aboutsummaryrefslogtreecommitdiff
path: root/haddock-library/src/Documentation/Haddock/Parser.hs
blob: 94750ad4ac9f81dabac9285f57dcd45313c976a8 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
{-# LANGUAGE OverloadedStrings #-}
-- |
-- Module      :  Documentation.Haddock.Parser
-- Copyright   :  (c) Mateusz Kowalczyk 2013-2014,
--                    Simon Hengel      2013
-- License     :  BSD-like
--
-- Maintainer  :  haddock@projects.haskell.org
-- Stability   :  experimental
-- Portability :  portable
--
-- Parser used for Haddock comments. For external users of this
-- library, the most commonly used combination of functions is going
-- to be
--
-- @'toRegular' . 'parseParas'@
module Documentation.Haddock.Parser ( parseString, parseParas
                                    , overIdentifier, toRegular, Identifier
                                    ) where

import           Control.Applicative
import           Control.Arrow (first)
import           Control.Monad (void, mfilter)
import           Data.Attoparsec.ByteString.Char8 hiding (parse, take, endOfLine)
import qualified Data.ByteString.Char8 as BS
import           Data.Char (chr, isAsciiUpper)
import           Data.List (stripPrefix, intercalate, unfoldr)
import           Data.Maybe (fromMaybe)
import           Data.Monoid
import           Documentation.Haddock.Doc
import           Documentation.Haddock.Parser.Util
import           Documentation.Haddock.Types
import           Documentation.Haddock.Utf8
import           Prelude hiding (takeWhile)

-- $setup
-- >>> :set -XOverloadedStrings

-- | Identifier string surrounded with opening and closing quotes/backticks.
type Identifier = (Char, String, Char)

-- | Drops the quotes/backticks around all identifiers, as if they
-- were valid but still 'String's.
toRegular :: DocH mod Identifier -> DocH mod String
toRegular = fmap (\(_, x, _) -> x)

-- | Maps over 'DocIdentifier's over 'String' with potentially failing
-- conversion using user-supplied function. If the conversion fails,
-- the identifier is deemed to not be valid and is treated as a
-- regular string.
overIdentifier :: (String -> Maybe a)
               -> DocH mod Identifier
               -> DocH mod a
overIdentifier f d = g d
  where
    g (DocIdentifier (o, x, e)) = case f x of
      Nothing -> DocString $ o : x ++ [e]
      Just x' -> DocIdentifier x'
    g DocEmpty = DocEmpty
    g (DocAppend x x') = DocAppend (g x) (g x')
    g (DocString x) = DocString x
    g (DocParagraph x) = DocParagraph $ g x
    g (DocIdentifierUnchecked x) = DocIdentifierUnchecked x
    g (DocModule x) = DocModule x
    g (DocWarning x) = DocWarning $ g x
    g (DocEmphasis x) = DocEmphasis $ g x
    g (DocMonospaced x) = DocMonospaced $ g x
    g (DocBold x) = DocBold $ g x
    g (DocUnorderedList x) = DocUnorderedList $ fmap g x
    g (DocOrderedList x) = DocOrderedList $ fmap g x
    g (DocDefList x) = DocDefList $ fmap (\(y, z) -> (g y, g z)) x
    g (DocCodeBlock x) = DocCodeBlock $ g x
    g (DocHyperlink x) = DocHyperlink x
    g (DocPic x) = DocPic x
    g (DocAName x) = DocAName x
    g (DocProperty x) = DocProperty x
    g (DocExamples x) = DocExamples x
    g (DocHeader (Header l x)) = DocHeader . Header l $ g x

parse :: Parser a -> BS.ByteString -> a
parse p = either err id . parseOnly (p <* endOfInput)
  where
    err = error . ("Haddock.Parser.parse: " ++)

-- | Main entry point to the parser. Appends the newline character
-- to the input string.
parseParas :: String -- ^ String to parse
           -> DocH mod Identifier
parseParas = parse (p <* skipSpace) . encodeUtf8 . (++ "\n")
  where
    p :: Parser (DocH mod Identifier)
    p = docConcat <$> paragraph `sepBy` many (skipHorizontalSpace *> "\n")

-- | Parse a text paragraph. Actually just a wrapper over 'parseStringBS' which
-- drops leading whitespace and encodes the string to UTF8 first.
parseString :: String -> DocH mod Identifier
parseString = parseStringBS . encodeUtf8 . dropWhile isSpace

parseStringBS :: BS.ByteString -> DocH mod Identifier
parseStringBS = parse p
  where
    p :: Parser (DocH mod Identifier)
    p = docConcat <$> many (monospace <|> anchor <|> identifier <|> moduleName
                            <|> picture <|> hyperlink <|> bold
                            <|> emphasis <|> encodedChar <|> string'
                            <|> skipSpecialChar)

-- | Parses and processes
-- <https://en.wikipedia.org/wiki/Numeric_character_reference Numeric character references>
--
-- >>> parseOnly encodedChar "&#65;"
-- Right (DocString "A")
encodedChar :: Parser (DocH mod a)
encodedChar = "&#" *> c <* ";"
  where
    c = DocString . return . chr <$> num
    num = hex <|> decimal
    hex = ("x" <|> "X") *> hexadecimal

-- | List of characters that we use to delimit any special markup.
-- Once we have checked for any of these and tried to parse the
-- relevant markup, we can assume they are used as regular text.
specialChar :: [Char]
specialChar = "_/<@\"&'`# "

-- | Plain, regular parser for text. Called as one of the last parsers
-- to ensure that we have already given a chance to more meaningful parsers
-- before capturing their characers.
string' :: Parser (DocH mod a)
string' = DocString . unescape . decodeUtf8 <$> takeWhile1_ (`notElem` specialChar)
  where
    unescape "" = ""
    unescape ('\\':x:xs) = x : unescape xs
    unescape (x:xs) = x : unescape xs

-- | Skips a single special character and treats it as a plain string.
-- This is done to skip over any special characters belonging to other
-- elements but which were not deemed meaningful at their positions.
skipSpecialChar :: Parser (DocH mod a)
skipSpecialChar = DocString . return <$> satisfy (`elem` specialChar)

-- | Emphasis parser.
--
-- >>> parseOnly emphasis "/Hello world/"
-- Right (DocEmphasis (DocString "Hello world"))
emphasis :: Parser (DocH mod Identifier)
emphasis = DocEmphasis . parseStringBS <$>
  mfilter ('\n' `BS.notElem`) ("/" *> takeWhile1_ (/= '/') <* "/")

-- | Bold parser.
--
-- >>> parseOnly bold "__Hello world__"
-- Right (DocBold (DocString "Hello world"))
bold :: Parser (DocH mod Identifier)
bold = DocBold . parseStringBS <$> disallowNewline ("__" *> takeUntil "__")

disallowNewline :: Parser BS.ByteString -> Parser BS.ByteString
disallowNewline = mfilter ('\n' `BS.notElem`)

-- | Like `takeWhile`, but unconditionally take escaped characters.
takeWhile_ :: (Char -> Bool) -> Parser BS.ByteString
takeWhile_ p = scan False p_
  where
    p_ escaped c
      | escaped = Just False
      | not $ p c = Nothing
      | otherwise = Just (c == '\\')

-- | Like `takeWhile1`, but unconditionally take escaped characters.
takeWhile1_ :: (Char -> Bool) -> Parser BS.ByteString
takeWhile1_ = mfilter (not . BS.null) . takeWhile_

-- | Text anchors to allow for jumping around the generated documentation.
--
-- >>> parseOnly anchor "#Hello world#"
-- Right (DocAName "Hello world")
anchor :: Parser (DocH mod a)
anchor = DocAName . decodeUtf8 <$>
         disallowNewline ("#" *> takeWhile1_ (/= '#') <* "#")

-- | Monospaced strings.
--
-- >>> parseOnly monospace "@cruel@"
-- Right (DocMonospaced (DocString "cruel"))
monospace :: Parser (DocH mod Identifier)
monospace = DocMonospaced . parseStringBS <$> ("@" *> takeWhile1_ (/= '@') <* "@")

moduleName :: Parser (DocH mod a)
moduleName = DocModule <$> (char '"' *> modid <* char '"')
  where
    modid = intercalate "." <$> conid `sepBy1` "."
    conid = (:)
      <$> satisfy isAsciiUpper
      -- NOTE: According to Haskell 2010 we should actually only
      -- accept {small | large | digit | ' } here.  But as we can't
      -- match on unicode characters, this is currently not possible.
      -- Note that we allow ‘#’ to suport anchors.
      <*> (decodeUtf8 <$> takeWhile (`notElem` " .&[{}(=*)+]!|@/;,^?\"\n"))

-- | Picture parser, surrounded by \<\< and \>\>. It's possible to specify
-- a title for the picture.
--
-- >>> parseOnly picture "<<hello.png>>"
-- Right (DocPic (Picture {pictureUri = "hello.png", pictureTitle = Nothing}))
-- >>> parseOnly picture "<<hello.png world>>"
-- Right (DocPic (Picture {pictureUri = "hello.png", pictureTitle = Just "world"}))
picture :: Parser (DocH mod a)
picture = DocPic . makeLabeled Picture . decodeUtf8
          <$> disallowNewline ("<<" *> takeUntil ">>")

-- | Paragraph parser, called by 'parseParas'.
paragraph :: Parser (DocH mod Identifier)
paragraph = examples <|> skipSpace *> (
      unorderedList
  <|> orderedList
  <|> definitionList
  <|> birdtracks
  <|> codeblock
  <|> property
  <|> header
  <|> textParagraph
  )

-- | Headers inside the comment denoted with @=@ signs, up to 6 levels
-- deep.
--
-- >>> parseOnly header "= Hello"
-- Right (DocHeader (Header {headerLevel = 1, headerTitle = DocString "Hello"}))
-- >>> parseOnly header "== World"
-- Right (DocHeader (Header {headerLevel = 2, headerTitle = DocString "World"}))
header :: Parser (DocH mod Identifier)
header = do
  let psers = map (string . encodeUtf8 . concat . flip replicate "=") [6, 5 .. 1]
      pser = foldl1 (<|>) psers
  delim <- decodeUtf8 <$> pser
  line <- skipHorizontalSpace *> nonEmptyLine >>= return . parseString
  rest <- paragraph <|> return DocEmpty
  return $ DocHeader (Header (length delim) line) `docAppend` rest

textParagraph :: Parser (DocH mod Identifier)
textParagraph = docParagraph . parseString . intercalate "\n" <$> many1 nonEmptyLine

-- | Parses unordered (bullet) lists.
unorderedList :: Parser (DocH mod Identifier)
unorderedList = DocUnorderedList <$> p
  where
    p = ("*" <|> "-") *> innerList p

-- | Parses ordered lists (numbered or dashed).
orderedList :: Parser (DocH mod Identifier)
orderedList = DocOrderedList <$> p
  where
    p = (paren <|> dot) *> innerList p
    dot = (decimal :: Parser Int) <* "."
    paren = "(" *> decimal <* ")"

-- | Generic function collecting any further lines belonging to the
-- list entry and recursively collecting any further lists in the
-- same paragraph. Usually used as
--
-- > someListFunction = listBeginning *> innerList someListFunction
innerList :: Parser [DocH mod Identifier] -> Parser [DocH mod Identifier]
innerList item = do
  c <- takeLine
  (cs, items) <- more item
  let contents = docParagraph . parseString . dropNLs . unlines $ c : cs
  return $ case items of
    Left p -> [contents `docAppend` p]
    Right i -> contents : i

-- | Parses definition lists.
definitionList :: Parser (DocH mod Identifier)
definitionList = DocDefList <$> p
  where
    p = do
      label <- "[" *> (parseStringBS <$> takeWhile1 (`notElem` "]\n")) <* "]"
      c <- takeLine
      (cs, items) <- more p
      let contents = parseString . dropNLs . unlines $ c : cs
      return $ case items of
        Left x -> [(label, contents `docAppend` x)]
        Right i -> (label, contents) : i

-- | Drops all trailing newlines.
dropNLs :: String -> String
dropNLs = reverse . dropWhile (== '\n') . reverse

-- | Main worker for 'innerList' and 'definitionList'.
-- We need the 'Either' here to be able to tell in the respective functions
-- whether we're dealing with the next list or a nested paragraph.
more :: Monoid a => Parser a
     -> Parser ([String], Either (DocH mod Identifier) a)
more item = innerParagraphs <|> moreListItems item
            <|> moreContent item <|> pure ([], Right mempty)

-- | Used by 'innerList' and 'definitionList' to parse any nested paragraphs.
innerParagraphs :: Parser ([String], Either (DocH mod Identifier) a)
innerParagraphs = (,) [] . Left <$> ("\n" *> indentedParagraphs)

-- | Attempts to fetch the next list if possibly. Used by 'innerList' and
-- 'definitionList' to recursively grab lists that aren't separated by a whole
-- paragraph.
moreListItems :: Parser a
              -> Parser ([String], Either (DocH mod Identifier) a)
moreListItems item = (,) [] . Right <$> (skipSpace *> item)

-- | Helper for 'innerList' and 'definitionList' which simply takes
-- a line of text and attempts to parse more list content with 'more'.
moreContent :: Monoid a => Parser a
            -> Parser ([String], Either (DocH mod Identifier) a)
moreContent item = first . (:) <$> nonEmptyLine <*> more item

-- | Runs the 'parseParas' parser on an indented paragraph.
-- The indentation is 4 spaces.
indentedParagraphs :: Parser (DocH mod Identifier)
indentedParagraphs = parseParas . concat <$> dropFrontOfPara "    "

-- | Grab as many fully indented paragraphs as we can.
dropFrontOfPara :: Parser BS.ByteString -> Parser [String]
dropFrontOfPara sp = do
  currentParagraph <- some (sp *> takeNonEmptyLine)
  followingParagraphs <-
    skipHorizontalSpace *> nextPar -- we have more paragraphs to take
    <|> skipHorizontalSpace *> nlList -- end of the ride, remember the newline
    <|> endOfInput *> return [] -- nothing more to take at all
  return (currentParagraph ++ followingParagraphs)
  where
    nextPar = (++) <$> nlList <*> dropFrontOfPara sp
    nlList = "\n" *> return ["\n"]

nonSpace :: BS.ByteString -> Parser BS.ByteString
nonSpace xs
  | not $ any (not . isSpace) $ decodeUtf8 xs = fail "empty line"
  | otherwise = return xs

-- | Takes a non-empty, not fully whitespace line.
--
--  Doesn't discard the trailing newline.
takeNonEmptyLine :: Parser String
takeNonEmptyLine = do
    (++ "\n") . decodeUtf8 <$> (takeWhile1 (/= '\n') >>= nonSpace) <* "\n"

-- | Blocks of text of the form:
--
-- >> foo
-- >> bar
-- >> baz
--
birdtracks :: Parser (DocH mod a)
birdtracks = DocCodeBlock . DocString . intercalate "\n" . stripSpace <$> many1 line
  where
    line = skipHorizontalSpace *> ">" *> takeLine

stripSpace :: [String] -> [String]
stripSpace = fromMaybe <*> mapM strip'
  where
    strip' (' ':xs') = Just xs'
    strip' "" = Just ""
    strip' _  = Nothing

-- | Parses examples. Examples are a paragraph level entitity (separated by an empty line).
-- Consecutive examples are accepted.
examples :: Parser (DocH mod a)
examples = DocExamples <$> (many (skipHorizontalSpace *> "\n") *> go)
  where
    go :: Parser [Example]
    go = do
      prefix <- decodeUtf8 <$> takeHorizontalSpace <* ">>>"
      expr <- takeLine
      (rs, es) <- resultAndMoreExamples
      return (makeExample prefix expr rs : es)
      where
        resultAndMoreExamples :: Parser ([String], [Example])
        resultAndMoreExamples = moreExamples <|> result <|> pure ([], [])
          where
            moreExamples :: Parser ([String], [Example])
            moreExamples = (,) [] <$> go

            result :: Parser ([String], [Example])
            result = first . (:) <$> nonEmptyLine <*> resultAndMoreExamples

    makeExample :: String -> String -> [String] -> Example
    makeExample prefix expression res =
      Example (strip expression) result
      where
        result = map (substituteBlankLine . tryStripPrefix) res

        tryStripPrefix xs = fromMaybe xs (stripPrefix prefix xs)

        substituteBlankLine "<BLANKLINE>" = ""
        substituteBlankLine xs = xs

nonEmptyLine :: Parser String
nonEmptyLine = mfilter (any (not . isSpace)) takeLine

takeLine :: Parser String
takeLine = decodeUtf8 <$> takeWhile (/= '\n') <* endOfLine

endOfLine :: Parser ()
endOfLine = void "\n" <|> endOfInput

-- | Property parser.
--
-- >>> parseOnly property "prop> hello world"
-- Right (DocProperty "hello world")
property :: Parser (DocH mod a)
property = DocProperty . strip . decodeUtf8 <$> ("prop>" *> takeWhile1 (/= '\n'))

-- |
-- Paragraph level codeblock. Anything between the two delimiting \@ is parsed
-- for markup.
codeblock :: Parser (DocH mod Identifier)
codeblock =
  DocCodeBlock . parseStringBS . dropSpaces
  <$> ("@" *> skipHorizontalSpace *> "\n" *> block' <* "@")
  where
    dropSpaces xs =
      let rs = decodeUtf8 xs
      in case splitByNl rs of
        [] -> xs
        ys -> case last ys of
          ' ':_ -> case mapM dropSpace ys of
            Nothing -> xs
            Just zs -> encodeUtf8 $ intercalate "\n" zs
          _ -> xs

    -- This is necessary because ‘lines’ swallows up a trailing newline
    -- and we lose information about whether the last line belongs to @ or to
    -- text which we need to decide whether we actually want to be dropping
    -- anything at all.
    splitByNl = unfoldr (\x -> case x of
                                 '\n':s -> Just (span (/= '\n') s)
                                 _      -> Nothing)
                . ('\n' :)

    dropSpace "" = Just ""
    dropSpace (' ':xs) = Just xs
    dropSpace _ = Nothing

    block' = scan False p
      where
        p isNewline c
          | isNewline && c == '@' = Nothing
          | isNewline && isSpace c = Just isNewline
          | otherwise = Just $ c == '\n'

-- | Parses links that were specifically marked as such.
hyperlink :: Parser (DocH mod a)
hyperlink = DocHyperlink . makeLabeled Hyperlink . decodeUtf8
              <$> disallowNewline ("<" *> takeUntil ">")
            <|> autoUrl

-- | Looks for URL-like things to automatically hyperlink even if they
-- weren't marked as links.
autoUrl :: Parser (DocH mod a)
autoUrl = mkLink <$> url
  where
    url = mappend <$> ("http://" <|> "https://" <|> "ftp://") <*> takeWhile1 (not . isSpace)
    mkLink :: BS.ByteString -> DocH mod a
    mkLink s = case unsnoc s of
      Just (xs, x) | x `elem` ",.!?" -> DocHyperlink (Hyperlink (decodeUtf8 xs) Nothing) `docAppend` DocString [x]
      _ -> DocHyperlink (Hyperlink (decodeUtf8 s) Nothing)

-- | Parses strings between identifier delimiters. Consumes all input that it
-- deems to be valid in an identifier. Note that it simply blindly consumes
-- characters and does no actual validation itself.
parseValid :: Parser String
parseValid = do
  vs' <- many' $ utf8String "⋆" <|> return <$> idChar
  let vs = concat vs'
  c <- peekChar
  case c of
    Just '`' -> return vs
    Just '\'' -> (\x -> vs ++ "'" ++ x) <$> ("'" *> parseValid)
                 <|> return vs
    _ -> fail "outofvalid"
  where
    idChar = satisfy (`elem` "_.!#$%&*+/<=>?@\\|-~:^")
             <|> digit <|> letter_ascii

-- | Parses UTF8 strings from ByteString streams.
utf8String :: String -> Parser String
utf8String x = decodeUtf8 <$> string (encodeUtf8 x)

-- | Parses identifiers with help of 'parseValid'. Asks GHC for 'String' from the
-- string it deems valid.
identifier :: Parser (DocH mod Identifier)
identifier = do
  o <- idDelim
  vid <- parseValid
  e <- idDelim
  return $ DocIdentifier (o, vid, e)
  where
    idDelim = char '\'' <|> char '`'