This package contains only unit tests for flipStandardChart and flipChart. Test outputs are machine-specific, so it is recommended that you use the image files included inside the repository rather than local versions.

Tests are run automatically in TravisCI, after a push to flipChartTests, flipChart, flipStandardCharts or other upstream dependencies. Snapshots will be found in tests/testthat/snapshots. Diffing files are in the diff directory and should be compared against files of the same name that are in the accepted directory.

Tests can be added to the R files in the testthat directory. Convenience functions have been provided in flipChartTests so the tests can be created easily. For example,

pp <- plot_ly(x = 1:10, y = 1:10, trace = "scatter")
expect_true(TestWidget(pp, "test1"))

Only the test file (e.g. tests/testthat/test-background.R) should be committed, not the test outputs (PNG files).

The output from running tests locally are not reliable, and should be compared against the output in TravisCI. If working locally, the dependencies can be set up by

install.packages("devtools")
install.packages("testthat")
require(devtools)
install_github("MangoTheCat/visualTest")
install.packages("webshot")
webshot::install_phantomjs()

Because the snapshots are machine-specific, the first time they are run locally, you may need to delete all files inside tests/testthat/snapshots/accepted before regenerating.

To accept new test outputs, simply delete the existing test output inside tests/testthat/snapshots/accepted and the new outputs will saved in that directory next time the tests are run using devtools::test() or testthat_test_file(<filename>) to run a subset of tests.

Note that you should use devtools::test() rather than devtools::check() as check() will give the same errors and warning but will not generate output PNG files in the repository.