Nothing
# Unit tests for the rule engine (R/detect_rules.R) and candidate
# proposer (R/detect_candidates.R).
# Reach private symbols via masque:::.
.cands <- masque:::.propose_candidates
.rules <- masque:::.rules_all
# --- candidate proposer ------------------------------------------------------
test_that(".propose_candidates() includes factor type at any cardinality", {
df <- data.frame(
trt = factor(rep(LETTERS[1:24], each = 3)),
yield = rnorm(72),
rep = factor(rep(1:3, times = 24)),
stringsAsFactors = FALSE
)
cands <- .cands(df)
expect_true("trt" %in% cands$factors)
expect_true("rep" %in% cands$factors)
expect_false("yield" %in% cands$factors)
})
test_that(".propose_candidates() excludes unique-per-row columns", {
df <- data.frame(id = 1:50, trt = factor(rep(1:5, each = 10)))
cands <- .cands(df)
expect_false("id" %in% cands$factors)
expect_true("trt" %in% cands$factors)
})
test_that(
".propose_candidates() respects user-roled outcomes / keeps / ignores",
{
df <- data.frame(
x = factor(rep(1:5, each = 10)),
y = rnorm(50),
keep_me = factor(rep(c("north", "south"), 25)),
num_keep = seq_len(50),
other = factor(rep(c("a", "b"), 25)),
stringsAsFactors = FALSE
)
roles <- propose_roles(df, detect = FALSE)
roles$role[roles$col == "x"] <- "outcome"
roles$role[roles$col == "keep_me"] <- "keep"
roles$role[roles$col == "num_keep"] <- "keep"
roles$role[roles$col == "other"] <- "ignore"
cands <- .cands(df, roles = roles)
expect_false("x" %in% cands$factors)
expect_false("keep_me" %in% cands$factors)
expect_false("num_keep" %in% cands$numerics)
expect_false("other" %in% cands$factors)
}
)
test_that(".detect_spatial_pair() returns NULL without integer-named row+col", {
df <- data.frame(
rep = 1:9, row = 1:9, what = "x",
stringsAsFactors = FALSE
)
expect_null(masque:::.detect_spatial_pair(
df, names(df),
vapply(df, masque:::col_kind, character(1L))
))
})
test_that(".detect_spatial_pair() finds row/col when both gridded", {
df <- expand.grid(row = 1:4, col = 1:5)
sp <- masque:::.detect_spatial_pair(
df, names(df),
vapply(df, masque:::col_kind, character(1L))
)
expect_equal(sp$row, "row")
expect_equal(sp$col, "col")
expect_equal(sp$n_row, 4L)
expect_equal(sp$n_col, 5L)
})
# --- per-rule unit tests ----------------------------------------------------
test_that("rule_crd scores high on a balanced 1-factor design", {
df <- data.frame(
trt = factor(rep(LETTERS[1:4], each = 10)),
y = rnorm(40),
stringsAsFactors = FALSE
)
cands <- .cands(df)
r <- .rules$CRD(df, cands)
expect_equal(r$class_label, "CRD")
expect_gte(r$score, 0.8)
})
test_that("rule_crd is demoted when a clear block exists", {
df <- expand.grid(rep = 1:4, trt = factor(letters[1:5]))
cands <- .cands(df)
r <- .rules$CRD(df, cands)
expect_lt(r$score, 0.5)
})
test_that("rule_rcbd requires a design-named block (factorial rejected)", {
# Two crossed treatments, neither named like a block.
df <- expand.grid(supp = factor(c("VC", "OJ")), dose = c(0.5, 1.0, 2.0))
df <- df[rep(seq_len(nrow(df)), 10), ]
cands <- .cands(df)
r <- .rules$RCBD(df, cands)
expect_equal(r$score, 0)
expect_match(r$evidence$reason, "design-named|named-block",
ignore.case = TRUE
)
})
test_that("rule_rcbd scores high with a named block + balanced treatment", {
df <- expand.grid(rep = 1:5, trt = factor(letters[1:4]))
cands <- .cands(df)
r <- .rules$RCBD(df, cands)
expect_gte(r$score, 0.8)
expect_equal(r$evidence$treatment_col, "trt")
expect_equal(r$evidence$block_col, "rep")
})
test_that("rule_ibd_alpha finds the pairwise interaction (rep:block)", {
skip_if_not_installed("agridat")
d <- agridat::john.alpha
cands <- .cands(d)
r <- .rules$`IBD/alpha-lattice`(d, cands)
expect_gte(r$score, 0.9)
expect_setequal(r$evidence$block_basis, c("rep", "block"))
expect_true(r$evidence$block_is_pairwise)
})
test_that("rule_row_column requires within-row/col balance, not just a grid", {
# CRD with row/col coordinates -> grid present but trt unbalanced per row/col
set.seed(1)
df <- data.frame(
trt = factor(sample(LETTERS[1:4], 32, replace = TRUE)),
row = rep(1:4, times = 8),
col = rep(1:8, each = 4)
)
cands <- .cands(df)
r <- .rules$`row-column`(df, cands)
# Score may be > 0 due to fill_pct base, but well below 0.85 ideal Latin.
expect_lt(r$score, 0.85)
})
test_that("rule_split_plot needs a named block + 2 non-block-named factors", {
# Only 1 non-block factor -> reject.
df <- expand.grid(rep = 1:3, trt = factor(letters[1:4]))
cands <- .cands(df)
r <- .rules$`split-plot`(df, cands)
expect_equal(r$score, 0)
})
test_that("rule_factorial excludes pairs where either factor is design-named", {
df <- expand.grid(rep = 1:3, trt = factor(letters[1:4]))
cands <- .cands(df)
r <- .rules$factorial(df, cands)
expect_equal(r$score, 0) # rep is block_named -> excluded
})
test_that("rule_factorial scores high on a 2-treatment crossing", {
df <- expand.grid(supp = factor(c("VC", "OJ")), dose = c(0.5, 1.0, 2.0))
df <- df[rep(seq_len(nrow(df)), 10), ]
cands <- .cands(df)
r <- .rules$factorial(df, cands)
expect_gte(r$score, 0.7)
})
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.