https://github.com/deepkit/deepkit
Product page: https://deepkit.ai/
It’s a machine learning platform for experiment tracking, execution, and debugging. The GUI is a real-time interface with native-like desktop interface which means all data is streamed using reactive RPC APIs via websockets. There’s no REST or polling involved.
Additional new Typescript open-source byproducts that were created specifically for this product and might be of interest to you:
-
Marshal.ts – The fastest universal Typescript data serializer and validator incl. Mongo-ORM abstraction
-
Glut.ts – A reactive real-time client-server framework with distributed entity/file abstraction, distributed data exchange, and automatic entity syncing, based on RxJs and websockets
-
angular-desktop-ui – A GUI library for native-like GUI widget based on Angular 9+
-
npm-local-development A `npm link` alternative that actually works with complex setups.
-
typedoc-plugin-lerna-packages – TS auto doc generator for lerna packages
Some notes:
-
Package management with Lerna
-
oclif for cli (with custom build), Angular 9+ for GUI, Electron 8+, custom client-server framework using Glut.ts, Mongo DB
-
Development time roughly 1 year (80h/week, total EUR ~450k), a single developer (for the main product + those OSS libs)
-
multi component build based on classic Makefile and webpack
I post it here to give people the opportunity to look how such a complex product is organised, that it is possible to write something like that in TS as a solo developer, and to ask questions about the development process and the like.