Condition classes signalled by metasurvey.core — metasurvey

All errors and warnings signalled by metasurvey.core are classed conditions: they carry a domain-specific subclass plus the package-wide base class ("metasurvey_error" or "metasurvey_warning"), so you can handle them with tryCatch()/withCallingHandlers() without matching message text.

Value

No return value. This topic documents the condition classes signalled by metasurvey.core; catch them with tryCatch() or withCallingHandlers() as shown in the examples.

Error classes

Every error carries "metasurvey_error" plus one of:

metasurvey_input_error: An argument of an exported function failed input validation (wrong type, wrong class, missing value).
metasurvey_error_step: A step could not be created or baked (unknown step type, missing variables, failed step_validate() checks with .action = "stop").
metasurvey_error_recipe: A recipe is invalid or cannot be applied (missing metadata, unmet variable dependencies, dependency cycles).
metasurvey_error_workflow: A workflow object or a workflow() estimation is invalid.
metasurvey_error_panel: A rotating-panel operation failed (inconsistent waves, invalid extraction).
metasurvey_error_survey: A survey-level operation failed (design construction, weight or replicate specification).
metasurvey_error_engine: The requested processing engine is not supported or not installed.
metasurvey_error_io: A file could not be loaded (unsupported format, missing reader package, failed download).
metasurvey_error_backend: A recipe/workflow backend operation failed.
metasurvey_error_backend_unavailable: The "api" backend was selected but no provider is registered (also carries metasurvey_error_backend). Install a provider package such as metasurvey.explorer.backend.

Warning classes

Every warning carries "metasurvey_warning" plus one of metasurvey_warning_step, metasurvey_warning_recipe, metasurvey_warning_workflow, metasurvey_warning_panel, metasurvey_warning_survey, metasurvey_warning_engine or metasurvey_warning_backend, following the same domains as the error classes.

Examples

# Catch any metasurvey error by its base class
tryCatch(
  set_engine("not-an-engine"),
  metasurvey_error = function(e) message("caught: ", conditionMessage(e))
)
#> caught: Engine 'not-an-engine' is not supported. Available: data.table,
#> tidyverse, dplyr

# Or target a specific domain
tryCatch(
  set_engine("not-an-engine"),
  metasurvey_error_engine = function(e) "engine problem"
)
#> [1] "engine problem"