Cloud Collaboration with Crunch
Crunch.io provides a cloud-based data store and analytic engine. It has a web client for interactive data exploration and visualization. The crunch package for R allows analysts to interact with and manipulate Crunch datasets from within R. Importantly, this allows technical researchers to collaborate naturally with team members, managers, and clients who prefer a point-and-click interface: because all connect to the same dataset in the cloud, there is no need to email files back and forth continually to share results.
Subscribe to the mailing list to receive notification of releases and to ask general support questions.
crunch can be installed from CRAN with
install.packages("crunch")
The pre-release version of the package can be pulled from GitHub using the remotes package (part of devtools):
# install.packages("remotes")
remotes::install_github("Crunch-io/rcrunch")The crunch R package needs to know what URL to use for the API and what your token is for authentication. For more details on how to get an API token, see the help center article.
The usethis package can help you set environment environment variables, the following code will
open a text editor:
$ R
> if (!require("usethis")) install.packages("usethis")
> usethis::edit_r_environ()
And then you can add environment variables by adding the following (filling in the correct values for for
https://my-awesome-brand.crunch.io/api/ and secret-key-goes-here, with one set of quotation marks
around the url and key).
R_CRUNCH_API="https://my-awesome-brand.crunch.io/api/"
R_CRUNCH_API_KEY="secret-key-goes-here"
Restart your R session, or run command readRenviron("~/.Renviron"), and then you will be authenticated automatically.
Check out listDatasets() to see the names of your existing datasets, which you can load like ds <- loadDataset("The name of my dataset"). New Crunch datasets can be made from a data.frame, a .csv or .sav file, or a URL to a file with newDataset(). See the help for these functions or vignette("crunch") for more information.
crunch requires R version 3.0 or greater. Rstudio has instructions for a variety of platforms
both installing from binaries (recommended) or
from source if necessary.
The tests for the rcrunch package relies on mocks that are compressed into a tarball
so that the package stays within the CRAN 5MB size limit. To avoid git merge conflicts
this file is not checked in so it must be created before you can run tests. It is
created for you if you run make test, or you can create it explicitly by running
make compress-fixtures or Rscript dev-misc/compress-fixtures.R.
See the contribution guidelines.
# make deps
# make test
This installs dependencies and then runs the tests, which installs crunch from your local checkout in the process. If the dependencies fail to install, check the error message. You may need to install libcurl on your system before installing the R packages.
You can run tests in an interactive session, or from the command line, $ make test is all you need. Requires the httptest package for R. You can also specify a specific test file or files to run by adding a "file=" argument, like $ make test file=auth. test_package will do a regular-expression pattern match within the file names. See its documentation in the testthat package.
Testing has two options: unit tests only, and tests that communicate with an API server. This is governed by an environment variable, INTEGRATION, which is false by default, meaning that API integration tests are not run. To modify this and test against the Crunch API, you can run $ make test INTEGRATION=TRUE.
To run integration tests, you will need to specify an API key and API server to communicate with. You can do this by setting the environment variables R_TEST_API and CRUNCH_TEST_API_KEY.
If you are a Crunch developer serving a version of the API/backend with Vagrant or Docker, you will have best results if your R_TEST_API/CRUNCH_TEST_API_KEY (1) is local.crunch.io, thanks to a mapping of localhost to that in your hosts file. In order to avoid self-signed certificate errors use the environment variable R_TEST_VERIFY_SSL=FALSE. You might point at "https://local.crunch.io:28443/api/", for example. Some tests that cannot run successfully in the local environment will be skipped when run against a local.crunch.io URL.
Example of local usage:
$ R_TEST_VERIFY_SSL=FALSE CRUNCH_TEST_API_KEY=t0pSecretK3y! R_TEST_API=https://local.crunch.io:28443/api/ make test INTEGRATION=TRUE file=variable-summaryRun $ make doc. Requires the roxygen2 package.