2e32c31 · streamly · git.iko.soy

git.iko.soy/streamly

Add wordFramedBy parser

commit2e32c313e8df27dd47b071967089e08d8c1f14fa authorHarendra Kumar <harendra@composewell.com> dateMon 9 May 2022, 13:36 (4 years ago) parent28d5a2f

2 files changed+127−3

compare with master

unifiedsplit

core/src/Streamly/Internal/Data/Parser.hs+36−2 core/src/Streamly/Internal/Data/Parser/ParserD.hs+91−1

core/src/Streamly/Internal/Data/Parser.hs+36−2

@@ -128,8 +128,7 @@ module Streamly.Internal.Data.Parser

128128 , takeFramedBy_

129129 , takeFramedByEsc_

130130 , takeFramedByGeneric

131− -- , wordByQuoted

132− -- , wordByQuotedEsc

131+ , wordFramedBy

133132

134133 -- | Unimplemented

135134 --

@@ -252,6 +251,7 @@ import qualified Streamly.Internal.Data.Parser.ParserK.Type as K

252251-- >>> :set -package streamly

253252-- >>> import Prelude hiding (any, all, dropWhile, take, takeWhile, sequence, concatMap, maybe, either, filter)

254253-- >>> import Control.Applicative ((<|>))

254+-- >>> import Data.Char (isSpace)

255255-- >>> import qualified Streamly.Prelude as Stream

256256-- >>> import qualified Streamly.Internal.Data.Stream.IsStream as Stream (parse, parseMany)

257257-- >>> import qualified Streamly.Internal.Data.Fold as Fold

@@ -661,6 +661,7 @@ dropWhile p = takeWhile p FL.drain

661661-- Separators

662662-------------------------------------------------------------------------------

663663

664+-- XXX We can remove Maybe from esc

664665{-# INLINE takeFramedByGeneric #-}

665666takeFramedByGeneric :: MonadCatch m =>

666667 Maybe (a -> Bool)

@@ -823,6 +824,8 @@ takeFramedBy_ isBegin isEnd f = D.toParserK $ D.takeFramedBy_ isBegin isEnd f

823824-- * Stops - when it finds a word separator after a non-word element

824825-- * Fails - never.

825826--

827+-- >>> wordBy = Parser.wordFramedBy (const False) (const False) (const False)

828+--

826829-- @

827830-- S.wordsBy pred f = S.parseMany (PR.wordBy pred f)

828831-- @

@@ -831,6 +834,37 @@ takeFramedBy_ isBegin isEnd f = D.toParserK $ D.takeFramedBy_ isBegin isEnd f

831834wordBy :: MonadCatch m => (a -> Bool) -> Fold m a b -> Parser m a b

832835wordBy f = D.toParserK . D.wordBy f

833836

837+-- | Like 'wordBy' but treats anything inside a pair of quotes as a single

838+-- word, the quotes can be escaped by an escape character. Recursive quotes

839+-- are possible if quote begin and end characters are different, quotes must be

840+-- balanced. Outermost quotes are stripped.

841+--

842+-- >>> braces = Parser.wordFramedBy (== '\\') (== '{') (== '}') isSpace Fold.toList

843+-- >>> Stream.parse braces $ Stream.fromList "{ab} cd"

844+-- "ab"

845+-- >>> Stream.parse braces $ Stream.fromList "{ab}{cd}"

846+-- "abcd"

847+-- >>> Stream.parse braces $ Stream.fromList "a{b} cd"

848+-- "ab"

849+-- >>> Stream.parse braces $ Stream.fromList "a{{b}} cd"

850+-- "a{b}"

851+--

852+-- >>> quotes = Parser.wordFramedBy (== '\\') (== '"') (== '"') isSpace Fold.toList

853+-- >>> Stream.parse quotes $ Stream.fromList "\"a\"\"b\""

854+-- "ab"

855+--

856+-- /Pre-release/

857+{-# INLINE wordFramedBy #-}

858+wordFramedBy :: MonadCatch m =>

859+ (a -> Bool) -- ^ Escape

860+ -> (a -> Bool) -- ^ left quote

861+ -> (a -> Bool) -- ^ right quote

862+ -> (a -> Bool) -- ^ word seperator

863+ -> Fold m a b

864+ -> Parser m a b

865+wordFramedBy isEsc isBegin isEnd isSpc =

866+ D.toParserK . D.wordFramedBy isEsc isBegin isEnd isSpc

867+

834868-- | Given an input stream @[a,b,c,...]@ and a comparison function @cmp@, the

835869-- parser assigns the element @a@ to the first group, then if @a \`cmp` b@ is

836870-- 'True' @b@ is also assigned to the same group. If @a \`cmp` c@ is 'True'

core/src/Streamly/Internal/Data/Parser/ParserD.hs+91−1

@@ -82,7 +82,7 @@ module Streamly.Internal.Data.Parser.ParserD

8282

8383 -- Words and grouping

8484 , wordBy

85− -- wordFramedBy

85+ , wordFramedBy

8686 , groupBy

8787 , groupByRolling

8888 , groupByRollingEither

@@ -1033,6 +1033,96 @@ wordBy predicate (Fold fstep finitial fextract) = Parser step initial extract

10331033 extract (WBWord s) = fextract s

10341034 extract (WBRight b) = return b

10351035

1036+data WordFramedState s b =

1037+ WordFramedSkipPre !s

1038+ | WordFramedWord !s !Int

1039+ | WordFramedEsc !s !Int

1040+ | WordFramedSkipPost !b

1041+

1042+-- | See 'Streamly.Internal.Data.Parser.wordFramedBy'

1043+--

1044+{-# INLINE wordFramedBy #-}

1045+wordFramedBy :: MonadCatch m =>

1046+ (a -> Bool) -- ^ Escape

1047+ -> (a -> Bool) -- ^ left quote

1048+ -> (a -> Bool) -- ^ right quote

1049+ -> (a -> Bool) -- ^ word seperator

1050+ -> Fold m a b

1051+ -> Parser m a b

1052+wordFramedBy isEsc isBegin isEnd isSep

1053+ (Fold fstep finitial fextract) =

1054+ Parser step initial extract

1055+

1056+ where

1057+

1058+ initial = do

1059+ res <- finitial

1060+ return $

1061+ case res of

1062+ FL.Partial s -> IPartial (WordFramedSkipPre s)

1063+ FL.Done _ ->

1064+ error "wordFramedBy: fold done without input"

1065+

1066+ {-# INLINE process #-}

1067+ process s a n = do

1068+ res <- fstep s a

1069+ return

1070+ $ case res of

1071+ FL.Partial s1 -> Continue 0 (WordFramedWord s1 n)

1072+ FL.Done b -> Done 0 b

1073+

1074+ step (WordFramedSkipPre s) a

1075+ | isEsc a = return $ Continue 0 $ WordFramedEsc s 0

1076+ | isSep a = return $ Partial 0 $ WordFramedSkipPre s

1077+ | isBegin a = return $ Continue 0 $ WordFramedWord s 1

1078+ | isEnd a =

1079+ return $ Error "wordFramedBy: missing frame start"

1080+ | otherwise = process s a 0

1081+ step (WordFramedWord s n) a

1082+ | isEsc a = return $ Continue 0 $ WordFramedEsc s n

1083+ | n == 0 && isSep a = do

1084+ b <- fextract s

1085+ return $ Partial 0 $ WordFramedSkipPost b

1086+ | otherwise = do

1087+ -- We need to use different order for checking begin and end for

1088+ -- the n == 0 and n == 1 case so that when the begin and end

1089+ -- character is the same we treat the one after begin as end.

1090+ if n == 0

1091+ then

1092+ -- Need to check isBegin first

1093+ if isBegin a

1094+ then return $ Continue 0 $ WordFramedWord s 1

1095+ else if isEnd a

1096+ then return $ Error "wordFramedBy: missing frame start"

1097+ else process s a n

1098+ else

1099+ -- Need to check isEnd first

1100+ if isEnd a

1101+ then

1102+ if n == 1

1103+ then return $ Continue 0 $ WordFramedWord s 0

1104+ else process s a (n - 1)

1105+ else if isBegin a

1106+ then process s a (n + 1)

1107+ else process s a n

1108+ step (WordFramedEsc s n) a = process s a n

1109+ step (WordFramedSkipPost b) a =

1110+ return

1111+ $ if not (isSep a)

1112+ then Done 1 b

1113+ else Partial 0 $ WordFramedSkipPost b

1114+

1115+ err = throwM . ParseError

1116+

1117+ extract (WordFramedSkipPre s) = fextract s

1118+ extract (WordFramedWord s n) =

1119+ if n == 0

1120+ then fextract s

1121+ else err "wordFramedBy: missing frame end"

1122+ extract (WordFramedEsc _ _) =

1123+ err "wordFramedBy: trailing escape"

1124+ extract (WordFramedSkipPost b) = return b

1125+

10361126{-# ANN type GroupByState Fuse #-}

10371127data GroupByState a s

10381128 = GroupByInit !s