Merge pull request #21 from velveteer/some-fixes

Adjust vector bounds and add Decoder instances

CHANGELOG.md

@@ -1,5 +1,12 @@
 # Revision history for hermes-json
 
+## 0.5.0.0 -- 2023-03-08
+
+* Convert `DecoderM` callbacks to `Decoder`. This is a breaking change.
+* Add useful instances for the `Decoder` monad
+* Remove some redundant functions
+* Adjust `vector` bounds
+
 ## 0.4.0.0 -- 2023-03-07
 
 * Update simdjson to 3.1.3

README.md

@@ -90,17 +90,17 @@ We benchmark the following operations using both `hermes-json` and `aeson` stric
 <!-- AUTO-GENERATED-CONTENT:START (BENCHES) -->
 | Name                                    | Mean (ps)     | 2*Stdev (ps) | Allocated  | Copied     | Peak Memory |
 | --------------------------------------- | ------------- | ------------ | ---------- | ---------- | ----------- |
-| All.Decode.Arrays.Hermes                | 267104550000  | 18274712758  | 503599934  | 439150544  | 541065216   |
-| All.Decode.Arrays.Aeson                 | 2205838200000 | 108871466542 | 7094759154 | 2392723275 | 1166016512  |
-| All.Decode.Persons.Hermes               | 47153700000   | 3880584170   | 144901928  | 57032737   | 1166016512  |
-| All.Decode.Persons.Aeson                | 134265700000  | 7195219536   | 357269946  | 188529734  | 1166016512  |
-| All.Decode.Partial Twitter.Hermes       | 246448046     | 20934312     | 348540     | 3088       | 1166016512  |
-| All.Decode.Partial Twitter.JsonStream   | 2105484765    | 73210262     | 15261108   | 273820     | 1166016512  |
-| All.Decode.Partial Twitter.Aeson        | 4297434375    | 139205712    | 12547656   | 4625157    | 1166016512  |
-| All.Decode.Persons (Aeson Value).Hermes | 108099550000  | 9966602188   | 303649194  | 138051155  | 1166016512  |
-| All.Decode.Persons (Aeson Value).Aeson  | 119240200000  | 9148201308   | 286148916  | 177027844  | 1166016512  |
-| All.Decode.Twitter (Aeson Value).Hermes | 4261312500    | 149205128    | 12555922   | 4151184    | 1166016512  |
-| All.Decode.Twitter (Aeson Value).Aeson  | 4832229687    | 242990712    | 12539421   | 5527422    | 1166016512  |
+| All.Decode.Arrays.Hermes                | 267914650000  | 10610366160  | 503599934  | 439150544  | 541065216   |
+| All.Decode.Arrays.Aeson                 | 2214928800000 | 160279563772 | 7094759111 | 2392723388 | 1166016512  |
+| All.Decode.Persons.Hermes               | 47338175000   | 4290343628   | 144901928  | 57032737   | 1166016512  |
+| All.Decode.Persons.Aeson                | 132864400000  | 9509102680   | 357269946  | 188529742  | 1166016512  |
+| All.Decode.Partial Twitter.Hermes       | 241083593     | 13856196     | 348540     | 3088       | 1166016512  |
+| All.Decode.Partial Twitter.JsonStream   | 2116192187    | 158907568    | 15259526   | 273821     | 1166016512  |
+| All.Decode.Partial Twitter.Aeson        | 4254060937    | 262619196    | 12538003   | 4634594    | 1166016512  |
+| All.Decode.Persons (Aeson Value).Hermes | 106420425000  | 3747538126   | 303886293  | 135388183  | 1166016512  |
+| All.Decode.Persons (Aeson Value).Aeson  | 119489550000  | 9713032080   | 286148916  | 177027852  | 1166016512  |
+| All.Decode.Twitter (Aeson Value).Hermes | 4164246875    | 240020934    | 12368752   | 4149211    | 1166016512  |
+| All.Decode.Twitter (Aeson Value).Aeson  | 4810817187    | 345165042    | 12539421   | 5527424    | 1166016512  |
 |                                         |
 <!-- AUTO-GENERATED-CONTENT:END (BENCHES) -->
 

hermes-aeson/hermes-aeson.cabal

@@ -12,8 +12,8 @@ library
   exposed-modules:
     Data.Hermes.Aeson
   build-depends:
-    aeson ^>= 2.1.2.1,
-    base  ^>= 4.16.4.0,
+    aeson  >= 2.0.1 && < 2.2,
+    base   >= 4.16.4.0,
     hermes-json
   hs-source-dirs:   src
   ghc-options:      -Wall -O2
@@ -26,7 +26,7 @@ test-suite test
   main-is:          test.hs
   ghc-options: -Wall
   build-depends:
-    aeson          >= 2.0.1 && < 2.2,
+    aeson,
     base,
     bytestring,
     hermes-aeson,

hermes-aeson/src/Data/Hermes/Aeson.hs

@@ -6,11 +6,12 @@ import qualified Data.Aeson.Key as K
 import qualified Data.Aeson.KeyMap as KM
 
 hValueToAeson :: H.Decoder A.Value
-hValueToAeson = H.withType $ \ty ->
+hValueToAeson = do
+  ty <- H.getType
   case ty of
-    H.VArray   -> H.withVector hValueToAeson (pure . A.Array)
-    H.VObject  -> H.withObjectAsMap (pure . K.fromText) hValueToAeson (pure . A.Object . KM.fromMap)
-    H.VNumber  -> H.withScientific (pure . A.Number)
-    H.VString  -> H.withText (pure . A.String)
-    H.VBoolean -> H.withBool (pure . A.Bool)
-    H.VNull    -> H.withNull (\isNil -> if isNil then pure A.Null else fail "expected null")
+    H.VArray   -> A.Array <$> H.vector hValueToAeson
+    H.VObject  -> A.Object . KM.fromMap <$> H.objectAsMap (pure . K.fromText) hValueToAeson
+    H.VNumber  -> A.Number <$> H.scientific
+    H.VString  -> A.String <$> H.text
+    H.VBoolean -> A.Bool <$> H.bool
+    H.VNull    -> pure A.Null

hermes-bench/bench.csv

@@ -1,12 +1,12 @@
 Name,Mean (ps),2*Stdev (ps),Allocated,Copied,Peak Memory
-All.Decode.Arrays.Hermes,267104550000,18274712758,503599934,439150544,541065216
-All.Decode.Arrays.Aeson,2205838200000,108871466542,7094759154,2392723275,1166016512
-All.Decode.Persons.Hermes,47153700000,3880584170,144901928,57032737,1166016512
-All.Decode.Persons.Aeson,134265700000,7195219536,357269946,188529734,1166016512
-All.Decode.Partial Twitter.Hermes,246448046,20934312,348540,3088,1166016512
-All.Decode.Partial Twitter.JsonStream,2105484765,73210262,15261108,273820,1166016512
-All.Decode.Partial Twitter.Aeson,4297434375,139205712,12547656,4625157,1166016512
-All.Decode.Persons (Aeson Value).Hermes,108099550000,9966602188,303649194,138051155,1166016512
-All.Decode.Persons (Aeson Value).Aeson,119240200000,9148201308,286148916,177027844,1166016512
-All.Decode.Twitter (Aeson Value).Hermes,4261312500,149205128,12555922,4151184,1166016512
-All.Decode.Twitter (Aeson Value).Aeson,4832229687,242990712,12539421,5527422,1166016512
+All.Decode.Arrays.Hermes,267914650000,10610366160,503599934,439150544,541065216
+All.Decode.Arrays.Aeson,2214928800000,160279563772,7094759111,2392723388,1166016512
+All.Decode.Persons.Hermes,47338175000,4290343628,144901928,57032737,1166016512
+All.Decode.Persons.Aeson,132864400000,9509102680,357269946,188529742,1166016512
+All.Decode.Partial Twitter.Hermes,241083593,13856196,348540,3088,1166016512
+All.Decode.Partial Twitter.JsonStream,2116192187,158907568,15259526,273821,1166016512
+All.Decode.Partial Twitter.Aeson,4254060937,262619196,12538003,4634594,1166016512
+All.Decode.Persons (Aeson Value).Hermes,106420425000,3747538126,303886293,135388183,1166016512
+All.Decode.Persons (Aeson Value).Aeson,119489550000,9713032080,286148916,177027852,1166016512
+All.Decode.Twitter (Aeson Value).Hermes,4164246875,240020934,12368752,4149211,1166016512
+All.Decode.Twitter (Aeson Value).Aeson,4810817187,345165042,12539421,5527424,1166016512

hermes-json.cabal

@@ -1,6 +1,6 @@
 cabal-version:      3.0
 name:               hermes-json
-version:            0.4.0.0
+version:            0.5.0.0
 category:           Text, Web, JSON, FFI
 synopsis:           Fast JSON decoding via simdjson C++ bindings
 description:
@@ -72,7 +72,7 @@ library
     transformers       >= 0.5.6 && < 0.6,
     time               >= 1.9.3 && < 1.13,
     time-compat        >= 1.9.5 && < 1.10,
-    vector             >= 0.13.0 && < 0.14
+    vector             >= 0.12.3.1 && < 0.14
 
   hs-source-dirs:   src
   default-language: Haskell2010

src/Data/Hermes.hs

@@ -69,14 +69,12 @@ module Data.Hermes
   , withBool
   , withDouble
   , withInt
-  , withNull
   , withObject
   , withObjectAsMap
   , withScientific
   , withString
   , withText
   , withType
-  , withTypeM
   , withVector
   -- * Raw ByteString access
   , withRawByteString

src/Data/Hermes/Decoder/Internal.hs

@@ -105,6 +105,34 @@ instance NFData Path
 
 newtype Decoder a = Decoder { runDecoder :: Value -> DecoderM a }
 
+instance Functor Decoder where
+  {-# INLINE fmap #-}
+  fmap f d = Decoder $ \val -> f <$> runDecoder d val
+
+instance Applicative Decoder where
+  {-# INLINE pure #-}
+  pure a = Decoder $ \_ -> pure a
+  {-# INLINE (<*>) #-}
+  (Decoder f) <*> (Decoder e) = Decoder $ \val -> f val <*> e val
+
+instance Monad Decoder where
+  {-# INLINE return #-}
+  return = pure
+  {-# INLINE (>>=) #-}
+  (Decoder d) >>= f = Decoder $ \val -> do
+    x <- d val
+    runDecoder (f x) val
+
+instance Alternative Decoder where
+  {-# INLINE (<|>) #-}
+  (Decoder a) <|> (Decoder b) = Decoder $ \val -> a val <|> b val
+  {-# INLINE empty #-}
+  empty = Decoder $ const empty
+
+instance MonadFail Decoder where
+  {-# INLINE fail #-}
+  fail e = Decoder $ \_ -> fail e
+
 -- | Decode a strict `ByteString` using the simdjson::ondemand bindings.
 -- Creates simdjson instances on each decode.
 decodeEither :: Decoder a -> BS.ByteString -> Either HermesException a
@@ -139,6 +167,8 @@ parseByteStringIO hEnv d bs =
 -- It is preferable to use `withHermesEnv` to keep foreign references in scope.
 -- Be careful using this, the foreign references can be finalized if the
 -- `HermesEnv` goes out of scope.
+--
+-- Do _not_ share a `HermesEnv` across multiple threads. Each thread should get its own.
 mkHermesEnv :: Maybe Int -> IO HermesEnv
 mkHermesEnv mCapacity = do
   parser   <- mkSIMDParser mCapacity