Checks the structure and domain of a center-configurable ancestry
compatibility rules table (issue #168). Each row is one unordered pair of
standardized ancestry levels plus a severity: block rules exclude
the pairing during breeding-group formation, flag rules annotate it
afterward. It mirrors checkKinshipOverrides: it stop()s
on structural or domain errors and returns the coerced table when the input
is acceptable. An empty table (zero rules) is valid.
Value
The validated rules data.frame with ancestry1 and
ancestry2 coerced to uppercase character and severity to
lowercase character.
Details
Ancestry levels must come from convertAncestry's standardized
vocabulary: CHINESE, INDIAN, HYBRID, JAPANESE, OTHER, UNKNOWN. Levels are
coerced to uppercase and severity to lowercase before validation, so
a hand-edited file's casing never matters. A rule may pair a level with
itself (e.g. HYBRID with HYBRID); duplicated unordered pairs are a data
error the user must resolve. Because convertAncestry maps a
blank ancestry to UNKNOWN but any unrecognized text – including a literal
re-standardized "UNKNOWN" string – to OTHER, a table that names one
of UNKNOWN/OTHER without the other draws a warning here: a center wanting
conservative treatment of animals without usable ancestry information
almost always wants both.
Examples
rules <- data.frame(
ancestry1 = c("INDIAN", "INDIAN"),
ancestry2 = c("CHINESE", "HYBRID"),
severity = c("block", "flag"), stringsAsFactors = FALSE
)
checkAncestryRules(rules)
#> ancestry1 ancestry2 severity
#> 1 INDIAN CHINESE block
#> 2 INDIAN HYBRID flag