitating answers to previously unmanageable complex questions. To create knowledge graphs from documents using LlamaParse and Neo4j, one needs to set up an environment, process PDF documents, design a graph model, store extracted data in Neo4j, generate text embeddings, and query and analyze the data. The integration of LlamaParse with Neo4j allows for building GraphRAG applications that can uncover insights and relationships hidden within PDF content.

The Neo4j Query Languages Standards and Research Team has been actively involved in the development of ISO/IEC 39075 Information technology — Database languages — GQL, a new standard for a property graph database language. The team's involvement was unique as it combined ideas from industry-proven graph query languages with SQL. The GQL standard was developed by SC32 WG3, an international standards committee responsible for developing and enhancing the SQL database language standard. The development process took about five years and involved 38 meetings, including face-to-face and web conferences. The final standard consists of eight key ideas that fuse concepts from openCypher, GSQL, and PGQL with SQL. It includes features such as querying, updating, and managing graph databases using the property graph model, user-friendly language syntax, consistent visual 'ASCII-art' style graph patterns, and natural composition of complex queries. The standard also incorporates 430 papers developed, reviewed, discussed, and accepted into the GQL standard. The development process required collaboration among experts from various countries, including those employed by database vendors. The Neo4j team's experience in working on the GQL standard was a steep learning curve, but it provided valuable insights and skills that can be applied to future projects.

Welcome to This Week in Neo4j, your weekly fix for news from the world of graph databases! The Knowledge Graph for RAG Course on Deeplearning is a perfect timing for next week's Knowledge Graph Conference. Additionally, this edition has a recap from an Open Source AI Meetup and instructions on adding RAG to your GraphQL API. The NODES 2024 Call for Papers is now open, inviting submissions of graph stories. A few links have been added for Graph Database Beginners, including a recent GraphAcademy Live where Neo4j & LLM Fundamentals were covered. The Knowledge Graph Conference and AWS Summit are coming up next week in New York on May 6 and Singapore on May 7 respectively. Meetups will take place in New York on May 7, Sydney on May 8, and Bengaluru on May 11. Neo4j Events include webinars and more. The GraphSummit Series is getting connected with graphs, with the next event taking place in Sydney, AU on May 9. GRAPHACADEMY offers Neo4j Fundamentals, a watchable version of the course, and a tryable version called Neo4j AuraDB Free. Johannes Jolkkonen, a data architect and consultant specializing in Azure and LLM applications, is featured as a community member. A recent livestream addressed entity resolution and deduplication with Neo4j and GenAI. Andreas Kollegger will teach a course on using knowledge graphs within RAG applications. Paco Nathan will deliver a masterclass at the Knowledge Graph Conference, constructing a knowledge graph in Neo4j and comparing the before/after impact of resolving duplicate records. Jesús Barrasa will deliver a keynote Exploring AI's Future: The Convergence of Knowledge Graphs and LLMs on Thursday. An open-source AI meetup took place during KubeCon Paris last month, gathering builders from the AI community. Adam Cowley looked into adding RAG to GraphQL APIs and found the generate resolver as a flexible approach. Finally, DirectoryRanger shared an Active Directory audit tool that leverages cypher queries to crunch data from BloodHound graph database.