1 {-# OPTIONS_GHC -fno-warn-unused-imports #-}
3 module Pipes.Text.Tutorial (
10 -- ** @Pipes.Text.Encoding@
15 -- ** @view@ \/ @(^.)@
18 -- ** @over@ \/ @(%~)@
24 -- * Special types: @Producer Text m (Producer Text m r)@ and @FreeT (Producer Text m) m r@
31 import Pipes.Text.Encoding
34 This package provides @pipes@ utilities for /character streams/,
35 realized as streams of 'Text' chunks. The individual chunks are uniformly /strict/,
36 and thus the @Text@ type we are using is the one from @Data.Text@, not @Data.Text.Lazy@
37 But the type @Producer Text m r@, as we are using it, is a sort of /pipes/ equivalent of
40 The main @Pipes.Text@ module provides many functions equivalent
41 in one way or another to the pure functions in
42 <https://hackage.haskell.org/package/text-1.1.0.0/docs/Data-Text-Lazy.html Data.Text.Lazy>
43 (and the corresponding @Prelude@ functions for @String@ s): they transform,
44 divide, group and fold text streams. Though @Producer Text m r@
45 is the type of \'effectful Text\', the functions in @Pipes.Text@ are \'pure\'
46 in the sense that they are uniformly monad-independent.
47 Simple /IO/ operations are defined in @Pipes.Text.IO@ - as lazy IO @Text@
48 operations are in @Data.Text.Lazy.IO@. Similarly, as @Data.Text.Lazy.Encoding@
49 handles inter-operation with @Data.ByteString.Lazy@, @Pipes.Text.Encoding@ provides for
50 interoperation with the \'effectful ByteStrings\' of @Pipes.ByteString@.
52 Remember that the @Text@ type exported by @Data.Text.Lazy@ is basically
53 that of a lazy list of strict @Text@: the implementation is arranged so that
54 the individual strict 'Text' chunks are kept to a reasonable size; the user
55 is not aware of the divisions between the connected 'Text' chunks, but uses
56 operations akin to those for strict text.
57 So also here: the functions in this module are designed to operate on character streams that
58 in a way that is independent of the boundaries of the underlying @Text@ chunks.
59 This means that they may freely split text into smaller texts and /discard empty texts/.
60 The objective, though, is that they should not /concatenate texts/ in order to provide strict upper
61 bounds on memory usage.
63 For example, to stream only the first three lines of 'stdin' to 'stdout' you
67 > import qualified Pipes.Text as Text
68 > import qualified Pipes.Text.IO as Text
69 > import Pipes.Group (takes')
70 > import Lens.Family (view)
72 > main = runEffect $ takeLines 3 Text.stdin >-> Text.stdout
74 > takeLines n = view Text.unlines . takes' n . view Text.lines
76 This program will never bring more into memory than what @Text.stdin@ considers
77 one chunk of text (~ 32 KB), even if individual lines are split across many chunks.
81 As the use of @view@ in this example shows, one superficial difference from @Data.Text.Lazy@
82 is that many of the operations, like 'lines', are \'lensified\'; this has a
83 number of advantages; in particular it facilitates their use with 'Parser's of Text
84 (in the general <http://hackage.haskell.org/package/pipes-parse-3.0.1/docs/Pipes-Parse-Tutorial.html pipes-parse>
85 sense.) The remarks that follow in this section are for non-lens adepts.
87 Each lens exported here, e.g. 'lines', 'chunksOf' or 'splitAt', reduces to the
88 intuitively corresponding function when used with @view@ or @(^.)@. Instead of
93 as we would with the Prelude or Text functions, we write
95 > view (splitAt 17) producer
99 > producer ^. splitAt 17
101 This may seem a little indirect, but note that many equivalents of
102 @Text -> Text@ functions are exported here as 'Pipe's. Here too we recover the intuitively
103 corresponding functions by prefixing them with @(>->)@. Thus something like
105 > stripLines = view Text.unlines . Group.maps (>-> Text.stripStart) . view Text.lines
107 would drop the leading white space from each line.
109 The lenses in this library are marked as /improper/; this just means that
110 they don't admit all the operations of an ideal lens, but only /getting/ and /focusing/.
111 Just for this reason, though, the magnificent complexities of the lens libraries
112 are a distraction. The lens combinators to keep in mind, the ones that make sense for
113 our lenses, are @view@ \/ @(^.)@), @over@ \/ @(%~)@ , and @zoom@.
115 One need only keep in mind that if @l@ is a @Lens' a b@, then:
119 @view l@ is a function @a -> b@ . Thus @view l a@ (also written @a ^. l@ )
120 is the corresponding @b@; as was said above, this function will typically be
121 the pipes equivalent of the function you think it is, given its name. So for example
124 > view (Text.splitAt 300) :: Producer Text m r -> Producer Text (Producer Text m r)
125 > Text.stdin ^. splitAt 300 :: Producer Text IO (Producer Text IO r)
127 I.e., it produces the first 300 characters, and returns the rest of the producer.
128 Thus to uppercase the first n characters
129 of a Producer, leaving the rest the same, we could write:
132 > upper n p = do p' <- p ^. Text.splitAt n >-> Text.toUpper
136 @over l@ is a function @(b -> b) -> a -> a@. Thus, given a function that modifies
137 @b@s, the lens lets us modify an @a@ by applying @f :: b -> b@ to
138 the @b@ that we can \"see\" through the lens. So @over l f :: a -> a@
139 (it can also be written @l %~ f@).
140 For any particular @a@, then, @over l f a@ or @(l %~ f) a@ is a revised @a@.
141 So above we might have written things like these:
143 > stripLines = Text.lines %~ maps (>-> Text.stripStart)
144 > stripLines = over Text.lines (maps (>-> Text.stripStart))
145 > upper n = Text.splitAt n %~ (>-> Text.toUpper)
149 @zoom l@, finally, is a function from a @Parser b m r@
150 to a @Parser a m r@ (or more generally a @StateT (Producer b m x) m r@).
151 Its use is easiest to see with an decoding lens like 'utf8', which
152 \"sees\" a Text producer hidden inside a ByteString producer:
153 @drawChar@ is a Text parser, returning a @Maybe Char@, @zoom utf8 drawChar@ is
154 a /ByteString/ parser, returning a @Maybe Char@. @drawAll@ is a Parser that returns
155 a list of everything produced from a Producer, leaving only the return value; it would
156 usually be unreasonable to use it. But @zoom (splitAt 17) drawAll@
157 returns a list of Text chunks containing the first seventeen Chars, and returns the rest of
158 the Text Producer for further parsing. Suppose that we want, inexplicably, to
159 modify the casing of a Text Producer according to any instruction it might
160 contain at the start. Then we might write something like this:
162 > obey :: Monad m => Producer Text m b -> Producer Text m b
163 > obey p = do (ts, p') <- lift $ runStateT (zoom (Text.splitAt 7) drawAll) p
164 > let seven = T.concat ts
165 > case T.toUpper seven of
166 > "TOUPPER" -> p' >-> Text.toUpper
167 > "TOLOWER" -> p' >-> Text.toLower
168 > _ -> do yield seven
172 > >>> let doc = each ["toU","pperTh","is document.\n"]
173 > >>> runEffect $ obey doc >-> Text.stdout
176 The purpose of exporting lenses is the mental economy achieved with this three-way
177 applicability. That one expression, e.g. @lines@ or @splitAt 17@ can have these
178 three uses is no more surprising than that a pipe can act as a function modifying
179 the output of a producer, namely by using @>->@ to its left: @producer >-> pipe@
180 -- but can /also/ modify the inputs to a consumer by using @>->@ to its right:
183 The three functions, @view@ \/ @(^.)@, @over@ \/ @(%~)@ and @zoom@ are supplied by
184 both <http://hackage.haskell.org/package/lens lens> and
185 <http://hackage.haskell.org/package/lens-family lens-family> The use of 'zoom' is explained
186 in <http://hackage.haskell.org/package/pipes-parse-3.0.1/docs/Pipes-Parse-Tutorial.html Pipes.Parse.Tutorial>
187 and to some extent in the @Pipes.Text.Encoding@ module here.
191 These simple 'lines' examples reveal a more important difference from @Data.Text.Lazy@ .
192 This is in the types that are most closely associated with our central text type,
193 @Producer Text m r@. In @Data.Text@ and @Data.Text.Lazy@ we find functions like
195 > splitAt :: Int -> Text -> (Text, Text)
196 > lines :: Text -> [Text]
197 > chunksOf :: Int -> Text -> [Text]
199 which relate a Text with a pair of Texts or a list of Texts.
200 The corresponding functions here (taking account of \'lensification\') are
202 > view . splitAt :: (Monad m, Integral n) => n -> Producer Text m r -> Producer Text m (Producer Text m r)
203 > view lines :: Monad m => Producer Text m r -> FreeT (Producer Text m) m r
204 > view . chunksOf :: (Monad m, Integral n) => n -> Producer Text m r -> FreeT (Producer Text m) m r
206 Some of the types may be more readable if you imagine that we have introduced
207 our own type synonyms
209 > type Text m r = Producer T.Text m r
210 > type Texts m r = FreeT (Producer T.Text m) m r
212 Then we would think of the types above as
214 > view . splitAt :: (Monad m, Integral n) => n -> Text m r -> Text m (Text m r)
215 > view lines :: (Monad m) => Text m r -> Texts m r
216 > view . chunksOf :: (Monad m, Integral n) => n -> Text m r -> Texts m r
218 which brings one closer to the types of the similar functions in @Data.Text.Lazy@
220 In the type @Producer Text m (Producer Text m r)@ the second
221 element of the \'pair\' of effectful Texts cannot simply be retrieved
222 with something like 'snd'. This is an \'effectful\' pair, and one must work
223 through the effects of the first element to arrive at the second Text stream, even
224 if you are proposing to throw the Text in the first element away.
225 Note that we use Control.Monad.join to fuse the pair back together, since it specializes to
227 > join :: Monad m => Producer Text m (Producer m r) -> Producer m r
229 The return type of 'lines', 'words', 'chunksOf' and the other /splitter/ functions,
230 @FreeT (Producer m Text) m r@ -- our @Texts m r@ -- is the type of (effectful)
231 lists of (effectful) texts. The type @([Text],r)@ might be seen to gather
232 together things of the forms:
237 > (Text, (Text, (Text, r)))
238 > (Text, (Text, (Text, (Text, r))))
241 (We might also have identified the sum of those types with @Free ((,) Text) r@
242 -- or, more absurdly, @FreeT ((,) Text) Identity r@.)
244 Similarly, our type @Texts m r@, or @FreeT (Text m) m r@ -- in fact called
245 @FreeT (Producer Text m) m r@ here -- encompasses all the members of the sequence:
250 > Text m (Text m (Text m r))
251 > Text m (Text m (Text m (Text m r)))
254 We might have used a more specialized type in place of @FreeT (Producer a m) m r@,
255 or indeed of @FreeT (Producer Text m) m r@, but it is clear that the correct
256 result type of 'lines' will be isomorphic to @FreeT (Producer Text m) m r@ .
260 > lines :: Monad m => Lens' (Producer Text m r) (FreeT (Producer Text m) m r)
261 > view . lines :: Monad m => Producer Text m r -> FreeT (Producer Text m) m r
263 should really have the type
265 > lines :: Monad m => Pipe Text Text m r
267 as e.g. 'toUpper' does. But this would spoil the control we are
268 attempting to maintain over the size of chunks. It is in fact just
269 as unreasonable to want such a pipe as to want
271 > Data.Text.Lazy.lines :: Text -> Text
273 to 'rechunk' the strict Text chunks inside the lazy Text to respect
274 line boundaries. In fact we have
276 > Data.Text.Lazy.lines :: Text -> [Text]
277 > Prelude.lines :: String -> [String]
279 where the elements of the list are themselves lazy Texts or Strings; the use
280 of @FreeT (Producer Text m) m r@ is simply the 'effectful' version of this.
282 The @Pipes.Group@ module, which can generally be imported without qualification,
283 provides many functions for working with things of type @FreeT (Producer a m) m r@.
284 In particular it conveniently exports the constructors for @FreeT@ and the associated
285 @FreeF@ type -- a fancy form of @Either@, namely
287 > data FreeF f a b = Pure a | Free (f b)
289 for pattern-matching. Consider the implementation of the 'words' function, or
290 of the part of the lens that takes us to the words; it is compact but exhibits many
291 of the points under discussion, including explicit handling of the @FreeT@ and @FreeF@
292 constuctors. Keep in mind that
294 > newtype FreeT f m a = FreeT (m (FreeF f a (FreeT f m a)))
295 > next :: Monad m => Producer a m r -> m (Either r (a, Producer a m r))
297 Thus the @do@ block after the @FreeT@ constructor is in the base monad, e.g. 'IO' or 'Identity';
298 the later subordinate block, opened by the @Free@ constructor, is in the @Producer@ monad:
300 > words :: Monad m => Producer Text m r -> FreeT (Producer Text m) m r
301 > words p = FreeT $ do -- With 'next' we will inspect p's first chunk, excluding spaces;
302 > x <- next (p >-> dropWhile isSpace) -- note that 'dropWhile isSpace' is a pipe, and is thus *applied* with '>->'.
303 > return $ case x of -- We use 'return' and so need something of type 'FreeF (Text m) r (Texts m r)'
304 > Left r -> Pure r -- 'Left' means we got no Text chunk, but only the return value; so we are done.
305 > Right (txt, p') -> Free $ do -- If we get a chunk and the rest of the producer, p', we enter the 'Producer' monad
306 > p'' <- view (break isSpace) -- When we apply 'break isSpace', we get a Producer that returns a Producer;
307 > (yield txt >> p') -- so here we yield everything up to the next space, and get the rest back.
308 > return (words p'') -- We then carry on with the rest, which is likely to begin with space.