December 2023 Summaries

SurrealDB's journey to develop a full-text search capability was driven by the need to create a user-friendly solution without compromising the integrity of its ACID SQL database. Initially considering Lucene-based solutions and the Rust-based Tantivy, SurrealDB faced challenges such as language inconsistency and index management in distributed environments. The absence of inherent ACID compliance in these technologies led SurrealDB to build its own search system to maintain atomicity, consistency, isolation, and durability, ensuring that transactions involving full-text searches are reliable and consistent. This development not only enhanced search functionality but also preserved the database's integrity, crucial for applications relying on data accuracy. Looking forward, SurrealDB plans to introduce vector search capabilities to further advance its database technology, demonstrating a commitment to innovation and user experience.

SurrealML is a tool designed to streamline the hosting, version control, and deployment of machine learning models by integrating with SurrealDB, allowing for model inference using SurrealQL statements. Unlike traditional machine learning libraries focused on model training, SurrealML emphasizes the packaging and deployment of trained models onto SurrealDB instances, enabling users to perform flexible model inferences with simple commands. This flexibility allows for comparisons between different model versions and the integration of inference results into database queries. The guide provides a practical walkthrough for setting up the environment, training models using PyTorch and scikit-learn, and integrating those models with SurrealDB for efficient computation. SurrealML facilitates local and remote model inference in ONNX format, ensuring language agnosticism and potential for future expansion to other language bindings. This approach marks a significant advancement in simplifying machine learning deployment processes, with SurrealQL providing adaptable interaction with multiple deployed models.

SurrealDB has revamped its documentation platform by transitioning to Docusaurus, aiming to enhance user experience through improved accessibility, content discovery, and search functionality. This transition supports Markdown and MDX for ease of use and allows customization, including a robust full-text search system powered by SurrealDB's own SurrealIX product. The new documentation features a user-friendly layout, versioning support, and a mobile-optimized interface, addressing past feedback regarding information findability and accessibility, including for users with disabilities. SurrealDB plans to expand language support and integrate interactive examples to foster hands-on learning, with future updates driven by user feedback and community contributions. This move is part of SurrealDB's commitment to providing comprehensive, efficient, and user-friendly documentation that meets the needs of its global user base.