This package provides search capability to RCloud by taking functionality that was previously provided by search.R from the rcloud.support package.
This package must be used alongside an RCloud version >=1.9
.
install rcloud.solr
with something like:
source("https://install-github.me/att/rcloud.solr")
or:
install.packages("remotes")
remotes::install_github("att/rcloud.solr")
Clearly we need solr running to test this out. A Docker image suitable for testing has been produced at mangothecat/docker-rcloud-solr This is also on Docker hub, so to install it run:
docker pull mangothecat/rcloud-solr
to start the service run something like:
sudo docker run --rm -d -p 8983:8983 mangothecat/rcloud-solr
Your RCloud config will then need the line:
solr.url: http://127.0.0.1:8983/solr/rcloudnotebooks
When you bring up RCloud it should pick up this service.
I recommend using Docker to setup the development environment. The repository contains a Docker file to build an RStudio image (mostly based on rocker/tidyverse). This also installs any dev dependencies that rcloud.solr
needs so rebuild it if these ever change. To build the RStudio image run (from project root):
sudo docker build . -t rcsolr-rstudio
Or you can pull the version that matches the development head from docker hub: docker pull mangothecat/rcsolr-rstudio
.
The repo also contains a Docker Compose file to bring up the whole environment. This includes the RStudio container and an instance of Solr. To run all this in the background run:
sudo docker-compose up -d
and then you can use stop
or down
depending how you want to tear down.
Some of the unit tests require a connection to a solr instance to work. The docker-compose.yml file lays out a configuration that will allow testing. There needs to be a solr instance, setup for RCloud, running on http://solr:8983/
. It is recommnded to use this docker image.
Also note that this package will not work in Windows due to use of parallel forks.