highjson


Intro
Hackage: highjson
Low boilerplate, easy to use and very fast Haskell JSON serialisation and parsing. WARNING: Work in progress!
Usage
{-# LANGUAGE OverloadedStrings #-}
data SomeDummy
   = SomeDummy
   { sd_int :: Int
   , sd_bool :: Bool
   , sd_text :: T.Text
   , sd_either :: Either Bool T.Text
   , sd_maybe :: Maybe Int
   } deriving (Show, Eq)
someDummySpec =
    JsonSpec SomeDummy $
    "int" .= sd_int
    :+: "bool" .= sd_bool
    :+: "text" .= sd_text
    :+: "either" .= sd_either
    :+: "maybe" .=? sd_maybe
    :+: EmptySpec
instance ToJson SomeDummy where
    toJson = makeSerialiser someDummySpec
instance JsonReadable SomeDummy where
    readJson = makeParser someDummySpec
test =
    parseJsonBs "{\"int\": 34, \"text\": \"Teext\", \"bool\": true, \"either\": false}"
        == Right (SomeDummy 34 True "Teext" (Left False) Nothing)
For more usage examples check the tests.
Install
- Using cabal: cabal install highjson
- From Source: git clone https://github.com/agrafix/highjson.git && cd highjson && cabal install
Todo
- Implement proper string parsing (handle escape charaters)
- Write more tests
- Generate typescript interfaces from object specs
- …
Benchmarks
To run the benchmarks, use cabal bench. Current results on my MacBook Pro:
$ cabal bench
Preprocessing library highjson-0.2.0.0...
In-place registering highjson-0.2.0.0...
Preprocessing benchmark 'highjson-benchmarks' for highjson-0.2.0.0...
Running 1 benchmarks...
Benchmark highjson-benchmarks: RUNNING...
benchmarking twitter/aeson
time                 2.148 ms   (2.102 ms .. 2.187 ms)
                     0.997 R²   (0.995 R² .. 0.999 R²)
mean                 2.137 ms   (2.112 ms .. 2.169 ms)
std dev              99.12 μs   (81.84 μs .. 120.1 μs)
variance introduced by outliers: 31% (moderately inflated)
benchmarking twitter/highjson
time                 2.196 ms   (2.162 ms .. 2.235 ms)
                     0.998 R²   (0.996 R² .. 0.999 R²)
mean                 2.222 ms   (2.195 ms .. 2.252 ms)
std dev              94.52 μs   (75.45 μs .. 125.1 μs)
variance introduced by outliers: 28% (moderately inflated)
The benchmarks are derived from aeson’s
twitter-json-parsing benchmarks and should probably more there when this library is in a
more complete state.