json-to-elm
Create Elm type aliases and decoders based on JSON input
This project allows you to automate the creation of:
- type aliases from JSON data
- decoders from type aliases and some union types
- encoders from type aliases and some union types
You can run this program as a command line tool:
Noahs-MacBook-Pro:json-to-elm noah$ python generate.py ../elm-static-site/examples/Users.elm
decodeUser : Decoder User
decodeUser =
succeed User, : ("name" := string), : ("location" := string), : ("age" := int)
encodeUser : User -> Json.Encode.Value
encodeUser record =
object
[ ("name", string record.name)
, ("location", string record.location)
, ("age", int record.age)
]
You can also give it a json file.
or you can use the following functions:
print_everything takes in a string containing a JSON object, and a top-level name for the alias it will generate.
It will then recursively generate aliases and decoders for all the JSON objects in the json object.
create_type_alias takes a json object, and returns a list of type aliases needed for that object
create_decoder takes a type alias as a string, along with a prefix to use for custom decoders, and if the encoded fields have snakecase or not, and generates just the decoder for that type alias
create_encoder takes a type alias as a string, along with a prefix to use for custom encoders, and if the encoded fields have snakecase or not, and generates just the encoder for that type alias
create_union_type_decoder takes a union type definition as a string, and will generate the decoder needed if the json value is a string
create_union_type_encoder takes a union type definition as a string, and will generate the encoder needed if the json value is a string
Example:
print_everything(
"""
{ "name" : "Noah"
, "age" : 23
, "location" :
{ "name" : "sweden"
, "days" : 45
}
}
"""
, alias_name = "Person")
generates
type alias Location =
{ name : String
, days : Int
}
type alias Person =
{ age : Int
, name : String
, location : Location
}
decodeLocation : Json.Decode.Decoder Location
decodeLocation =
Json.Decode.succeed Location, : ("name" := Json.Decode.string), : ("days" := Json.Decode.int)
decodePerson : Json.Decode.Decoder Person
decodePerson =
Json.Decode.succeed Person, : ("age" := Json.Decode.int), : ("name" := Json.Decode.string), : ("location" := decodeLocation)
encodeLocation : Location -> Json.Encode.Value
encodeLocation record =
Json.Encode.object
[ ("name", Json.Decode.string record.name)
, ("days", Json.Decode.int record.days)
]
encodePerson : Person -> Json.Encode.Value
encodePerson record =
Json.Encode.object
[ ("age", Json.Decode.int record.age)
, ("name", Json.Decode.string record.name)
, ("location", encodeLocation record.location)
]
for more examples of this, see the test function
Union types
print(create_union_type_decoder('type Suit = Hearts, Diamonds, Spades, Clubs'))
print(create_union_type_encoder('type Suit = Hearts, Diamonds, Spades, Clubs'))
will print
decodeSuit : Json.Decode.Decoder Suit
decodeSuit =
let
decodeToType string =
case string of
"Hearts" -> Result.Ok Hearts
"Diamonds" -> Result.Ok Diamonds
"Spades" -> Result.Ok Spades
"Clubs" -> Result.Ok Clubs
_ -> Result.Err ("Not valid pattern for decoder to Suit. Pattern: " ++ (toString string))
in
Json.Decode.customDecoder Json.Decode.string decodeToType
encodeSuit : Suit -> Json.Encode.Value
encodeSuit item =
case item of
Hearts -> Json.Encode.string "Hearts"
Diamonds -> Json.Encode.string "Diamonds"
Spades -> Json.Encode.string "Spades"
Clubs -> Json.Encode.string "Clubs"