Check a set of keys and their values

check_values(
  x,
  annotations,
  allowlist_keys = NULL,
  allowlist_values = NULL,
  success_msg = "All annotation values are valid",
  fail_msg = "Some annotation values are invalid",
  return_valid = FALSE,
  annots_link = "https://shinypro.synapse.org/users/nsanati/annotationUI/",
  syn
)

Arguments

x	A data frame of annotation data
annotations	A data frame of annotations to check against
allowlist_keys	A character vector of annotation keys to allowlist. If provided, all values for the given keys will be treated as valid.
allowlist_values	A named list of keys (as the names) and values (as vectors) to allowlist
success_msg	Message indicating the check succeeded.
fail_msg	Message indicating the check failed.
return_valid	Should the function return valid values? Defaults to `FALSE` (i.e. the function will return invalid values).
annots_link	Link to a definition of the annotations being used in the project
syn	Synapse client object

Value

If return_valid = FALSE: a condition object indicating whether all annotation values are valid. Invalid annotation values are included as data within the object: a named list where each element corresponds to a key that contains invalid values, and the contents of each element is a vector of invalid values. If return_valid = TRUE: a named list of the valid annotation keys and values.

Examples

annots <- data.frame(
  key = c("fileFormat", "fileFormat"),
  value = c("txt", "csv"),
  columnType = c("STRING", "STRING"),
  stringsAsFactors = FALSE
)
dat <- data.frame(
  fileFormat = c("wrong", "txt", "csv", "wrong again"),
  stringsAsFactors = FALSE
)
check_values(dat, annots)
#> <error/check_fail>
#> Some annotation values are invalid

## Comma-separated and json-style strings for keys with enumerated values
## are separated and checked individually
dat <- data.frame(
  fileFormat = c("wrong, txt, csv", "[\"wrong again\",\"csv\"]"),
  stringsAsFactors = FALSE
)
check_values(dat, annots)
#> <error/check_fail>
#> Some annotation values are invalid