1 {-# LANGUAGE RankNTypes, TypeFamilies, CPP #-}
3 {-| This module provides @pipes@ utilities for \"text streams\", which are
4 streams of 'Text' chunks. The individual chunks are uniformly @strict@, but
5 a 'Producer' can be converted to and from lazy 'Text's; an 'IO.Handle' can
6 be associated with a 'Producer' or 'Consumer' according as it is read or written to.
8 To stream to or from 'IO.Handle's, one can use 'fromHandle' or 'toHandle'. For
9 example, the following program copies a document from one file to another:
12 > import qualified Data.Text.Pipes as Text
16 > withFile "inFile.txt" ReadMode $ \hIn ->
17 > withFile "outFile.txt" WriteMode $ \hOut ->
18 > runEffect $ Text.fromHandle hIn >-> Text.toHandle hOut
20 To stream from files, the following is perhaps more Prelude-like (note that it uses Pipes.Safe):
23 > import qualified Data.Text.Pipes as Text
26 > main = runSafeT $ runEffect $ Text.readFile "inFile.txt" >-> Text.writeFile "outFile.txt"
28 You can stream to and from 'stdin' and 'stdout' using the predefined 'stdin'
29 and 'stdout' proxies, as with the following \"echo\" program:
31 > main = runEffect $ Text.stdin >-> Text.stdout
33 You can also translate pure lazy 'TL.Text's to and from proxies:
35 > main = runEffect $ Text.fromLazy (TL.pack "Hello, world!\n") >-> Text.stdout
37 In addition, this module provides many functions equivalent to lazy
38 'Text' functions so that you can transform or fold text streams. For
39 example, to stream only the first three lines of 'stdin' to 'stdout' you
43 > import qualified Pipes.Text as Text
44 > import qualified Pipes.Parse as Parse
46 > main = runEffect $ takeLines 3 Text.stdin >-> Text.stdout
48 > takeLines n = Text.unlines . Parse.takeFree n . Text.lines
50 The above program will never bring more than one chunk of text (~ 32 KB) into
51 memory, no matter how long the lines are.
53 Note that functions in this library are designed to operate on streams that
54 are insensitive to text boundaries. This means that they may freely split
55 text into smaller texts and /discard empty texts/. However, they will
56 /never concatenate texts/ in order to provide strict upper bounds on memory
84 #if MIN_VERSION_text(0,11,4)
122 #if MIN_VERSION_text(0,11,4)
134 -- * Character Parsers
148 import Control.Exception (throwIO, try)
149 import Control.Monad (liftM, unless)
150 import Control.Monad.Trans.State.Strict (StateT(..))
151 import qualified Data.Text as T
152 import qualified Data.Text.IO as T
153 import qualified Data.Text.Encoding as TE
154 import qualified Data.Text.Encoding.Error as TE
155 import Data.Text (Text)
156 import qualified Data.Text.Lazy as TL
157 import qualified Data.Text.Lazy.IO as TL
158 import Data.Text.Lazy.Internal (foldrChunks, defaultChunkSize)
159 import Data.ByteString.Unsafe (unsafeTake, unsafeDrop)
160 import Data.ByteString (ByteString)
161 import qualified Data.ByteString as B
162 import Data.Char (ord, isSpace)
163 import Data.Functor.Identity (Identity)
164 import qualified Data.List as List
165 import Foreign.C.Error (Errno(Errno), ePIPE)
166 import qualified GHC.IO.Exception as G
168 import qualified Pipes.ByteString.Parse as PBP
169 import Pipes.Text.Parse (
170 nextChar, drawChar, unDrawChar, peekChar, isEndOfChars )
171 import Pipes.Core (respond, Server')
172 import qualified Pipes.Parse as PP
173 import Pipes.Parse (input, concat, FreeT)
174 import qualified Pipes.Safe.Prelude as Safe
175 import qualified Pipes.Safe as Safe
176 import Pipes.Safe (MonadSafe(..), Base(..))
177 import qualified Pipes.Prelude as P
178 import qualified System.IO as IO
179 import Data.Char (isSpace)
180 import Data.Word (Word8)
181 import Prelude hiding (
210 -- | Convert a lazy 'TL.Text' into a 'Producer' of strict 'Text's
211 fromLazy :: (Monad m) => TL.Text -> Producer' Text m ()
212 fromLazy = foldrChunks (\e a -> yield e >> a) (return ())
213 {-# INLINABLE fromLazy #-}
215 -- | Stream text from 'stdin'
216 stdin :: MonadIO m => Producer' Text m ()
217 stdin = fromHandle IO.stdin
218 {-# INLINABLE stdin #-}
220 {-| Convert a 'IO.Handle' into a text stream using a text size
221 determined by the good sense of the text library.
225 fromHandle :: MonadIO m => IO.Handle -> Producer' Text m ()
226 fromHandle h = go where
227 go = do txt <- liftIO (T.hGetChunk h)
228 unless (T.null txt) $ do yield txt
230 {-# INLINABLE fromHandle#-}
232 {-| Stream text from a file using Pipes.Safe
234 >>> runSafeT $ runEffect $ Text.readFile "hello.hs" >-> Text.map toUpper >-> hoist lift Text.stdout
235 MAIN = PUTSTRLN "HELLO WORLD"
238 readFile :: (MonadSafe m, Base m ~ IO) => FilePath -> Producer' Text m ()
239 readFile file = Safe.withFile file IO.ReadMode fromHandle
240 {-# INLINABLE readFile #-}
242 {-| Stream lines of text from stdin (for testing in ghci etc.)
244 >>> let safely = runSafeT . runEffect
245 >>> safely $ for Text.stdinLn (lift . lift . print . T.length)
252 stdinLn :: MonadIO m => Producer' Text m ()
255 eof <- liftIO (IO.hIsEOF IO.stdin)
257 txt <- liftIO (T.hGetLine IO.stdin)
262 {-| Stream text to 'stdout'
264 Unlike 'toHandle', 'stdout' gracefully terminates on a broken output pipe.
266 Note: For best performance, use @(for source (liftIO . putStr))@ instead of
267 @(source >-> stdout)@ in suitable cases.
269 stdout :: MonadIO m => Consumer' Text m ()
274 x <- liftIO $ try (T.putStr txt)
276 Left (G.IOError { G.ioe_type = G.ResourceVanished
277 , G.ioe_errno = Just ioe })
280 Left e -> liftIO (throwIO e)
282 {-# INLINABLE stdout #-}
284 stdoutLn :: (MonadIO m) => Consumer' Text m ()
289 x <- liftIO $ try (T.putStrLn str)
291 Left (G.IOError { G.ioe_type = G.ResourceVanished
292 , G.ioe_errno = Just ioe })
295 Left e -> liftIO (throwIO e)
297 {-# INLINABLE stdoutLn #-}
299 {-| Convert a text stream into a 'Handle'
301 Note: again, for best performance, where possible use
302 @(for source (liftIO . hPutStr handle))@ instead of @(source >-> toHandle handle)@.
304 toHandle :: MonadIO m => IO.Handle -> Consumer' Text m r
305 toHandle h = for cat (liftIO . T.hPutStr h)
306 {-# INLINABLE toHandle #-}
308 {-# RULES "p >-> toHandle h" forall p h .
309 p >-> toHandle h = for p (\bs -> liftIO (T.hPutStr h bs))
313 -- | Stream text into a file. Uses @pipes-safe@.
314 writeFile :: (MonadSafe m, Base m ~ IO) => FilePath -> Consumer' Text m ()
315 writeFile file = Safe.withFile file IO.WriteMode toHandle
317 -- | Apply a transformation to each 'Char' in the stream
318 map :: (Monad m) => (Char -> Char) -> Pipe Text Text m r
319 map f = P.map (T.map f)
320 {-# INLINABLE map #-}
322 -- | Map a function over the characters of a text stream and concatenate the results
324 :: (Monad m) => (Char -> Text) -> Pipe Text Text m r
325 concatMap f = P.map (T.concatMap f)
326 {-# INLINABLE concatMap #-}
329 -- | Transform a Pipe of 'Text' into a Pipe of 'ByteString's using UTF-8
330 -- encoding; @encodeUtf8 = Pipes.Prelude.map TE.encodeUtf8@ so more complex
331 -- encoding pipes can easily be constructed with the functions in @Data.Text.Encoding@
332 encodeUtf8 :: Monad m => Pipe Text ByteString m r
333 encodeUtf8 = P.map TE.encodeUtf8
334 {-# INLINEABLE encodeUtf8 #-}
336 -- | Transform a Pipe of 'String's into one of 'Text' chunks
337 pack :: Monad m => Pipe String Text m r
339 {-# INLINEABLE pack #-}
341 -- | Transforma a Pipes of 'Text' chunks into one of 'String's
342 unpack :: Monad m => Pipe Text String m r
343 unpack = for cat (\t -> yield (T.unpack t))
344 {-# INLINEABLE unpack #-}
347 -- | @toCaseFold@, @toLower@, @toUpper@ and @stripStart@ are standard 'Text' utility,
348 -- here acting on a 'Text' pipe, rather as they would on a lazy text
349 toCaseFold :: Monad m => Pipe Text Text m ()
350 toCaseFold = P.map T.toCaseFold
351 {-# INLINEABLE toCaseFold #-}
353 -- | lowercase incoming 'Text'
354 toLower :: Monad m => Pipe Text Text m ()
355 toLower = P.map T.toLower
356 {-# INLINEABLE toLower #-}
358 -- | uppercase incoming 'Text'
359 toUpper :: Monad m => Pipe Text Text m ()
360 toUpper = P.map T.toUpper
361 {-# INLINEABLE toUpper #-}
363 -- | Remove leading white space from an incoming succession of 'Text's
364 stripStart :: Monad m => Pipe Text Text m r
367 let text = T.stripStart chunk
371 {-# INLINEABLE stripStart #-}
373 -- | @(take n)@ only allows @n@ individual characters to pass;
374 -- contrast @Pipes.Prelude.take@ which would let @n@ chunks pass.
375 take :: (Monad m, Integral a) => a -> Pipe Text Text m ()
376 take n0 = go n0 where
381 let len = fromIntegral (T.length txt)
383 then yield (T.take (fromIntegral n) txt)
387 {-# INLINABLE take #-}
389 -- | @(drop n)@ drops the first @n@ characters
390 drop :: (Monad m, Integral a) => a -> Pipe Text Text m r
391 drop n0 = go n0 where
396 let len = fromIntegral (T.length txt)
399 yield (T.drop (fromIntegral n) txt)
402 {-# INLINABLE drop #-}
404 -- | Take characters until they fail the predicate
405 takeWhile :: (Monad m) => (Char -> Bool) -> Pipe Text Text m ()
406 takeWhile predicate = go
410 let (prefix, suffix) = T.span predicate txt
416 {-# INLINABLE takeWhile #-}
418 -- | Drop characters until they fail the predicate
419 dropWhile :: (Monad m) => (Char -> Bool) -> Pipe Text Text m r
420 dropWhile predicate = go where
423 case T.findIndex (not . predicate) txt of
428 {-# INLINABLE dropWhile #-}
430 -- | Only allows 'Char's to pass if they satisfy the predicate
431 filter :: (Monad m) => (Char -> Bool) -> Pipe Text Text m r
432 filter predicate = P.map (T.filter predicate)
433 {-# INLINABLE filter #-}
436 -- | Strict left scan over the characters
439 => (Char -> Char -> Char) -> Char -> Pipe Text Text m r
440 scan step begin = go begin
444 let txt' = T.scanl step c txt
448 {-# INLINABLE scan #-}
450 {-| Fold a pure 'Producer' of strict 'Text's into a lazy
453 toLazy :: Producer Text Identity () -> TL.Text
454 toLazy = TL.fromChunks . P.toList
455 {-# INLINABLE toLazy #-}
457 {-| Fold an effectful 'Producer' of strict 'Text's into a lazy
460 Note: 'toLazyM' is not an idiomatic use of @pipes@, but I provide it for
461 simple testing purposes. Idiomatic @pipes@ style consumes the chunks
462 immediately as they are generated instead of loading them all into memory.
464 toLazyM :: (Monad m) => Producer Text m () -> m TL.Text
465 toLazyM = liftM TL.fromChunks . P.toListM
466 {-# INLINABLE toLazyM #-}
468 -- | Reduce the text stream using a strict left fold over characters
471 => (x -> Char -> x) -> x -> (x -> r) -> Producer Text m () -> m r
472 fold step begin done = P.fold (T.foldl' step) begin done
473 {-# INLINABLE fold #-}
475 -- | Retrieve the first 'Char'
476 head :: (Monad m) => Producer Text m () -> m (Maybe Char)
482 Left _ -> return Nothing
483 Right (c, _) -> return (Just c)
484 {-# INLINABLE head #-}
486 -- | Retrieve the last 'Char'
487 last :: (Monad m) => Producer Text m () -> m (Maybe Char)
497 else go (Just $ T.last txt) p'
498 {-# INLINABLE last #-}
500 -- | Determine if the stream is empty
501 null :: (Monad m) => Producer Text m () -> m Bool
503 {-# INLINABLE null #-}
505 -- | Count the number of characters in the stream
506 length :: (Monad m, Num n) => Producer Text m () -> m n
507 length = P.fold (\n txt -> n + fromIntegral (T.length txt)) 0 id
508 {-# INLINABLE length #-}
510 -- | Fold that returns whether 'M.Any' received 'Char's satisfy the predicate
511 any :: (Monad m) => (Char -> Bool) -> Producer Text m () -> m Bool
512 any predicate = P.any (T.any predicate)
513 {-# INLINABLE any #-}
515 -- | Fold that returns whether 'M.All' received 'Char's satisfy the predicate
516 all :: (Monad m) => (Char -> Bool) -> Producer Text m () -> m Bool
517 all predicate = P.all (T.all predicate)
518 {-# INLINABLE all #-}
520 -- | Return the maximum 'Char' within a text stream
521 maximum :: (Monad m) => Producer Text m () -> m (Maybe Char)
522 maximum = P.fold step Nothing id
527 else Just $ case mc of
528 Nothing -> T.maximum txt
529 Just c -> max c (T.maximum txt)
530 {-# INLINABLE maximum #-}
532 -- | Return the minimum 'Char' within a text stream (surely very useful!)
533 minimum :: (Monad m) => Producer Text m () -> m (Maybe Char)
534 minimum = P.fold step Nothing id
540 Nothing -> Just (T.minimum txt)
541 Just c -> Just (min c (T.minimum txt))
542 {-# INLINABLE minimum #-}
544 -- | Find the first element in the stream that matches the predicate
547 => (Char -> Bool) -> Producer Text m () -> m (Maybe Char)
548 find predicate p = head (p >-> filter predicate)
549 {-# INLINABLE find #-}
551 -- | Index into a text stream
553 :: (Monad m, Integral a)
554 => a-> Producer Text m () -> m (Maybe Char)
555 index n p = head (p >-> drop n)
556 {-# INLINABLE index #-}
559 -- | Store a tally of how many segments match the given 'Text'
560 count :: (Monad m, Num n) => Text -> Producer Text m () -> m n
561 count c p = P.fold (+) 0 id (p >-> P.map (fromIntegral . T.count c))
562 {-# INLINABLE count #-}
564 #if MIN_VERSION_text(0,11,4)
565 -- | Transform a Pipe of 'ByteString's expected to be UTF-8 encoded
566 -- into a Pipe of Text
569 => Producer ByteString m r -> Producer Text m (Producer ByteString m r)
570 decodeUtf8 = go TE.streamDecodeUtf8
574 Left r -> return (return r)
575 Right (chunk, p') -> do
576 let TE.Some text l dec' = dec chunk
584 {-# INLINEABLE decodeUtf8 #-}
586 -- | Transform a Pipe of 'ByteString's expected to be UTF-8 encoded
587 -- into a Pipe of Text with a replacement function of type @String -> Maybe Word8 -> Maybe Char@
588 -- E.g. 'Data.Text.Encoding.Error.lenientDecode', which simply replaces bad bytes with \"�\"
592 -> Producer ByteString m r -> Producer Text m (Producer ByteString m r)
593 decodeUtf8With onErr = go (TE.streamDecodeUtf8With onErr)
597 Left r -> return (return r)
598 Right (chunk, p') -> do
599 let TE.Some text l dec' = dec chunk
607 {-# INLINEABLE decodeUtf8With #-}
609 -- | A simple pipe from 'ByteString' to 'Text'; a decoding error will arise
610 -- with any chunk that contains a sequence of bytes that is unreadable. Otherwise
611 -- only few bytes will only be moved from one chunk to the next before decoding.
612 pipeDecodeUtf8 :: Monad m => Pipe ByteString Text m r
613 pipeDecodeUtf8 = go TE.streamDecodeUtf8
614 where go dec = do chunk <- await
616 TE.Some text l dec' -> do yield text
618 {-# INLINEABLE pipeDecodeUtf8 #-}
620 -- | A simple pipe from 'ByteString' to 'Text' using a replacement function.
624 -> Pipe ByteString Text m r
625 pipeDecodeUtf8With onErr = go (TE.streamDecodeUtf8With onErr)
626 where go dec = do chunk <- await
628 TE.Some text l dec' -> do yield text
630 {-# INLINEABLE pipeDecodeUtf8With #-}
633 -- | Splits a 'Producer' after the given number of characters
635 :: (Monad m, Integral n)
638 -> Producer' Text m (Producer Text m r)
645 Left r -> return (return r)
646 Right (txt, p') -> do
647 let len = fromIntegral (T.length txt)
653 let (prefix, suffix) = T.splitAt (fromIntegral n) txt
655 return (yield suffix >> p')
656 {-# INLINABLE splitAt #-}
658 -- | Split a text stream into 'FreeT'-delimited text streams of fixed size
660 :: (Monad m, Integral n)
661 => n -> Producer Text m r -> FreeT (Producer Text m) m r
662 chunksOf n p0 = PP.FreeT (go p0)
668 Right (txt, p') -> PP.Free $ do
669 p'' <- splitAt n (yield txt >> p')
670 return $ PP.FreeT (go p'')
671 {-# INLINABLE chunksOf #-}
673 {-| Split a text stream in two, where the first text stream is the longest
674 consecutive group of text that satisfy the predicate
680 -> Producer' Text m (Producer Text m r)
686 Left r -> return (return r)
687 Right (txt, p') -> do
688 let (prefix, suffix) = T.span predicate txt
695 return (yield suffix >> p')
696 {-# INLINABLE span #-}
698 {-| Split a text stream in two, where the first text stream is the longest
699 consecutive group of characters that don't satisfy the predicate
705 -> Producer Text m (Producer Text m r)
706 break predicate = span (not . predicate)
707 {-# INLINABLE break #-}
709 {-| Split a text stream into sub-streams delimited by characters that satisfy the
716 -> PP.FreeT (Producer Text m) m r
717 splitWith predicate p0 = PP.FreeT (go0 p0)
722 Left r -> return (PP.Pure r)
726 else return $ PP.Free $ do
727 p'' <- span (not . predicate) (yield txt >> p')
728 return $ PP.FreeT (go1 p'')
733 Right (_, p') -> PP.Free $ do
734 p'' <- span (not . predicate) p'
735 return $ PP.FreeT (go1 p'')
736 {-# INLINABLE splitWith #-}
738 -- | Split a text stream using the given 'Char' as the delimiter
742 -> FreeT (Producer Text m) m r
743 split c = splitWith (c ==)
744 {-# INLINABLE split #-}
746 {-| Group a text stream into 'FreeT'-delimited text streams using the supplied
751 => (Char -> Char -> Bool)
753 -> FreeT (Producer Text m) m r
754 groupBy equal p0 = PP.FreeT (go p0)
759 Left r -> return (PP.Pure r)
760 Right (txt, p') -> case (T.uncons txt) of
763 return $ PP.Free $ do
764 p'' <- span (equal c) (yield txt >> p')
765 return $ PP.FreeT (go p'')
766 {-# INLINABLE groupBy #-}
768 -- | Group a text stream into 'FreeT'-delimited text streams of identical characters
770 :: (Monad m) => Producer Text m r -> FreeT (Producer Text m) m r
772 {-# INLINABLE group #-}
774 {-| Split a text stream into 'FreeT'-delimited lines
777 :: (Monad m) => Producer Text m r -> FreeT (Producer Text m) m r
778 lines p0 = PP.FreeT (go0 p0)
783 Left r -> return (PP.Pure r)
787 else return $ PP.Free $ go1 (yield txt >> p')
789 p' <- break ('\n' ==) p
790 return $ PP.FreeT $ do
793 Left r -> return $ PP.Pure r
794 Right (_, p'') -> go0 p''
795 {-# INLINABLE lines #-}
799 -- | Split a text stream into 'FreeT'-delimited words
801 :: (Monad m) => Producer Text m r -> FreeT (Producer Text m) m r
805 x <- next (p >-> dropWhile isSpace)
808 Right (bs, p') -> PP.Free $ do
809 p'' <- break isSpace (yield bs >> p')
811 {-# INLINABLE words #-}
814 -- | Intersperse a 'Char' in between the characters of the text stream
816 :: (Monad m) => Char -> Producer Text m r -> Producer Text m r
823 Right (txt, p') -> do
824 yield (T.intersperse c txt)
830 Right (txt, p') -> do
831 yield (T.singleton c)
832 yield (T.intersperse c txt)
834 {-# INLINABLE intersperse #-}
836 {-| 'intercalate' concatenates the 'FreeT'-delimited text streams after
837 interspersing a text stream in between them
841 => Producer Text m ()
842 -> FreeT (Producer Text m) m r
847 x <- lift (PP.runFreeT f)
849 PP.Pure r -> return r
854 x <- lift (PP.runFreeT f)
856 PP.Pure r -> return r
861 {-# INLINABLE intercalate #-}
863 {-| Join 'FreeT'-delimited lines into a text stream
866 :: (Monad m) => FreeT (Producer Text m) m r -> Producer Text m r
870 x <- lift (PP.runFreeT f)
872 PP.Pure r -> return r
875 yield $ T.singleton '\n'
877 {-# INLINABLE unlines #-}
879 {-| Join 'FreeT'-delimited words into a text stream
882 :: (Monad m) => FreeT (Producer Text m) m r -> Producer Text m r
883 unwords = intercalate (yield $ T.pack " ")
884 {-# INLINABLE unwords #-}
887 The following parsing utilities are single-character analogs of the ones found
892 @Pipes.Text.Parse@ re-exports 'nextChar', 'drawChar', 'unDrawChar', 'peekChar', and 'isEndOfChars'.
894 @Data.Text@ re-exports the 'Text' type.
896 @Pipes.Parse@ re-exports 'input', 'concat', and 'FreeT' (the type).