Create a vetiver Dockerfile that mounts the pin inside

make_dockerfile.md

First in R:

library(tidymodels)
data(Sacramento, package = "modeldata")

rf_spec <- rand_forest(mode = "regression")
rf_form <- price ~ type + sqft + beds + baths

rf_fit <-
  workflow(rf_form, rf_spec) %>%
  fit(Sacramento)

library(vetiver)
#> 
#> Attaching package: 'vetiver'
#> The following object is masked from 'package:tune':
#> 
#>     load_pkgs
v <- vetiver_model(rf_fit, "sacramento-houses")

library(pins)
b <- board_folder("my-pins")
b %>% vetiver_pin_write(v)
#> Creating new version '20220921T205544Z-52d58'
#> Writing to pin 'sacramento-houses'
#> 
#> Create a Model Card for your published model
#> • Model Cards provide a framework for transparent, responsible reporting
#> • Use the vetiver `.Rmd` template as a place to start

vetiver_write_plumber(b, "sacramento-houses", rsconnect = FALSE)
vetiver_write_docker(v)
#> * Lockfile written to 'vetiver_renv.lock'.

^{Created on 2022-09-21 with reprex v2.0.2}

plumber.R looks like this:

# Generated by the vetiver package; edit with care

library(pins)
library(plumber)
library(rapidoc)
library(vetiver)
b <- board_folder(path = "my-pins")
v <- vetiver_pin_read(b, "sacramento-houses")

#* @plumber
function(pr) {
    pr %>% vetiver_api(v)
}

Open and edit the Dockerfile to mount the pin directory inside the container:

# Generated by the vetiver package; edit with care

FROM rocker/r-ver:4.2.1
ENV RENV_CONFIG_REPOS_OVERRIDE https://packagemanager.rstudio.com/cran/latest

RUN apt-get update -qq && apt-get install -y --no-install-recommends \
  libcurl4-openssl-dev \
  libicu-dev \
  libsodium-dev \
  libssl-dev \
  make \
  zlib1g-dev \
  && apt-get clean

COPY vetiver_renv.lock renv.lock
RUN Rscript -e "install.packages('renv')"
RUN Rscript -e "renv::restore()"
ADD my-pins /opt/ml/my-pins
COPY plumber.R /opt/ml/plumber.R
EXPOSE 8000
ENTRYPOINT ["R", "-e", "pr <- plumber::plumb('/opt/ml/plumber.R'); pr$run(host = '0.0.0.0', port = 8000)"]

What I changed was ADD my-pins /opt/ml/my-pins; this mounts the directory where your pins are stored inside the container.

Now you can build and then run the container:

docker build --platform linux/amd64 -t pin-inside-demo .   
docker run --rm -p 8000:8000 pin-inside-demo

You don't need to authenticate in this case because the model object is actually inside the container.

This is awesome!! Thank you so much for this!! With that one little change I was able to get it to work and was able to play around with the predictions. Thanks a million!!