rdf4h 1.3.2 → 1.3.3

raw patch · 15 files changed

+106/−120 lines, 15 filesdep ~fgldep ~textPVP: major bump suggested

API removals or changes: PVP suggests a major version bump

Dependency ranges changed: fgl, text

API changes (from Hackage documentation)

Files

• bench/MainCriterion.hs +46/−47
• rdf4h.cabal +12/−37
• src/Data/RDF/Namespace.hs +1/−1
• src/Data/RDF/PatriciaTreeGraph.hs +2/−8
• src/Data/RDF/Query.hs +21/−4
• src/Data/RDF/Types.hs +1/−2
• src/Rdf4hParseMain.hs +2/−2
• src/Text/RDF/RDF4H/Interact.hs +1/−1
• src/Text/RDF/RDF4H/NTriplesParser.hs +3/−3
• src/Text/RDF/RDF4H/NTriplesSerializer.hs +4/−3
• src/Text/RDF/RDF4H/ParserUtils.hs +3/−3
• src/Text/RDF/RDF4H/TurtleParser.hs +3/−3
• src/Text/RDF/RDF4H/TurtleSerializer.hs +2/−2
• src/Text/RDF/RDF4H/XmlParser.hs +4/−3
• testsuite/tests/W3C/TurtleTest.hs +1/−1

@@ -1,15 +1,16 @@-{-# LANGUAGE OverloadedStrings, LambdaCase #-}+{-# LANGUAGE OverloadedStrings, LambdaCase, RankNTypes #-}  module Main where  import Criterion import Criterion.Main import Data.RDF-import qualified Data.Text as T+import qualified Data.Text.Lazy as T --- The `countries.ttl` Turtle file is needed to run this benchmark suite+-- The `bills.102.rdf` XML file is needed to run this benchmark suite ----- $ wget http://telegraphis.net/data/countries/countries.ttl+-- $ wget https://www.govtrack.us/data/rdf/bills.102.rdf.gz+-- $ gzip -d bills.102.rdf.gz  parseTurtle :: RDF rdf => String -> rdf parseTurtle s =@@ -24,7 +25,7 @@  main :: IO () main = defaultMain [-   env (readFile "countries.ttl") $ \ ~(ttl_countries) ->+   env (readFile "bills.102.ttl") $ \ ~(ttl_countries) ->    bgroup "parse" [      bench "TriplesGraph" $        nf (parseTurtle  :: String -> TriplesGraph) ttl_countries@@ -35,58 +36,56 @@    ]     ,-   env (do ttl_countries <- readFile "countries.ttl"+   env (do ttl_countries <- readFile "bills.102.ttl"            let (Right rdf1) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            let (Right rdf2) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            let (Right rdf3) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            return (rdf1 :: PatriciaTreeGraph,rdf2 :: TriplesGraph,rdf3 :: MGraph) )      $ \ ~(patriciaTreeCountries,triplesGraph,mGraph) ->-   bgroup "query" [-     bench "TriplesGraph" $-       nf (queryGr  :: (Maybe Node,Maybe Node,Maybe Node,TriplesGraph) -> [Triple])-              (Just (UNode "http://telegraphis.net/data/countries/AF#AF"),-               Just (UNode "geographis:capital"),-               Just (UNode "http://telegraphis.net/data/capitals/AF/Kabul#Kabul"),-               triplesGraph)-   , bench "MGraph" $-       nf (queryGr  :: (Maybe Node,Maybe Node,Maybe Node,MGraph) -> [Triple])-              (Just (UNode "http://telegraphis.net/data/countries/AF#AF"),-               Just (UNode "geographis:capital"),-               Just (UNode "http://telegraphis.net/data/capitals/AF/Kabul#Kabul"),-               mGraph)-   , bench "PatriciaTreeGraph" $-       nf (queryGr  :: (Maybe Node,Maybe Node,Maybe Node,PatriciaTreeGraph) -> [Triple])-              (Just (UNode "http://telegraphis.net/data/countries/AF#AF"),-               Just (UNode "geographis:capital"),-               Just (UNode "http://telegraphis.net/data/capitals/AF/Kabul#Kabul"),-               patriciaTreeCountries)-   ]+   bgroup "query"+     (queryBench "TriplesGraph" triplesGraph+     ++ queryBench "MGraph" mGraph+     ++ queryBench "PatriciaTreeGraph" patriciaTreeCountries)     ,-   env (do ttl_countries <- readFile "countries.ttl"+   env (do ttl_countries <- readFile "bills.102.ttl"            let (Right rdf1) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            let (Right rdf2) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            let (Right rdf3) = parseString (TurtleParser Nothing Nothing) (T.pack ttl_countries)            return (rdf1 :: PatriciaTreeGraph,rdf2 :: TriplesGraph,rdf3 :: MGraph) )      $ \ ~(patriciaTreeCountries,triplesGraph,mGraph) ->-   bgroup "select" [-     bench "TriplesGraph" $-       nf (selectGr  :: (NodeSelector,NodeSelector,NodeSelector,TriplesGraph) -> [Triple])-              (Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               triplesGraph)-   , bench "MGraph" $-       nf (selectGr  :: (NodeSelector,NodeSelector,NodeSelector,MGraph) -> [Triple])-              (Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               mGraph)-   , bench "PatriciaTreeGraph" $-       nf (selectGr  :: (NodeSelector,NodeSelector,NodeSelector,PatriciaTreeGraph) -> [Triple])-              (Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               Just (\case { (UNode n) -> T.length n > 12 ; _ -> False } ),-               patriciaTreeCountries)-   ]+   bgroup "select"+     (selectBench "TriplesGraph" triplesGraph+     ++ selectBench "MGraph" mGraph+     ++ selectBench "PatriciaTreeGraph" patriciaTreeCountries)  ]++selectBench :: forall rdf. RDF rdf => String -> rdf -> [Benchmark]+selectBench label gr =+   [ bench (label ++ " SPO") $ nf selectGr (subjSelect,predSelect,objSelect,gr)+   , bench (label ++ " SP")  $ nf selectGr (subjSelect,predSelect,selectNothing,gr)+   , bench (label ++ " S")   $ nf selectGr (subjSelect,selectNothing,selectNothing,gr)+   , bench (label ++ " PO")  $ nf selectGr (selectNothing,predSelect,objSelect,gr)+   , bench (label ++ " O")   $ nf selectGr (selectNothing,selectNothing,objSelect,gr)+   ]++subjSelect, predSelect, objSelect, selectNothing :: Maybe (Node -> Bool)+subjSelect = Just (\case { (UNode n) -> T.length n > 12 ; _ -> False })+predSelect = Just (\case { (UNode n) -> T.length n > 12 ; _ -> False })+objSelect  = Just (\case { (UNode n) -> T.length n > 12 ; _ -> False })+selectNothing = Nothing++subjQuery, predQuery, objQuery, queryNothing :: Maybe Node+subjQuery = Just (UNode "http://www.rdfabout.com/rdf/usgov/congress/102/bills/h5694")+predQuery = Just (UNode "bill:congress")+objQuery  = Just (LNode (PlainL (T.pack "102")))+queryNothing = Nothing++queryBench :: forall rdf. RDF rdf => String -> rdf -> [Benchmark]+queryBench label gr =+   [ bench (label ++ " SPO") $ nf queryGr (subjQuery,predQuery,objQuery,gr)+   , bench (label ++ " SP")  $ nf queryGr (subjQuery,predQuery,queryNothing,gr)+   , bench (label ++ " S")   $ nf queryGr (subjQuery,queryNothing,queryNothing,gr)+   , bench (label ++ " PO")  $ nf queryGr (queryNothing,predQuery,objQuery,gr)+   , bench (label ++ " O")   $ nf queryGr (queryNothing,queryNothing,objQuery,gr)+   ]

@@ -1,7 +1,7 @@ name:            rdf4h-version:         1.3.2+version:         1.3.3 synopsis:        A library for RDF processing in Haskell-description:     +description:   'RDF for Haskell' is a library for working with RDF in Haskell.   At present it includes parsers and serializers for RDF in the N-Triples   and Turtle, and parsing support for RDF/XML. It provides abilities such as querying@@ -49,10 +49,10 @@   else     build-depends: base < 3   build-depends:   parsec >= 3-                 , fgl+                 , fgl >= 5.5.2.0                  , HTTP >= 4000.0.0                  , hxt >= 9.3.1.2-                 , text+                 , text >= 1.2.1.0                  , unordered-containers                  , hashable                  , deepseq@@ -73,20 +73,14 @@   ghc-options:     -Wall -fno-warn-unused-do-bind -funbox-strict-fields -O2  executable rdf4h-  main-is:         Rdf4hParseMain.hs +  main-is:         src/Rdf4hParseMain.hs   if flag(small_base)     build-depends: base >= 3, bytestring   else     build-depends: base >= 4 && < 6-  build-depends:   parsec >= 3-                 , HTTP >= 4000.0.0-                 , hxt >= 9.3.1.2+  build-depends:   rdf4h                  , containers-                 , text-                 , hashable-                 , deepseq-                 , binary-                 , text-binary+                 , text >= 1.2.1.0    if impl(ghc < 7.6)     build-depends: ghc-prim@@ -96,7 +90,6 @@   else     build-depends: network-uri < 2.6, network < 2.6 -  hs-source-dirs:  src   extensions:      BangPatterns RankNTypes ScopedTypeVariables MultiParamTypeClasses OverloadedStrings   ghc-options:     -Wall -fno-warn-unused-do-bind -funbox-strict-fields @@ -106,24 +99,16 @@   ghc-options:   -Wall -fno-warn-unused-do-bind -fno-warn-orphans -fno-warn-name-shadowing -funbox-strict-fields   extensions:    RankNTypes MultiParamTypeClasses Arrows FlexibleContexts OverloadedStrings   build-depends: base >= 4 && < 6-               , parsec >= 3+               , rdf4h                , test-framework >= 0.2.3                , test-framework-quickcheck2                , test-framework-hunit >= 0.2.7-               , HTTP >= 4000.0.0                , QuickCheck >= 1.2.0.0                , HUnit >= 1.2.2.1                , bytestring-               , hxt >= 9.3.1.2                , containers-               , text+               , text >= 1.2.1.0                , knob-               , unordered-containers-               , hashable-               , fgl-               , deepseq-               , binary-               , text-binary    if impl(ghc < 7.6)     build-depends: ghc-prim@@ -133,18 +118,8 @@   else     build-depends: network-uri < 2.6, network < 2.6 -  other-modules: Data.RDF-               , Data.RDF.Namespace-               , Data.RDF.MGraph-               , Data.RDF.PatriciaTreeGraph-               , Data.RDF.TriplesGraph-               , Text.RDF.RDF4H.NTriplesParser-               , Text.RDF.RDF4H.NTriplesSerializer-               , Text.RDF.RDF4H.TurtleParser-               , Text.RDF.RDF4H.TurtleSerializer-               , Text.RDF.RDF4H.XmlParser-               , W3C.TurtleTest-  hs-source-dirs: src, testsuite/tests+  other-modules: W3C.TurtleTest+  hs-source-dirs: testsuite/tests  benchmark rdf4h-bench   type:             exitcode-stdio-1.0@@ -153,7 +128,7 @@   build-depends:    base,                     criterion,                     rdf4h,-                    text+                    text >= 1.2.1.0   ghc-options:      -Wall                     -O2 

@@ -14,7 +14,7 @@   standard_ns_mappings, ns_mappings ) where -import qualified Data.Text as T+import qualified Data.Text.Lazy as T import Data.RDF.Types import qualified Data.Map as Map 

@@ -1,4 +1,4 @@-{-# LANGUAGE GeneralizedNewtypeDeriving, FlexibleInstances, BangPatterns, CPP #-}+{-# LANGUAGE GeneralizedNewtypeDeriving, FlexibleInstances, BangPatterns #-}  module Data.RDF.PatriciaTreeGraph where @@ -16,16 +16,10 @@ import Data.Maybe  newtype PatriciaTreeGraph = PatriciaTreeGraph (PT.Gr Node Node,IntMap.IntMap Node, Maybe BaseUrl, PrefixMappings)-                            deriving (Show,NFData)+                            deriving (Show) -#if MIN_VERSION_fgl(5,5,2)--- fgl 7616e7135c9401e98f4c7350c9f60108ea73e456--- adds those instances but they are not released yet-#else--- fgl-5.5.1.0 and older don't have that instance instance NFData (PT.Gr Node Node)   where rnf x = seq x ()-#endif  instance RDF PatriciaTreeGraph where   baseUrl           = baseUrl'

@@ -20,9 +20,10 @@ import Data.List import Data.RDF.Types import qualified Data.RDF.Namespace as NS (toPMList, uriOf, rdf)-import qualified Data.Text as T+import qualified Data.Text.Lazy as T import Data.Maybe (catMaybes) + -- |Answer the subject node of the triple. {-# INLINE subjectOf #-} subjectOf :: Triple -> Node@@ -101,9 +102,25 @@ -- node names, triple order and prefixes.  In math terms, this is the \simeq -- latex operator, or ~= isIsomorphic :: forall rdf1 rdf2. (RDF rdf1, RDF rdf2) => rdf1 -> rdf2 -> Bool-isIsomorphic g1 g2 = normalize g1 == normalize g2-  where normalize :: forall rdf. (RDF rdf) => rdf -> Triples-        normalize = sort . nub . expandTriples+isIsomorphic g1 g2 = and $ zipWith compareTripleUnlessBlank (normalize g1) (normalize g2)+  where+    compareNodeUnlessBlank :: Node -> Node -> Bool+    compareNodeUnlessBlank (BNode _)     (BNode _)     = True+    compareNodeUnlessBlank (UNode n1)    (UNode n2)    = n1 == n2+    compareNodeUnlessBlank (BNodeGen i1) (BNodeGen i2) = i1 == i2+    compareNodeUnlessBlank (LNode l1)    (LNode l2)    = l1 == l2+    compareNodeUnlessBlank (BNodeGen _)  (BNode _)     = True+    compareNodeUnlessBlank (BNode _)     (BNodeGen _)  = True+    compareNodeUnlessBlank _             _             = False++    compareTripleUnlessBlank :: Triple -> Triple -> Bool+    compareTripleUnlessBlank (Triple s1 p1 o1) (Triple s2 p2 o2) =+        compareNodeUnlessBlank s1 s2 &&+        compareNodeUnlessBlank p1 p2 &&+        compareNodeUnlessBlank o1 o2++    normalize :: forall rdf. (RDF rdf) => rdf -> Triples+    normalize = sort . nub . expandTriples  -- |Expand the triples in a graph with the prefix map and base URL for that -- graph.

@@ -36,7 +36,7 @@ ) where  import Prelude hiding (pred)-import qualified Data.Text as T+import qualified Data.Text.Lazy as T import System.IO import Text.Printf import Data.Binary@@ -45,7 +45,6 @@ import Data.Hashable(Hashable) import qualified Data.List as List import qualified Data.Map as Map-import Data.Text.Binary () import qualified Network.URI as Network (isURI) import Control.DeepSeq (NFData,rnf) import GHC.Generics ()

@@ -12,8 +12,8 @@ import Text.RDF.RDF4H.TurtleParser import Text.RDF.RDF4H.TurtleSerializer -import qualified Data.Text as T-import qualified Data.Text.IO as TIO+import qualified Data.Text.Lazy as T+import qualified Data.Text.Lazy.IO as TIO  import System.Environment import System.IO

@@ -14,7 +14,7 @@  module Text.RDF.RDF4H.Interact where -import qualified Data.Text as T+import qualified Data.Text.Lazy as T  import Data.RDF.Types hiding (baseUrl) import Data.RDF.TriplesGraph()

@@ -11,9 +11,9 @@ import Data.Char(isLetter, isDigit, isLower) import qualified Data.Map as Map import Text.Parsec-import Text.Parsec.Text-import qualified Data.Text as T-import qualified Data.Text.IO as TIO+import Text.Parsec.Text.Lazy+import qualified Data.Text.Lazy as T+import qualified Data.Text.Lazy.IO as TIO import Control.Monad (liftM,void)  -- |NTriplesParser is an 'RdfParser' implementation for parsing RDF in the

@@ -7,8 +7,9 @@  import Control.Monad (void) import Data.RDF.Types-import qualified Data.Text as T-import qualified Data.Text.IO as T+import Data.RDF.Query (expandTriples)+import qualified Data.Text.Lazy as T+import qualified Data.Text.Lazy.IO as T import System.IO  data NTriplesSerializer = NTriplesSerializer@@ -26,7 +27,7 @@   writeN    _     = _writeNode stdout  _writeRdf :: RDF rdf => Handle -> rdf -> IO ()-_writeRdf h = _writeTriples h . triplesOf+_writeRdf h = _writeTriples h . expandTriples  _writeTriples :: Handle -> Triples -> IO () _writeTriples h = mapM_ (_writeTriple h)

@@ -7,9 +7,9 @@ import Network.URI import Network.HTTP import Data.Char (intToDigit)-import Data.Text.Encoding (decodeUtf8)-import qualified Data.ByteString.Char8 as B-import qualified Data.Text as T+import Data.Text.Lazy.Encoding (decodeUtf8)+import qualified Data.ByteString.Lazy.Char8 as B+import qualified Data.Text.Lazy as T -- import qualified Data.Map as Map import Data.Maybe (fromMaybe) 

@@ -12,10 +12,10 @@ import Data.RDF.Namespace import Text.RDF.RDF4H.ParserUtils import Text.Parsec-import Text.Parsec.Text+import Text.Parsec.Text.Lazy import qualified Data.Map as Map-import qualified Data.Text as T-import qualified Data.Text.IO as TIO+import qualified Data.Text.Lazy as T+import qualified Data.Text.Lazy.IO as TIO import Data.Sequence(Seq, (|>)) import qualified Data.Sequence as Seq import qualified Data.Foldable as F

@@ -10,8 +10,8 @@ import Data.RDF.Types import Data.RDF.Query import Data.RDF.Namespace hiding (rdf)-import qualified Data.Text as T-import qualified Data.Text.IO as T+import qualified Data.Text.Lazy as T+import qualified Data.Text.Lazy.IO as T import Data.Map(Map) import qualified Data.Map as Map import Data.List

@@ -13,8 +13,8 @@ import qualified Data.Map as Map (fromList) import Text.RDF.RDF4H.ParserUtils import Data.RDF.Types (RDF,RdfParser(..),Node(BNodeGen),BaseUrl(..),Triple(..),Triples,Subject,Predicate,Object,PrefixMappings(..),ParseFailure(ParseFailure),mkRdf,lnode,plainL,plainLL,typedL,unode,bnode)-import qualified Data.Text as T (Text,pack,unpack)-import qualified Data.Text.IO as TIO+import qualified Data.Text.Lazy as T (Text,pack,unpack)+import qualified Data.Text.Lazy.IO as TIO import Text.XML.HXT.Core (ArrowXml,XmlTree,IfThen((:->)),(>.),(>>.),first,neg,(<+>),expandURI,getName,getAttrValue,getAttrValue0,getAttrl,hasAttrValue,hasAttr,constA,choiceA,getChildren,ifA,arr2A,second,hasName,isElem,isWhiteSpace,xshow,listA,isA,isText,getText,this,unlistA,orElse,sattr,mkelem,xreadDoc,runSLA)  -- TODO: write QuickCheck tests for XmlParser instance for RdfParser.@@ -116,7 +116,7 @@ parseDescription' :: forall a. (ArrowXml a, ArrowState GParseState a) => a (BaseUrl, XmlTree) Triple parseDescription' = proc (bUrl, rdf) -> do                          desc <- isElem <<< getChildren -< rdf-                         state <- arr (\(s, o) -> s { stateSubject = o }) <<< arr fst &&& arr2A mkNode -< (LParseState bUrl Nothing undefined, desc)+                         state <- arr (\(s, o) -> s { stateSubject = o }) <<< arr fst &&& arr2A mkNode <<< updateState -< (LParseState bUrl Nothing undefined, desc)                          triple <- parseDescription -< (state, desc)                          returnA -< triple @@ -154,6 +154,7 @@          <+> isA (== "rdf:ID")          <+> isA (== "xml:lang")          <+> isA (== "rdf:parseType")+         <+> isA (== "xml:base")  -- |Read a children of an rdf:Description element.  These correspond to the Predicate portion of the Triple parsePredicatesFromChildren :: forall a. (ArrowXml a, ArrowState GParseState a)

@@ -3,7 +3,7 @@ import Test.Framework.Providers.API import Test.Framework.Providers.HUnit import qualified Test.HUnit as TU-import qualified Data.Text as T+import qualified Data.Text.Lazy as T  import W3C.Manifest