No comments yet.
My name is Martin Braun. I am a Co-Founder and CTO of NeuroForge (https://neuroforge.de/) - a Germany based AI & Big Data startup.
As part of our services we also provide support for building integrations between (our) systems and external systems. As we didn't want to keep building the same REST APIs every time, we set out to build a standardized data integratin platform that allows for a quick way to generate user specified REST API definitions via a REST API. This has become NF Compose (https://github.com/neuroforgede/nfcompose).
I distinctively remember defining a REST API for a customer project 10 minutes before the first meeting and finalizing the final contract during the meeting together with the customer. The same job would have taken hours, if not 1-2 days, if we had built it the traditional way.
After using NF Compose for mulitple years internally and in customer projects, we recently decided to release the whole project under MPL-2.0 on GitHub and to iterate on it in public.
We have a companion repo over at https://github.com/neuroforgede/nfcompose-examples where we already have a small sample project that shows how to setup a REST API in 5-10 minutes (or if you know the concepts - seconds) in a reproducible manner. We will be updating that repository with more content over the next weeks.
Some notable features include:
- Ability to define DataSeries (REST APIs) via a browsable REST API
- Multi-Tenancy support
- Consumers (Webhooks) for changes in DataSeries with guaranteed delivery and best effort ordering of events
- Multiple PostgreSQL "backends" for DataSeries (different amount of historization)
- Ability to generate PostgreSQL views in the database so that a BI tool like Superset or Metabase can work with the data
- Ability to build custom REST APIs via pluggable Engines. Let NF Compose handle the authentication
I am looking forward to people trying the project out and giving feedback, feature requests, bug reports.
