Port over pipeline module.

This commit is contained in:
Dillon Kearns 2020-04-17 21:24:23 -07:00
parent 9226fdeeb5
commit 435724de98
2 changed files with 334 additions and 0 deletions

View File

@ -8,6 +8,7 @@
"Head",
"Head.Seo",
"OptimizedDecoder",
"OptimizedDecoder.Pipeline",
"Pages.Document",
"Pages.ImagePath",
"Pages.PagePath",

View File

@ -0,0 +1,333 @@
module OptimizedDecoder.Pipeline exposing
( required, requiredAt, optional, optionalAt, hardcoded, custom
, decode, resolve
)
{-|
# Json.Decode.Pipeline
Use the `(|>)` operator to build JSON decoders.
## Decoding fields
@docs required, requiredAt, optional, optionalAt, hardcoded, custom
## Beginning and ending pipelines
@docs decode, resolve
### Verified docs
The examples all expect imports set up like this:
import Json.Decode.Exploration exposing (..)
import Json.Decode.Exploration.Pipeline exposing (..)
import Json.Decode.Exploration.Located exposing (Located(..))
import Json.Encode as Encode
import List.Nonempty as Nonempty
For automated verification of these examples, this import is also required.
Please ignore it.
import DocVerificationHelpers exposing (User)
-}
import OptimizedDecoder as Decode exposing (Decoder)
{-| Decode a required field.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> required "name" string
|> required "email" string
""" {"id": 123, "email": "sam@example.com", "name": "Sam"} """
|> decodeString userDecoder
--> Success { id = 123, name = "Sam", email = "sam@example.com" }
-}
required : String -> Decoder a -> Decoder (a -> b) -> Decoder b
required key valDecoder decoder =
decoder |> Decode.andMap (Decode.field key valDecoder)
{-| Decode a required nested field.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> requiredAt [ "profile", "name" ] string
|> required "email" string
"""
{
"id": 123,
"email": "sam@example.com",
"profile": { "name": "Sam" }
}
"""
|> decodeString userDecoder
--> Success { id = 123, name = "Sam", email = "sam@example.com" }
-}
requiredAt : List String -> Decoder a -> Decoder (a -> b) -> Decoder b
requiredAt path valDecoder decoder =
decoder |> Decode.andMap (Decode.at path valDecoder)
{-| Decode a field that may be missing or have a null value. If the field is
missing, then it decodes as the `fallback` value. If the field is present,
then `valDecoder` is used to decode its value. If `valDecoder` fails on a
`null` value, then the `fallback` is used as if the field were missing
entirely.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> optional "name" string "blah"
|> required "email" string
""" { "id": 123, "email": "sam@example.com" } """
|> decodeString userDecoder
--> Success { id = 123, name = "blah", email = "sam@example.com" }
Because `valDecoder` is given an opportunity to decode `null` values before
resorting to the `fallback`, you can distinguish between missing and `null`
values if you need to:
userDecoder2 =
decode User
|> required "id" int
|> optional "name" (oneOf [ string, null "NULL" ]) "MISSING"
|> required "email" string
Note also that this behaves _slightly_ different than the stock pipeline
package.
In the stock pipeline package, running the following decoder with an array as
the input would _succeed_.
fooDecoder =
decode identity
|> optional "foo" (maybe string) Nothing
In this package, such a decoder will error out instead, saying that it expected
the input to be an object. The _key_ `"foo"` is optional, but it really does
have to be an object before we even consider trying your decoder or returning
the fallback.
-}
optional : String -> Decoder a -> a -> Decoder (a -> b) -> Decoder b
optional key valDecoder fallback decoder =
-- source: https://github.com/NoRedInk/elm-json-decode-pipeline/blob/d9c10a2b388176569fe3e88ef0e2b6fc19d9beeb/src/Json/Decode/Pipeline.elm#L113
custom (optionalDecoder (Decode.field key Decode.value) valDecoder fallback) decoder
{-| Decode an optional nested field.
-}
optionalAt : List String -> Decoder a -> a -> Decoder (a -> b) -> Decoder b
optionalAt path valDecoder fallback decoder =
custom (optionalDecoder (Decode.at path Decode.value) valDecoder fallback) decoder
-- source: https://github.com/NoRedInk/elm-json-decode-pipeline/blob/d9c10a2b388176569fe3e88ef0e2b6fc19d9beeb/src/Json/Decode/Pipeline.elm#L116-L148
optionalDecoder : Decode.Decoder Decode.Value -> Decoder a -> a -> Decoder a
optionalDecoder pathDecoder valDecoder fallback =
let
nullOr decoder =
Decode.oneOf [ decoder, Decode.null fallback ]
handleResult input =
case Decode.decodeValue pathDecoder input of
Ok rawValue ->
-- The field was present, so now let's try to decode that value.
-- (If it was present but fails to decode, this should and will fail!)
case Decode.decodeValue (nullOr valDecoder) rawValue of
Ok finalResult ->
Decode.succeed finalResult
Err finalErr ->
-- TODO is there some way to preserve the structure
-- of the original error instead of using toString here?
Decode.fail (Decode.errorToString finalErr)
Err _ ->
-- The field was not present, so use the fallback.
Decode.succeed fallback
in
Decode.value
|> Decode.andThen handleResult
{-| Rather than decoding anything, use a fixed value for the next step in the
pipeline. `harcoded` does not look at the JSON at all.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> hardcoded "Alex"
|> required "email" string
""" { "id": 123, "email": "sam@example.com" } """
|> decodeString userDecoder
--> Success { id = 123, name = "Alex", email = "sam@example.com" }
-}
hardcoded : a -> Decoder (a -> b) -> Decoder b
hardcoded =
Decode.andMap << Decode.succeed
{-| Run the given decoder and feed its result into the pipeline at this point.
Consider this example.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> custom (at [ "profile", "name" ] string)
|> required "email" string
"""
{
"id": 123,
"email": "sam@example.com",
"profile": {"name": "Sam"}
}
"""
|> decodeString userDecoder
--> Success { id = 123, name = "Sam", email = "sam@example.com" }
-}
custom : Decoder a -> Decoder (a -> b) -> Decoder b
custom =
Decode.andMap
{-| Convert a `Decoder (Result x a)` into a `Decoder a`. Useful when you want
to perform some custom processing just before completing the decoding operation.
import Json.Decode.Exploration exposing (..)
type alias User =
{ id : Int
, name : String
, email : String
}
userDecoder : Decoder User
userDecoder =
let
-- toDecoder gets run *after* all the
-- (|> required ...) steps are done.
toDecoder : Int -> String -> String -> Int -> Decoder User
toDecoder id name email version =
if version >= 2 then
succeed (User id name email)
else
fail "This JSON is from a deprecated source. Please upgrade!"
in
decode toDecoder
|> required "id" int
|> required "name" string
|> required "email" string
|> required "version" int
-- version is part of toDecoder,
-- but it is not a part of User
|> resolve
"""
{
"id": 123,
"name": "Sam",
"email": "sam@example.com",
"version": 3
}
"""
|> decodeString userDecoder
--> Success { id = 123, name = "Sam", email = "sam@example.com" }
-}
resolve : Decoder (Decoder a) -> Decoder a
resolve =
Decode.andThen identity
{-| Begin a decoding pipeline. This is a synonym for [Json.Decode.succeed](http://package.elm-lang.org/packages/elm-lang/core/latest/Json-Decode#succeed),
intended to make things read more clearly.
type alias User =
{ id : Int
, email : String
, name : String
}
userDecoder : Decoder User
userDecoder =
decode User
|> required "id" int
|> required "email" string
|> optional "name" string ""
-}
decode : a -> Decoder a
decode =
Decode.succeed