Skip to content

Latest commit

 

History

History
37 lines (22 loc) · 2.25 KB

README.md

File metadata and controls

37 lines (22 loc) · 2.25 KB
Raw

rocker-win demo: Plumber

About

Plumber is an "R package that converts your existing R code to a web API using a handful of special one-line comments."

Build and run

Run the following commands in the directory of this file to build and run a small example app in a Windows Container.

docker build --tag plumber-win .

docker run --rm -it -p 8000:8000 plumber-win

Use

Now visit the following URLs in your browser to see the mean of 10 random numbers: http://localhost:8000/mean

Or run the following command in PowerShell to send two numbers to the service with HTTP POST and retrieve the sum:

Invoke-WebRequest -UseBasicParsing -Uri http://localhost:8000/sum -Method POST -Body @{a='5';b='2'}

Look for the [7] in the RawContent section.

In the Dockerfile the default command starts an R script, which is added to the image wit COPY. The script creates a Plumber server based on an API definition script included in the installed R package.

Alternatively you might use a combination of ENTRYPOINT and CMD to run different Plumber API implementations.

What next?

  • You can run more plumber examples by selecting a differnt file in start.R, e.g. C:/R/library/plumber/examples/11-car-inventory/plumber.R to run this API. You must also adjust the working directory to .../11-car-inventory, because R on Windows resolves paths differently. Then re-build the container and open http://localhost:8000/car/
  • Write your own API and start script, bind mount them into the container, and run it from there, e.g. in PowerShell docker run --rm -it -p 8000:8000 -v ${PWD}:C:\demo plumber-win C:\demo\api_start.R (in cmd the current directory is %cd%). Open http://localhost:8000/echo?msg=Hello! or http://localhost:8000/plot or http://localhost:8000/platform.
  • You can run docker run --rm -it -p 8000:8000 --entrypoint R.exe plumber-win to get an R session inside the container and manually start your API.
  • Start multiple API endpoints under different ports by adjusting the -p parameter.