Testing suite for PoolPrev and HierPoolPrev

[caitlinch]

Develop testing for PoolPrev() and HierPoolPrev(), including checking input data structure and format.

[caitlinch]

Implemented tests for input data frame in tests/testthat/test-input_checks.R:

• Warnings:
  • Check whether empty rows are present
  • Check whether rows containing only NA are present
  • Check whether rows have one or more missing value
• Errors:
  • Check that the result column is present
  • Check that the poolSize column is present
  • Check that results column values are numeric or integer
  • Check that poolSize column values are numeric or integer
  • Check that result column contains only 0 and 1

[caitlinch]

Completed checks:

• HierPoolPrev() raises no errors for trimmed SimpleExampleData
• PoolPrev() returns correct ML point estimate for SimpleExampleData
• PoolPrev() returns correct ML and Bayesian estimates for SimpleExampleData

[caitlinch]

To do:

1. Add tests for bad data input in PoolPrev(), HierPoolPrev()
2. Correct test for HierPoolPrev() to handle errors within the statistical modelling section. The current test for HierPoolPrev() sometimes fails with the following message:

  Expected `expect_no_warning(...)` to run without any errors.
  i Actually got a <expectation_failure> with text:
      Expected `HierPoolPrev(...)` to run without any warnings.
  i Actually got a <simpleWarning> with text:
    There were 8 divergent transitions after warmup. 
    See https://mc-stan.org/misc/warnings.html#divergent-transitions-after-warmup
    to find out why this is a problem and how to eliminate them.

This test needs to be updated to check for PoolTestR errors only (if divergent transitions occur, the test should still pass).