This library provides a way to express validations through Validator objects holding the validation logic and allowing for composition.
A Validator
holds validations and corresponding errors for a model.
The internals of this data type are not exposed, instead see the constructors
for more details.
model -> Basics.Bool
A Validation
is a function that validates a model.
validate : Validator model error -> model -> List error
Validate a model using a validator, returning a list of errors. The list
will be empty when the model is valid. If you only care about the validity of
the model, consider using the more efficient isValid
instead.
atLeast8Chars = simple (\s -> String.length s >= 8) "should be at least 8 chars long"
validate atLeast8Chars "2short" == ["should be at least 8 chars long"]
validate atLeast8Chars "long enough" == []
isValid : Validator model error -> model -> Basics.Bool
Checks wether a model is valid according to a validator. If you care about
the reasons why the model is invalid, consider using validate
instead.
bigNumber = simple (flip (>) 1000) "Not such a big number"
isValid bigNumber 1001 == True
isValid bigNumber 1000 == False
simple : Validation model -> error -> Validator model error
A simple validator that returns an error constant.
stayPositive = (simple (flip (>) 0) "only positive numbers are allowed here stranger")
validate stayPositive 0 == ["only positive numbers are allowed here stranger"]
parameterized : Validation model -> (model -> error) -> Validator model error
A parameterized validator that composes a single error from the model.
justMonika = parameterized ((==) "Monika") (\name -> "Not " ++ name ++ "! Just Monika.")
validate justMonika "Yuri" == ["Not Yuri! Just Monika."]
invalid : error -> Validator model error
A validator that is always invalid. It will always provide the same error.
validate (invalid "Bad, bad number!") 111 == [ "Bad, bad number!" ]
isValid (invalid "Some error message") "blah" == True
valid : Validator model error
A validator that is always valid. It will always pass, never provide error.
validate valid 123 == []
isValid valid "something" == True
all : List (Validator model error) -> Validator model error
Converts a list of validator into one new validator of the same type. This is useful to aggregate multiple validations on the same model (fields in an record, more than one check on a value, etc.).
moreThan2 = simple (flip (>) 2) "must be greater than 2"
notFactorOf3 = simple (\n -> n % 3 /= 0) "must not be a factor of 3"
lessThan1000 = simple (flip (<) 1000) "must be lower than 1000"
allThose =
all
[ moreThan2
, notFactorOf3
, lessThan1000
]
validate allThose 1 == ["must be greater than 2"]
validate allThose 333 == ["must not be a factor of 3"]
validate allThose 1000 == ["must be lower than 1000"]
validate allThose 12345 == ["must not be a factor of 3", "must be lower than 1000"]
validate allThose 124 == []
isValid allThose 778 == True
isValid allTHose 12 == False
mapError : (errorA -> errorB) -> Validator model errorA -> Validator model errorB
Decorates a validator by modifying the errors it returns. The transformation function handles the transformation from one error type to the other.
(fans of functional programming will recognize profunctor's rmap)
onlyTrue = simple ((==) True) "This is simply not True!"
onlyMoreTrue = onlyTrue |> mapError (\error -> (error, "It is simply False!"))
validate onlyTrue False == ["This is simply not True!"]
validate onlyMoreTrue False == ("This is simply not True!", "It is simply False!")
mapErrorWithModel : (model -> errorA -> errorB) -> Validator model errorA -> Validator model errorB
Same as mapError but with the model as an additional firt parameter to the iterator function.
onlyTrue = simple ((==) True) "This is simply not True!"
onlyMoreTrue = onlyTrue |> mapErrorWithModel (\model error -> (error, "It is simply " ++ toString model ++ !"))
validate onlyTrue False == ["This is simply not True!"]
validate onlyMoreTrue False == ("This is simply not True!", "It is simply False!")
focus : (modelB -> modelA) -> Validator modelA error -> Validator modelB error
Creates a new validator to check another model. The transformation function takes the model provided to the new validator, extracts the model expected by the original validator and feeds it instead. Useful to focus on a field of a record, part of a list, etc.
(fans of functional programming will recognize profunctor's lmap)
type alias Fighter = { name: String, strength : Int }
under9000 = simple (flip (<=) 9000) "It's over 9000!!!"
onlyHuman = focus .strength under9000
validate onlyHuman { name = "Satan", strength = 9 } == []
validate onlyHuman { name = "Goku", strength = 999999 } == [ "It's over 9000!!!" ]
focusError : (modelB -> modelA) -> (errorA -> errorB) -> Validator modelA errorA -> Validator modelB errorB
Takes a validator and transform it to work on another model and change the
errors. Shortcut for focus
then mapError
.
(fans of functional programming will recognize profunctor's dimap)
type alias User =
{ login : String, password : String }
passwordValidator =
simple ((/=) "password") "'password' is not a very good password"
userValidator =
focusError .password (\{login} error -> "for realz, " ++ login ++ " " ++ error) passwordValidator
user = { name = "Carl Streator", password = "password" }
validate userValidator user == ["for realz Carl Streator 'password' is not a very good password"]
focusInside : (model -> Validator model error) -> Validator model error
Focuses inside a model to make validations that depend on more than one field (typically, validations that require to check the value of another field to be validated).
under value = simple (\n -> n < value) "is not under (" ++ toString value ++ ")"
minUnderMax = focusInside (\{ max } -> focusError .min ((++) "min ") (under max))
isValid minUnderMax { min = 1, max = 5 } == True
validate minUnderMax { min = 1, max = 5 } == []
isValid minUnderMax { min = 5, max = 5 } == False
validate minUnderMax { min = 5, max = 5 } == ["min is not under (5)"]
maybe : (modelB -> Maybe modelA) -> Validator modelA error -> Validator modelB error
Focuses on a value that may or may not be available for validation. When there is nothing, the validation succeeds by default.
notBlank = simple ((/=) "") "filled or nothing"
nameNotBlank = maybe .name notBlank
isValid notBlank { name = Nothing } == True
validate notBlank { name = Nothing } == []
isValid notBlank { name = Just "Someone" } == True
validate notBlank { name = Just "Me" } == []
isValid notBlank { name = Just "" } == False
validate notBlank { name = Just "" } == ["filled or nothing"]
list : (Basics.Int -> errorA -> errorB) -> Validator model errorA -> Validator (List model) errorB
Makes a validator that applies another validator to a list of elements. The transformation function receives the index of the element under scrutiny as well as the error obtained by the internal validator.
model Cup = {owner: String, temperature: Int}
cups =
[ {owner = "Mama Bear", temperature = 100}
, {owner = "Papa Bear", temperature = 20}
, {owner = "Little Bear", temperature = 30}
]
tooHot = simple (\cup -> cup.temperature > 30) "it's too hot!"
tooCold = simple (\cup -> cup.temperature < 30) "it's too cold!"
justRight = all [tooHot, tooCold]
goldilocks = list
(\index model error -> (index, error))
justRight
validate goldilocks cups ==
[ (1, "it's too hot!")
, (2, "it's too cold!")
]