README.md
In fcyprowski/microservicer: R microservice template creator

microservicer

R is really hard to deploy in comparison to Python, so - for now - most of data scientist would prefer to work with R only in the "experimentation mode". And well, let's be honest - it has to do with the language design, lack of standarization and years of littering the R world with the "scientist's code" and "analytics experience". We want to change that. I want to change that. R with the whole tidyverse is one of the best languages to work with complicated data structures. We just need the tools to make it more accessible for machines.

Package should provide not only the Dockerfiles and kube manifests builders, but also a full standarized working method with that kind of environment. So you should expect here some functions that will make a use of specified working directory trees and enforce good practices. And I really mean "ENFORCE" because the R is dying of the lack of standarization. So - first the structure. Then we'll march to Kubeflow.

Yeah, there are some of them. In fact, through my whole carrier I've met only two packages that can in some way help you to build a microservice. Packrat - the most annoying package ever with a big middle finger aimed at a declarative management and loooong build times and unintuitive resource management. And containerit which looks good, but it has some flaws (like not givin' a damn about versions) and is focused solely on