May 2023 Summaries

Last week's Flutter Hackathon showcased a variety of innovative projects. The Best Overall Project was Syno by @samuelP09301972, which provides concise video summaries. Study Genius by @salomon_diei came in as runner-up, offering a personalized learning experience for students. Lord Of The Memes by @mohonishc won Best FlutterFlow project, an interactive game using meme gifs. Colabity by @hichamics was the runner-up, matching creative minds for collaborative projects. ChatGlobe by @tmhk_tnht won Best Dart Edge Project, allowing users to chat globally in their native language. The Most Technically Impressive project was Accident Detection by @Guneetsingh02, using machine learning and sensor data for road accident detection. IntelliTask by @noga_dev was the runner-up, a semi-automated todo app with AI assistance. Caffeio by @caurregoz won Most Visually Pleasing, an app for coffee aficionados exploring brewing methods. Decision Tales by @JudoUergens was the runner-up, offering interactive storytelling based on user decisions. The winners received special Flutter Hackathon swag packs and a limited Flutter-themed Supabase keyboard. For those who missed the hackathon, resources are available to get started with Flutter and Supabase.

Vecs is a Python library for managing embeddings in Postgres databases using the pgvector extension, providing an interface that lets users work with their data like a dedicated vector store. It was designed with ease-of-use and exploratory data analysis in mind, but works equally well as a search workhorse. The library makes it easy to create a collection of vectors, insert records, query vectors by distance measures, filter based on user-defined metadata, and index collections for efficient querying.

Retrieval plugins allow ChatGPT to access information from a database, helping it answer questions from private datasets. Supabase recently contributed to the OpenAI repo with Postgres and Supabase implementations for building plugins using pgvector. A retrieval plugin turns documents into smaller chunks, converts these chunks into embeddings using OpenAI's text-embedding-ada-002 model, stores the embeddings in a vector database, and queries the vector database for relevant documents when a question is asked. Supabase added two vector provider options: one for Postgres and one for Supabase. The main differences are that the Postgres version uses the psycopg2 python library to directly connect to the database, while the Supabase version interacts with the database via PostgREST. Both have the same schema so you can easily switch between them.

ChatGPT Plugins support is being released in beta this week with a TypeScript plugin template running on Supabase Edge Runtime. The ai-plugin.json manifest file is required for ChatGPT to identify the plugin, understand its authentication mechanism, and locate the OpenAPI definition. Swagger-jsdoc can be used to generate the OpenAPI definition from JSDoc comments in the /functions/chatgpt-plugin/index.ts file. The CORS headers are added using a shared cors.ts file. A Dockerfile and docker-compose file are provided for running Edge Runtime locally, and users can test their plugins on localhost using the ChatGPT UI. Finally, deployment to Fly.io is outlined.

The Flutter hackathon, co-sponsored by FlutterFlow and Invertase, invites developers to build beautiful apps using the hybrid framework Flutter/Dart and Supabase within a 10-day timeframe. The competition has five categories with prizes for various aspects of the projects, including technical impressiveness, visual appeal, and creativity. Participants can submit their projects individually or in teams of up to five people, with a deadline of May 21st, 2023. A panel of judges, including representatives from FlutterFlow, Invertase, and Supabase, will evaluate submissions based on criteria such as functionality, design, technical skill, and use of features like Supabase. The event is open to developers who want to showcase their skills and creativity with Flutter and Supabase.

Supabase has released its Beta Update Launch Week 7, which includes several feature-packed updates such as open-sourcing Logflare, a self-hosted Deno Functions runtime, and improvements to Supabase Storage and Auth. The update also introduced new features in the Supabase Studio 2.0, including AI-powered tools, and announced the launch of dbdev, a PostgreSQL package manager. Furthermore, the community has been actively contributing to the platform through various projects, such as FlutterFlow supporting Supabase Authentication, and Guillaume creating an incredible course about Supabase. The update also highlighted several successful use cases, including Mendable.ai switching from Pinecone to Supabase for PostgreSQL vector embeddings.

In this tutorial, we built a Flutter app that uses Supabase for authentication and implements multi-factor authentication (MFA) using time-based one-time passwords (TOTP). The app includes the following features: 1. Registration page with email verification 2. Login page with MFA setup option 3. Home page with secure contents that require MFA to view 4. List of MFA factors and unenrollment page 5. Go router for handling routing and redirection based on user authentication status To achieve this, we used the following packages: - `flutter_dotenv`: To manage environment variables in Flutter apps - `go_router`: For managing app navigation and routing - `supabase_flutter`: For integrating Supabase with our Flutter app We also utilized some key features of Supabase, such as email verification, MFA setup, and row-level security (RLS) to secure the data.

PostgreSQL's pluggable storage allows developers to use different storage engines for different tables within the same database, enabling optimized performance for specific workloads. The technology has been introduced in PostgreSQL 12 with initial support for ZHeap, but current progress has stalled due to limitations in the table access method API. Despite this, several initiatives are underway, including community updates and upstreaming code from OrioleDB, which aims to bring pluggable storage to the Postgres core. With ambitious timelines, pluggable storage could land in PostgreSQL 17, making it an even more attractive option for developers.