Check that |a - b| < eps. Works with scalars and vectors on any input. For vector input, one may defined, how many |a - b| >= eps are acceptable with r argument. Given it is a "workhorse" used in almost all test-files, this function is also very well tested.

Usage

expect_eps(a, b, eps, r = 0, relative = FALSE, note = NULL, debug = FALSE)

Arguments

a: numeric scalar or vector a to be compared
b: numeric scalar or vector b to be compared
eps: numeric scalar or vector, setting the max differences, eps > 0
r: optional numeric scalar (r = 0 in default), relative number of values, that may have an difference > eps.
relative: bool argument, if set will take the normale difference with euler metric. Default = FALSE
note: optional parameter used for debugging.
debug: bool argument, if set will printout the difference, in case of failure. Default = FALSE Used internally, too calculate acceptable amount of deviances. Calculated absolute value will be floored.

Value

success or failure with message

Details

For vector/scalar combinations, allowed are (r has to always be a scalar!):

all scalar
a vector, b scalar, eps scalar -> compare a values to be close enough to b
a scalar, b vector, eps scalar -> compare b values to be close enough to a
a scalar, b scalar, eps vector -> compare the difference a, b too all eps -> case not intended, but would work
a vector, b scalar, eps vector -> compare each vector-difference entry against each eps
all vectors -> each entry of |a - b| is compared to the same entry in eps -> different vector lengths != 1 dissallowed!expect_brms_family

Examples

print(bayesfam:::expect_eps(1, 1.1, 0.2)) # should pass
#> <expectation_success/expectation/condition>
#> As expected
# print(expect_error(bayesfam:::expect_eps(c(0, 1, 3), c(1, 1, 2), 1e-4, 1 / 3)))
# should fail (2/3 were wrong, but only 1/3 was allowed)