R-ereefs-debugging

This respository contains a docker environment definition for use in debugging problems accessing eReefs datasets from R

Usage

Step 1. Set up a bunch of environment variables to define your test environment

These will be referenced in docker commands later on.

export CURL_VERSION="7.88.1"
export HDF5_VERSION="1.14.0"
export NETCDF_VERSION="4.9.2"
export R_VERSION="4.3.0"
export BASE_IMAGE="r-base:${R_VERSION}"
export R_SCRIPT="opendap_ncml_test.r"

 

Step 2. Build the Docker Image

Setting explicit build arguments according to the test environment you're trying to build.

This example uses the environment variables exported in the previous step.

docker build \
  --build-arg "CURL_VERSION=${CURL_VERSION}" \
  --build-arg "HDF5_VERSION=${HDF5_VERSION}" \
  --build-arg "NETCDF_VERSION=${NETCDF_VERSION}" \
  --build-arg "R_VERSION=${R_VERSION}" \
  --build-arg "BASE_IMAGE=${BASE_IMAGE}" \
  --tag "r-ereefs-debugging:r${R_VERSION}-curl${CURL_VERSION}-hdf5${HDF5_VERSION}-netcdf${NETCDF_VERSION}" \
  .

The full set of supported build arguments are:

• CURL_VERSION => The version of curl (and libcurl) that should be used for testing. Default is 8.1.2, released 2023-05-30
• HDF5_VERSION => The version of the HDF 5 library that should be compiled for testing. Default is 1.14.0, released 2023-02-08. (Note for HDF5 v1.14+, you need netcdf 4.9.2 or later)
• NETCDF_VERSION => The version of the netCDF-C library that should be compiled for testing. Default is 4.9.2, released 2023-03-14
• R_LIBRARIES => Space-seperated list of R library packages that should be installed from CRAN. Default is dplyr ncdf4 tidync
• R_VERSION => The version of R that should be used. Default is 4.3.0
• BASE_IMAGE => The base docker image that should be used. Default is r-base:${R_VERSION}

The build may take quite a while, as it will attempt to compile the CURL, HDF5 and netCDF-C libraries from source, which is slow!

 

Step 3. Use the docker image to run your test script

Bind-mount the R script you want to run into the /app working directory, and also specify the name of that script as the run command. Save the results to a logfile for later comparison.

This example uses the environment variables exported in Step 1.

docker run \
    --volume ${PWD}/${R_SCRIPT}:/app/${R_SCRIPT} \
    r-ereefs-debugging:r${R_VERSION}-netcdf${NETCDF_VERSION}-curl${CURL_VERSION} \
    ${R_SCRIPT} \
    > "${R_SCRIPT}_r${R_VERSION}-netcdf${NETCDF_VERSION}-curl${CURL_VERSION}.log" 2>&1