aboutsummaryrefslogtreecommitdiff
path: root/src/Text/Pandoc/Readers/HTML.hs
blob: eb78979a3d6ddf6277eb846874a0f3b5e7e216a2 (plain)
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
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
{-# LANGUAGE FlexibleContexts     #-}
{-# LANGUAGE FlexibleInstances     #-}
{-# LANGUAGE LambdaCase            #-}
{-# LANGUAGE OverloadedStrings     #-}
{-# LANGUAGE ViewPatterns          #-}
{- |
   Module      : Text.Pandoc.Readers.HTML
   Copyright   : Copyright (C) 2006-2020 John MacFarlane
   License     : GNU GPL, version 2 or above

   Maintainer  : John MacFarlane <jgm@berkeley.edu>
   Stability   : alpha
   Portability : portable

Conversion of HTML to 'Pandoc' document.
-}
module Text.Pandoc.Readers.HTML ( readHtml
                                , htmlTag
                                , htmlInBalanced
                                , isInlineTag
                                , isBlockTag
                                , NamedTag(..)
                                , isTextTag
                                , isCommentTag
                                ) where

import Control.Applicative ((<|>))
import Control.Arrow (first)
import Control.Monad (guard, msum, mzero, unless, void)
import Control.Monad.Except (throwError)
import Control.Monad.Reader (ask, asks, lift, local, runReaderT)
import Data.ByteString.Base64 (encode)
import Data.Char (isAlphaNum, isLetter)
import Data.Default (Default (..), def)
import Data.Foldable (for_)
import Data.List.Split (splitWhen)
import qualified Data.Map as M
import Data.Maybe (fromMaybe, isJust, isNothing)
import Data.Monoid (First (..))
import qualified Data.Set as Set
import Data.Text (Text)
import qualified Data.Text as T
import Network.URI (nonStrictRelativeTo, parseURIReference)
import Text.HTML.TagSoup
import Text.HTML.TagSoup.Match
import Text.Pandoc.Builder (Blocks, Inlines, trimInlines)
import qualified Text.Pandoc.Builder as B
import Text.Pandoc.Class.PandocMonad (PandocMonad (..))
import Text.Pandoc.CSS (pickStyleAttrProps)
import qualified Text.Pandoc.UTF8 as UTF8
import Text.Pandoc.Definition
import Text.Pandoc.Readers.HTML.Parsing
import Text.Pandoc.Readers.HTML.Table (pTable)
import Text.Pandoc.Readers.HTML.TagCategories
import Text.Pandoc.Readers.HTML.Types
import Text.Pandoc.Readers.LaTeX (rawLaTeXInline)
import Text.Pandoc.Error
import Text.Pandoc.Logging
import Text.Pandoc.Options (
    Extension (Ext_epub_html_exts, Ext_empty_paragraphs, Ext_native_divs,
               Ext_native_spans, Ext_raw_html, Ext_line_blocks, Ext_raw_tex),
    ReaderOptions (readerExtensions, readerStripComments),
    extensionEnabled)
import Text.Pandoc.Parsing hiding ((<|>))
import Text.Pandoc.Shared (addMetaField, blocksToInlines', crFilter, escapeURI,
                           extractSpaces, htmlSpanLikeElements, safeRead, tshow)
import Text.Pandoc.Walk
import Text.Parsec.Error
import Text.TeXMath (readMathML, writeTeX)

-- | Convert HTML-formatted string to 'Pandoc' document.
readHtml :: PandocMonad m
         => ReaderOptions -- ^ Reader options
         -> Text        -- ^ String to parse (assumes @'\n'@ line endings)
         -> m Pandoc
readHtml opts inp = do
  let tags = stripPrefixes . canonicalizeTags $
             parseTagsOptions parseOptions{ optTagPosition = True }
             (crFilter inp)
      parseDoc = do
        blocks <- fixPlains False . mconcat <$> manyTill block eof
        meta <- stateMeta . parserState <$> getState
        bs' <- replaceNotes (B.toList blocks)
        reportLogMessages
        return $ Pandoc meta bs'
      getError (errorMessages -> ms) = case ms of
                                         []    -> ""
                                         (m:_) -> messageString m
  result <- flip runReaderT def $
       runParserT parseDoc
       (HTMLState def{ stateOptions = opts }
         [] Nothing Set.empty [] M.empty opts)
       "source" tags
  case result of
    Right doc -> return doc
    Left  err -> throwError $ PandocParseError $ T.pack $ getError err

replaceNotes :: PandocMonad m => [Block] -> TagParser m [Block]
replaceNotes bs = do
  st <- getState
  return $ walk (replaceNotes' (noteTable st)) bs

replaceNotes' :: [(Text, Blocks)] -> Inline -> Inline
replaceNotes' noteTbl (RawInline (Format "noteref") ref) =
  maybe (Str "") (Note . B.toList) $ lookup ref noteTbl
replaceNotes' _ x = x

setInChapter :: PandocMonad m => HTMLParser m s a -> HTMLParser m s a
setInChapter = local (\s -> s {inChapter = True})

setInPlain :: PandocMonad m => HTMLParser m s a -> HTMLParser m s a
setInPlain = local (\s -> s {inPlain = True})

pHtml :: PandocMonad m => TagParser m Blocks
pHtml = try $ do
  (TagOpen "html" attr) <- lookAhead pAny
  for_ (lookup "lang" attr) $
    updateState . B.setMeta "lang" . B.text
  pInTags "html" block

pBody :: PandocMonad m => TagParser m Blocks
pBody = pInTags "body" block

pHead :: PandocMonad m => TagParser m Blocks
pHead = pInTags "head" $ pTitle <|> pMetaTag <|> pBaseTag <|> (mempty <$ pAny)
  where pTitle = pInTags "title" inline >>= setTitle . trimInlines
        setTitle t = mempty <$ updateState (B.setMeta "title" t)
        pMetaTag = do
          mt <- pSatisfy (matchTagOpen "meta" [])
          let name = fromAttrib "name" mt
          if T.null name
             then return mempty
             else do
               let content = fromAttrib "content" mt
               updateState $ \s ->
                 let ps = parserState s in
                 s{ parserState = ps{
                      stateMeta = addMetaField name (B.text content)
                                   (stateMeta ps) } }
               return mempty
        pBaseTag = do
          bt <- pSatisfy (matchTagOpen "base" [])
          updateState $ \st -> st{ baseHref =
               parseURIReference $ T.unpack $ fromAttrib "href" bt }
          return mempty

block :: PandocMonad m => TagParser m Blocks
block = do
  res <- choice
            [ eSection
            , eSwitch B.para block
            , mempty <$ eFootnote
            , mempty <$ eTOC
            , mempty <$ eTitlePage
            , pPara
            , pHeader
            , pBlockQuote
            , pCodeBlock
            , pList
            , pHrule
            , pTable block
            , pHtml
            , pHead
            , pBody
            , pLineBlock
            , pDiv
            , pPlain
            , pFigure
            , pIframe
            , pRawHtmlBlock
            ]
  trace (T.take 60 $ tshow $ B.toList res)
  return res

namespaces :: PandocMonad m => [(Text, TagParser m Inlines)]
namespaces = [(mathMLNamespace, pMath True)]

mathMLNamespace :: Text
mathMLNamespace = "http://www.w3.org/1998/Math/MathML"

eSwitch :: (PandocMonad m, Monoid a)
        => (Inlines -> a)
        -> TagParser m a
        -> TagParser m a
eSwitch constructor parser = try $ do
  guardEnabled Ext_epub_html_exts
  pSatisfy (matchTagOpen "switch" [])
  cases <- getFirst . mconcat <$>
            manyTill (First <$> (eCase <* skipMany pBlank) )
              (lookAhead $ try $ pSatisfy (matchTagOpen "default" []))
  skipMany pBlank
  fallback <- pInTags "default" (skipMany pBlank *> parser <* skipMany pBlank)
  skipMany pBlank
  pSatisfy (matchTagClose "switch")
  return $ maybe fallback constructor cases

eCase :: PandocMonad m => TagParser m (Maybe Inlines)
eCase = do
  skipMany pBlank
  TagOpen _ attr' <- lookAhead $ pSatisfy (matchTagOpen "case" [])
  let attr = toStringAttr attr'
  case flip lookup namespaces =<< lookup "required-namespace" attr of
    Just p -> Just <$> pInTags "case" (skipMany pBlank *> p <* skipMany pBlank)
    Nothing -> Nothing <$ manyTill pAny (pSatisfy (matchTagClose "case"))

eFootnote :: PandocMonad m => TagParser m ()
eFootnote = try $ do
  let notes = ["footnote", "rearnote"]
  guardEnabled Ext_epub_html_exts
  (TagOpen tag attr') <- lookAhead pAny
  let attr = toStringAttr attr'
  guard $ maybe False (`elem` notes)
          (lookup "type" attr <|> lookup "epub:type" attr)
  let ident = fromMaybe "" (lookup "id" attr)
  content <- pInTags tag block
  addNote ident content

addNote :: PandocMonad m => Text -> Blocks -> TagParser m ()
addNote uid cont = updateState (\s -> s {noteTable = (uid, cont) : noteTable s})

eNoteref :: PandocMonad m => TagParser m Inlines
eNoteref = try $ do
  guardEnabled Ext_epub_html_exts
  TagOpen tag attr <-
    pSatisfy (\case
                 TagOpen _ as
                    -> (lookup "type" as <|> lookup "epub:type" as)
                        == Just "noteref"
                 _  -> False)
  ident <- case lookup "href" attr >>= T.uncons of
             Just ('#', rest) -> return rest
             _ -> mzero
  _ <- manyTill pAny (pSatisfy (\case
                                   TagClose t -> t == tag
                                   _          -> False))
  return $ B.rawInline "noteref" ident

-- Strip TOC if there is one, better to generate again
eTOC :: PandocMonad m => TagParser m ()
eTOC = try $ do
  guardEnabled Ext_epub_html_exts
  (TagOpen tag attr) <- lookAhead pAny
  guard $ (lookup "type" attr <|> lookup "epub:type" attr) == Just "toc"
  void (pInTags tag block)

pList :: PandocMonad m => TagParser m Blocks
pList = pBulletList <|> pOrderedList <|> pDefinitionList

pBulletList :: PandocMonad m => TagParser m Blocks
pBulletList = try $ do
  pSatisfy (matchTagOpen "ul" [])
  let nonItem = pSatisfy (\t ->
                  not (tagOpen (`elem` ["li","ol","ul","dl"]) (const True) t) &&
                  not (matchTagClose "ul" t))
  -- note: if they have an <ol> or <ul> not in scope of a <li>,
  -- treat it as a list item, though it's not valid xhtml...
  skipMany nonItem
  items <- manyTill (pListItem nonItem) (pCloses "ul")
  return $ B.bulletList $ map (fixPlains True) items

pListItem :: PandocMonad m => TagParser m a -> TagParser m Blocks
pListItem nonItem = do
  TagOpen _ attr' <- lookAhead $ pSatisfy (matchTagOpen "li" [])
  let attr = toStringAttr attr'
  let addId ident bs = case B.toList bs of
                           (Plain ils:xs) -> B.fromList (Plain
                                [Span (ident, [], []) ils] : xs)
                           _ -> B.divWith (ident, [], []) bs
  maybe id addId (lookup "id" attr) <$>
    pInTags "li" block <* skipMany nonItem

parseListStyleType :: Text -> ListNumberStyle
parseListStyleType "lower-roman" = LowerRoman
parseListStyleType "upper-roman" = UpperRoman
parseListStyleType "lower-alpha" = LowerAlpha
parseListStyleType "upper-alpha" = UpperAlpha
parseListStyleType "decimal"     = Decimal
parseListStyleType _             = DefaultStyle

parseTypeAttr :: Text -> ListNumberStyle
parseTypeAttr "i" = LowerRoman
parseTypeAttr "I" = UpperRoman
parseTypeAttr "a" = LowerAlpha
parseTypeAttr "A" = UpperAlpha
parseTypeAttr "1" = Decimal
parseTypeAttr _   = DefaultStyle

pOrderedList :: PandocMonad m => TagParser m Blocks
pOrderedList = try $ do
  TagOpen _ attribs' <- pSatisfy (matchTagOpen "ol" [])
  let attribs = toStringAttr attribs'
  let start = fromMaybe 1 $ lookup "start" attribs >>= safeRead
  let style = fromMaybe DefaultStyle
         $  (parseTypeAttr      <$> lookup "type" attribs)
        <|> (parseListStyleType <$> lookup "class" attribs)
        <|> (parseListStyleType <$> (lookup "style" attribs >>= pickListStyle))
        where
          pickListStyle = pickStyleAttrProps ["list-style-type", "list-style"]

  let nonItem = pSatisfy (\t ->
                  not (tagOpen (`elem` ["li","ol","ul","dl"]) (const True) t) &&
                  not (matchTagClose "ol" t))
  -- note: if they have an <ol> or <ul> not in scope of a <li>,
  -- treat it as a list item, though it's not valid xhtml...
  skipMany nonItem
  items <- manyTill (pListItem nonItem) (pCloses "ol")
  return $ B.orderedListWith (start, style, DefaultDelim) $ map (fixPlains True) items

pDefinitionList :: PandocMonad m => TagParser m Blocks
pDefinitionList = try $ do
  pSatisfy (matchTagOpen "dl" [])
  items <- manyTill pDefListItem (pCloses "dl")
  return $ B.definitionList items

pDefListItem :: PandocMonad m => TagParser m (Inlines, [Blocks])
pDefListItem = try $ do
  let nonItem = pSatisfy (\t -> not (matchTagOpen "dt" [] t) &&
                  not (matchTagOpen "dd" [] t) && not (matchTagClose "dl" t))
  terms <- many1 (try $ skipMany nonItem >> pInTags "dt" inline)
  defs  <- many1 (try $ skipMany nonItem >> pInTags "dd" block)
  skipMany nonItem
  let term = foldl1 (\x y -> x <> B.linebreak <> y) $ map trimInlines terms
  return (term, map (fixPlains True) defs)

fixPlains :: Bool -> Blocks -> Blocks
fixPlains inList bs = if any isParaish bs'
                         then B.fromList $ map plainToPara bs'
                         else bs
  where isParaish Para{}           = True
        isParaish CodeBlock{}      = True
        isParaish Header{}         = True
        isParaish BlockQuote{}     = True
        isParaish BulletList{}     = not inList
        isParaish OrderedList{}    = not inList
        isParaish DefinitionList{} = not inList
        isParaish _                = False
        plainToPara (Plain xs) = Para xs
        plainToPara x          = x
        bs' = B.toList bs

pRawTag :: PandocMonad m => TagParser m Text
pRawTag = do
  tag <- pAny
  let ignorable x = x `elem` ["html","head","body","!DOCTYPE","?xml"]
  if tagOpen ignorable (const True) tag || tagClose ignorable tag
     then return mempty
     else return $ renderTags' [tag]

pLineBlock :: PandocMonad m => TagParser m Blocks
pLineBlock = try $ do
  guardEnabled Ext_line_blocks
  _ <- pSatisfy $ tagOpen (=="div") (== [("class","line-block")])
  ils <- trimInlines . mconcat <$> manyTill inline (pSatisfy (tagClose (=="div")))
  let lns = map B.fromList $
            splitWhen (== LineBreak) $ filter (/= SoftBreak) $
            B.toList ils
  return $ B.lineBlock lns

pDiv :: PandocMonad m => TagParser m Blocks
pDiv = try $ do
  guardEnabled Ext_native_divs
  let isDivLike "div"     = True
      isDivLike "section" = True
      isDivLike "main"    = True
      isDivLike _         = False
  TagOpen tag attr' <- lookAhead $ pSatisfy $ tagOpen isDivLike (const True)
  let (ident, classes, kvs) = toAttr attr'
  contents <- pInTags tag block
  let classes' = if tag == "section"
                    then "section":classes
                    else classes
      kvs' = if tag == "main" && isNothing (lookup "role" kvs)
               then ("role", "main"):kvs
               else kvs
  return $ B.divWith (ident, classes', kvs') contents

pIframe :: PandocMonad m => TagParser m Blocks
pIframe = try $ do
  guardDisabled Ext_raw_html
  tag <- pSatisfy (tagOpen (=="iframe") (isJust . lookup "src"))
  pCloses "iframe" <|> eof
  url <- canonicalizeUrl $ fromAttrib "src" tag
  (bs, _) <- openURL url
  let inp = UTF8.toText bs
  opts <- readerOpts <$> getState
  Pandoc _ contents <- readHtml opts inp
  return $ B.divWith ("",["iframe"],[]) $ B.fromList contents

pRawHtmlBlock :: PandocMonad m => TagParser m Blocks
pRawHtmlBlock = do
  raw <- pHtmlBlock "script" <|> pHtmlBlock "style" <|> pHtmlBlock "textarea"
          <|> pRawTag
  exts <- getOption readerExtensions
  if extensionEnabled Ext_raw_html exts && not (T.null raw)
     then return $ B.rawBlock "html" raw
     else ignore raw

ignore :: (Monoid a, PandocMonad m) => Text -> TagParser m a
ignore raw = do
  pos <- getPosition
  -- raw can be null for tags like <!DOCTYPE>; see paRawTag
  -- in this case we don't want a warning:
  unless (T.null raw) $
    logMessage $ SkippedContent raw pos
  return mempty

pHtmlBlock :: PandocMonad m => Text -> TagParser m Text
pHtmlBlock t = try $ do
  open <- pSatisfy (matchTagOpen t [])
  contents <- manyTill pAny (pSatisfy (matchTagClose t))
  return $ renderTags' $ [open] <> contents <> [TagClose t]

-- Sets chapter context
eSection :: PandocMonad m => TagParser m Blocks
eSection = try $ do
  let matchChapter as = maybe False (T.isInfixOf "chapter")
                        (lookup "type" as <|> lookup "epub:type" as)
  let sectTag = tagOpen (`elem` sectioningContent) matchChapter
  TagOpen tag _ <- lookAhead $ pSatisfy sectTag
  setInChapter (pInTags tag block)

headerLevel :: Text -> TagParser m Int
headerLevel tagtype =
  case safeRead (T.drop 1 tagtype) of
        Just level ->
--          try (do
--            guardEnabled Ext_epub_html_exts
--            asks inChapter >>= guard
--            return (level - 1))
--            <|>
              return level
        Nothing -> Prelude.fail "Could not retrieve header level"

eTitlePage :: PandocMonad m => TagParser m ()
eTitlePage = try $ do
  let isTitlePage as = maybe False (T.isInfixOf "titlepage")
       (lookup "type" as <|> lookup "epub:type" as)
  let groupTag = tagOpen (\x -> x `elem` groupingContent || x == "section")
                          isTitlePage
  TagOpen tag _ <- lookAhead $ pSatisfy groupTag
  () <$ pInTags tag block

pHeader :: PandocMonad m => TagParser m Blocks
pHeader = try $ do
  TagOpen tagtype attr' <- pSatisfy $
                           tagOpen (`elem` ["h1","h2","h3","h4","h5","h6"])
                           (const True)
  let attr = toStringAttr attr'
  let bodyTitle = TagOpen tagtype attr' ~== TagOpen ("h1" :: Text)
                                               [("class","title")]
  level <- headerLevel tagtype
  contents <- trimInlines . mconcat <$> manyTill inline (pCloses tagtype <|> eof)
  let ident = fromMaybe "" $ lookup "id" attr
  let classes = maybe [] T.words $ lookup "class" attr
  let keyvals = [(k,v) | (k,v) <- attr, k /= "class", k /= "id"]
  attr'' <- registerHeader (ident, classes, keyvals) contents
  return $ if bodyTitle
              then mempty  -- skip a representation of the title in the body
              else B.headerWith attr'' level contents

pHrule :: PandocMonad m => TagParser m Blocks
pHrule = do
  pSelfClosing (=="hr") (const True)
  return B.horizontalRule

pBlockQuote :: PandocMonad m => TagParser m Blocks
pBlockQuote = do
  contents <- pInTags "blockquote" block
  return $ B.blockQuote $ fixPlains False contents

pPlain :: PandocMonad m => TagParser m Blocks
pPlain = do
  contents <- setInPlain $ trimInlines . mconcat <$> many1 inline
  if null contents
     then return mempty
     else return $ B.plain contents

pPara :: PandocMonad m => TagParser m Blocks
pPara = do
  contents <- trimInlines <$> pInTags "p" inline
  (do guardDisabled Ext_empty_paragraphs
      guard (null contents)
      return mempty)
    <|> return (B.para contents)

pFigure :: PandocMonad m => TagParser m Blocks
pFigure = try $ do
  TagOpen _ _ <- pSatisfy (matchTagOpen "figure" [])
  skipMany pBlank
  let pImg  = (\x -> (Just x, Nothing)) <$>
               (pInTag TagsOmittable "p" pImage <* skipMany pBlank)
      pCapt = (\x -> (Nothing, Just x)) <$> do
                bs <- pInTags "figcaption" block
                return $ blocksToInlines' $ B.toList bs
      pSkip = (Nothing, Nothing) <$ pSatisfy (not . matchTagClose "figure")
  res <- many (pImg <|> pCapt <|> pSkip)
  let mbimg = msum $ map fst res
  let mbcap = msum $ map snd res
  TagClose _ <- pSatisfy (matchTagClose "figure")
  let caption = fromMaybe mempty mbcap
  case B.toList <$> mbimg of
       Just [Image attr _ (url, tit)] ->
         return $ B.para $ B.imageWith attr url ("fig:" <> tit) caption
       _ -> mzero

pCodeBlock :: PandocMonad m => TagParser m Blocks
pCodeBlock = try $ do
  TagOpen _ attr' <- pSatisfy (matchTagOpen "pre" [])
  let attr = toAttr attr'
  contents <- manyTill pAny (pCloses "pre" <|> eof)
  let rawText = T.concat $ map tagToText contents
  -- drop leading newline if any
  let result' = case T.uncons rawText of
                     Just ('\n', xs) -> xs
                     _               -> rawText
  -- drop trailing newline if any
  let result = case T.unsnoc result' of
                    Just (result'', '\n') -> result''
                    _                     -> result'
  return $ B.codeBlockWith attr result

tagToText :: Tag Text -> Text
tagToText (TagText s)      = s
tagToText (TagOpen "br" _) = "\n"
tagToText _                = ""

inline :: PandocMonad m => TagParser m Inlines
inline = choice
           [ eNoteref
           , eSwitch id inline
           , pTagText
           , pQ
           , pEmph
           , pStrong
           , pSuperscript
           , pSubscript
           , pSpanLike
           , pSmall
           , pStrikeout
           , pUnderline
           , pLineBreak
           , pLink
           , pImage
           , pSvg
           , pBdo
           , pCode
           , pCodeWithClass [("samp","sample"),("var","variable")]
           , pSpan
           , pMath False
           , pScriptMath
           , pRawHtmlInline
           ]

pSelfClosing :: PandocMonad m
             => (Text -> Bool) -> ([Attribute Text] -> Bool)
             -> TagParser m (Tag Text)
pSelfClosing f g = do
  open <- pSatisfy (tagOpen f g)
  optional $ pSatisfy (tagClose f)
  return open

pQ :: PandocMonad m => TagParser m Inlines
pQ = choice $ map try [citedQuote, normalQuote]
  where citedQuote = do
          tag <- pSatisfy $ tagOpenLit "q" (any ((=="cite") . fst))

          url <- canonicalizeUrl $ fromAttrib "cite" tag
          let uid = fromMaybe (fromAttrib "name" tag) $
                       maybeFromAttrib "id" tag
          let cls = T.words $ fromAttrib "class" tag

          makeQuote $ B.spanWith (uid, cls, [("cite", escapeURI url)])
        normalQuote = do
          pSatisfy $ tagOpenLit "q" (const True)
          makeQuote id
        makeQuote wrapper = do
          ctx <- asks quoteContext
          let (constructor, innerContext) = case ctx of
                        InDoubleQuote -> (B.singleQuoted, InSingleQuote)
                        _             -> (B.doubleQuoted, InDoubleQuote)

          content <- withQuoteContext innerContext (mconcat <$> manyTill inline (pCloses "q"))
          return $ extractSpaces (constructor . wrapper) content

pEmph :: PandocMonad m => TagParser m Inlines
pEmph = pInlinesInTags "em" B.emph <|> pInlinesInTags "i" B.emph

pStrong :: PandocMonad m => TagParser m Inlines
pStrong = pInlinesInTags "strong" B.strong <|> pInlinesInTags "b" B.strong

pSuperscript :: PandocMonad m => TagParser m Inlines
pSuperscript = pInlinesInTags "sup" B.superscript

pSubscript :: PandocMonad m => TagParser m Inlines
pSubscript = pInlinesInTags "sub" B.subscript

pSpanLike :: PandocMonad m => TagParser m Inlines
pSpanLike =
  Set.foldr
    (\tagName acc -> acc <|> parseTag tagName)
    mzero
    htmlSpanLikeElements
  where
    parseTag tagName = do
      TagOpen _ attrs <- pSatisfy $ tagOpenLit tagName (const True)
      let (ids, cs, kvs) = toAttr attrs
      content <- mconcat <$> manyTill inline (pCloses tagName <|> eof)
      return $ B.spanWith (ids, tagName : cs, kvs) content

pSmall :: PandocMonad m => TagParser m Inlines
pSmall = pInlinesInTags "small" (B.spanWith ("",["small"],[]))

pStrikeout :: PandocMonad m => TagParser m Inlines
pStrikeout =
  pInlinesInTags "s" B.strikeout <|>
    pInlinesInTags "strike" B.strikeout <|>
    pInlinesInTags "del" B.strikeout <|>
    try (do pSatisfy (matchTagOpen "span" [("class","strikeout")])
            contents <- mconcat <$> manyTill inline (pCloses "span")
            return $ B.strikeout contents)

pUnderline :: PandocMonad m => TagParser m Inlines
pUnderline = pInlinesInTags "u" B.underline <|> pInlinesInTags "ins" B.underline

pLineBreak :: PandocMonad m => TagParser m Inlines
pLineBreak = do
  pSelfClosing (=="br") (const True)
  return B.linebreak

pLink :: PandocMonad m => TagParser m Inlines
pLink = try $ do
  tag <- pSatisfy $ tagOpenLit "a" (const True)
  let title = fromAttrib "title" tag
  -- take id from id attribute if present, otherwise name
  let uid = fromMaybe (fromAttrib "name" tag) $
               maybeFromAttrib "id" tag
  let cls = T.words $ fromAttrib "class" tag
  lab <- mconcat <$> manyTill inline (pCloses "a")
  -- check for href; if href, then a link, otherwise a span
  case maybeFromAttrib "href" tag of
       Nothing   ->
         return $ extractSpaces (B.spanWith (uid, cls, [])) lab
       Just url' -> do
         url <- canonicalizeUrl url'
         return $ extractSpaces (B.linkWith (uid, cls, []) (escapeURI url) title) lab

pImage :: PandocMonad m => TagParser m Inlines
pImage = do
  tag <- pSelfClosing (=="img") (isJust . lookup "src")
  url <- canonicalizeUrl $ fromAttrib "src" tag
  let title = fromAttrib "title" tag
  let alt = fromAttrib "alt" tag
  let uid = fromAttrib "id" tag
  let cls = T.words $ fromAttrib "class" tag
  let getAtt k = case fromAttrib k tag of
                   "" -> []
                   v  -> [(k, v)]
  let kvs = concatMap getAtt ["width", "height", "sizes", "srcset"]
  return $ B.imageWith (uid, cls, kvs) (escapeURI url) title (B.text alt)

pSvg :: PandocMonad m => TagParser m Inlines
pSvg = do
  guardDisabled Ext_raw_html
  -- if raw_html enabled, parse svg tag as raw
  opent@(TagOpen _ attr') <- pSatisfy (matchTagOpen "svg" [])
  let (ident,cls,_) = toAttr attr'
  contents <- many (notFollowedBy (pCloses "svg") >> pAny)
  closet <- TagClose "svg" <$ (pCloses "svg" <|> eof)
  let rawText = T.strip $ renderTags' (opent : contents ++ [closet])
  let svgData = "data:image/svg+xml;base64," <>
                   UTF8.toText (encode $ UTF8.fromText rawText)
  return $ B.imageWith (ident,cls,[]) svgData mempty mempty

pCodeWithClass :: PandocMonad m => [(T.Text,Text)] -> TagParser m Inlines
pCodeWithClass elemToClass = try $ do
  let tagTest = flip elem . fmap fst $ elemToClass
  TagOpen open attr' <- pSatisfy $ tagOpen tagTest (const True)
  result <- manyTill pAny (pCloses open)
  let (ids,cs,kvs) = toAttr attr'
      cs'          = maybe cs (:cs) . lookup open $ elemToClass
  return . B.codeWith (ids,cs',kvs) .
    T.unwords . T.lines . innerText $ result

pCode :: PandocMonad m => TagParser m Inlines
pCode = try $ do
  (TagOpen open attr') <- pSatisfy $ tagOpen (`elem` ["code","tt"]) (const True)
  let attr = toAttr attr'
  result <- manyTill pAny (pCloses open)
  return $ B.codeWith attr $ T.unwords $ T.lines $ innerText result

-- https://developer.mozilla.org/en-US/docs/Web/HTML/Element/bdo
-- Bidirectional Text Override
pBdo :: PandocMonad m => TagParser m Inlines
pBdo = try $ do
  TagOpen _ attr' <- lookAhead $ pSatisfy $ tagOpen (=="bdo") (const True)
  let attr = toStringAttr attr'
  contents <- pInTags "bdo" inline
  return $ case lookup "dir" attr of
    -- Only bdo with a direction matters
    Just dir -> B.spanWith ("", [], [("dir",T.toLower dir)]) contents
    Nothing  -> contents

pSpan :: PandocMonad m => TagParser m Inlines
pSpan = try $ do
  guardEnabled Ext_native_spans
  TagOpen _ attr' <- lookAhead $ pSatisfy $ tagOpen (=="span") (const True)
  let attr = toAttr attr'
  contents <- pInTags "span" inline
  let isSmallCaps = fontVariant == "small-caps" || "smallcaps" `elem` classes
                    where styleAttr   = fromMaybe "" $ lookup "style" attr'
                          fontVariant = fromMaybe "" $ pickStyleAttrProps ["font-variant"] styleAttr
                          classes     = maybe []
                                          T.words $ lookup "class" attr'
  let tag = if isSmallCaps then B.smallcaps else B.spanWith attr
  return $ tag contents

pRawHtmlInline :: PandocMonad m => TagParser m Inlines
pRawHtmlInline = do
  inplain <- asks inPlain
  result <- pSatisfy (tagComment (const True))
            <|> if inplain
                   then pSatisfy (not . isBlockTag)
                   else pSatisfy isInlineTag
  exts <- getOption readerExtensions
  let raw = renderTags' [result]
  if extensionEnabled Ext_raw_html exts
     then return $ B.rawInline "html" raw
     else ignore raw

mathMLToTeXMath :: Text -> Either Text Text
mathMLToTeXMath s = writeTeX <$> readMathML s

pScriptMath :: PandocMonad m => TagParser m Inlines
pScriptMath = try $ do
  TagOpen _ attr' <- pSatisfy $ tagOpen (=="script") (const True)
  isdisplay <- case lookup "type" attr' of
                    Just x | "math/tex" `T.isPrefixOf` x
                      -> return $ "display" `T.isSuffixOf` x
                    _ -> mzero
  contents <- innerText <$> manyTill pAny (pSatisfy (matchTagClose "script"))
  return $ (if isdisplay then B.displayMath else B.math) contents

pMath :: PandocMonad m => Bool -> TagParser m Inlines
pMath inCase = try $ do
  open@(TagOpen _ attr') <- pSatisfy $ tagOpen (=="math") (const True)
  -- we'll assume math tags are MathML unless specially marked
  -- otherwise...
  let attr = toStringAttr attr'
  unless inCase $
    guard (maybe True (== mathMLNamespace) (lookup "xmlns" attr))
  contents <- manyTill pAny (pSatisfy (matchTagClose "math"))
  case mathMLToTeXMath (renderTags $
          [open] <> contents <> [TagClose "math"]) of
       Left _   -> return $ B.spanWith ("",["math"],attr) $ B.text $
                             innerText contents
       Right "" -> return mempty
       Right x  -> return $ case lookup "display" attr of
                                 Just "block" -> B.displayMath x
                                 _            -> B.math x

pInlinesInTags :: PandocMonad m => Text -> (Inlines -> Inlines)
               -> TagParser m Inlines
pInlinesInTags tagtype f = extractSpaces f <$> pInTags tagtype inline

pTagText :: PandocMonad m => TagParser m Inlines
pTagText = try $ do
  (TagText str) <- pSatisfy isTagText
  st <- getState
  qu <- ask
  parsed <- lift $ lift $
            flip runReaderT qu $ runParserT (many pTagContents) st "text" str
  case parsed of
       Left _        -> throwError $ PandocParseError $
                        "Could not parse `" <> str <> "'"
       Right result  -> return $ mconcat result

type InlinesParser m = HTMLParser m Text

pTagContents :: PandocMonad m => InlinesParser m Inlines
pTagContents =
      B.displayMath <$> mathDisplay
  <|> B.math        <$> mathInline
  <|> pStr
  <|> pSpace
  <|> smartPunctuation pTagContents
  <|> pRawTeX
  <|> pSymbol
  <|> pBad

pRawTeX :: PandocMonad m => InlinesParser m Inlines
pRawTeX = do
  lookAhead $ try $ do
    char '\\'
    choice $ map (try . string) ["begin", "eqref", "ref"]
  guardEnabled Ext_raw_tex
  inp <- getInput
  st <- getState
  res <- lift $ runParserT (withRaw rawLaTeXInline) st "chunk" inp
  case res of
       Left _                -> mzero
       Right (contents, raw) -> do
         _ <- count (T.length raw) anyChar
         return $ B.rawInline "tex" contents

pStr :: PandocMonad m => InlinesParser m Inlines
pStr = do
  result <- many1 $ satisfy $ \c ->
                     not (isSpace c) && not (isSpecial c) && not (isBad c)
  updateLastStrPos
  return $ B.str $ T.pack result

isSpecial :: Char -> Bool
isSpecial '"'     = True
isSpecial '\''    = True
isSpecial '.'     = True
isSpecial '-'     = True
isSpecial '$'     = True
isSpecial '\\'    = True
isSpecial '\8216' = True
isSpecial '\8217' = True
isSpecial '\8220' = True
isSpecial '\8221' = True
isSpecial _       = False

pSymbol :: PandocMonad m => InlinesParser m Inlines
pSymbol = B.str . T.singleton <$> satisfy isSpecial

isBad :: Char -> Bool
isBad c = c >= '\128' && c <= '\159' -- not allowed in HTML

pBad :: PandocMonad m => InlinesParser m Inlines
pBad = do
  c <- satisfy isBad
  let c' = case c of
                '\128' -> '\8364'
                '\130' -> '\8218'
                '\131' -> '\402'
                '\132' -> '\8222'
                '\133' -> '\8230'
                '\134' -> '\8224'
                '\135' -> '\8225'
                '\136' -> '\710'
                '\137' -> '\8240'
                '\138' -> '\352'
                '\139' -> '\8249'
                '\140' -> '\338'
                '\142' -> '\381'
                '\145' -> '\8216'
                '\146' -> '\8217'
                '\147' -> '\8220'
                '\148' -> '\8221'
                '\149' -> '\8226'
                '\150' -> '\8211'
                '\151' -> '\8212'
                '\152' -> '\732'
                '\153' -> '\8482'
                '\154' -> '\353'
                '\155' -> '\8250'
                '\156' -> '\339'
                '\158' -> '\382'
                '\159' -> '\376'
                _      -> '?'
  return $ B.str $ T.singleton c'

pSpace :: PandocMonad m => InlinesParser m Inlines
pSpace = many1 (satisfy isSpace) >>= \xs ->
            if '\n' `elem` xs
               then return B.softbreak
               else return B.space

class NamedTag a where
  getTagName :: a -> Maybe Text

instance NamedTag (Tag Text) where
  getTagName (TagOpen t _) = Just t
  getTagName (TagClose t)  = Just t
  getTagName _             = Nothing

instance NamedTag (Tag String) where
  getTagName (TagOpen t _) = Just (T.pack t)
  getTagName (TagClose t)  = Just (T.pack t)
  getTagName _             = Nothing

isInlineTag :: NamedTag (Tag a) => Tag a -> Bool
isInlineTag t =
  isCommentTag t || case getTagName t of
                           Nothing  -> False
                           Just x   -> x `Set.notMember` blockTags ||
                                       T.take 1 x == "?" -- processing instr.

isBlockTag :: NamedTag (Tag a) => Tag a -> Bool
isBlockTag t = isBlockTagName || isTagComment t
                 where isBlockTagName =
                         case getTagName t of
                              Just x
                                | "?" `T.isPrefixOf` x -> True
                                | "!" `T.isPrefixOf` x -> True
                                | otherwise -> x `Set.member` blockTags
                                    || x `Set.member` eitherBlockOrInline
                              Nothing -> False

isTextTag :: Tag a -> Bool
isTextTag = tagText (const True)

isCommentTag :: Tag a -> Bool
isCommentTag = tagComment (const True)

--- parsers for use in markdown, textile readers

-- | Matches a stretch of HTML in balanced tags.
htmlInBalanced :: Monad m
               => (Tag Text -> Bool)
               -> ParserT Text st m Text
htmlInBalanced f = try $ do
  lookAhead (char '<')
  inp <- getInput
  let ts = canonicalizeTags $
        parseTagsOptions parseOptions{ optTagWarning = True,
                                       optTagPosition = True } inp
  case ts of
    (TagPosition sr sc : t@(TagOpen tn _) : rest) -> do
       guard $ f t
       guard $ not $ hasTagWarning (t : take 1 rest)
       case htmlInBalanced' tn (t:rest) of
            []  -> mzero
            xs  -> case reverse xs of
                        (TagClose _ : TagPosition er ec : _) -> do
                          let ls = er - sr
                          let cs = ec - sc
                          lscontents <- T.unlines <$> count ls anyLine
                          cscontents <- count cs anyChar
                          closetag <- do
                            x <- many (satisfy (/='>'))
                            char '>'
                            return (x <> ">")
                          return $ lscontents <> T.pack cscontents <> T.pack closetag
                        _ -> mzero
    _ -> mzero

htmlInBalanced' :: Text
                -> [Tag Text]
                -> [Tag Text]
htmlInBalanced' tagname ts = fromMaybe [] $ go 0 ts
  where go :: Int -> [Tag Text] -> Maybe [Tag Text]
        go n (t@(TagOpen tn' _):rest) | tn' == tagname =
              (t :) <$> go (n + 1) rest
        go 1 (t@(TagClose tn'):_) | tn' == tagname =
              return [t]
        go n (t@(TagClose tn'):rest)  | tn' == tagname =
              (t :) <$> go (n - 1) rest
        go n (t:ts') = (t :) <$> go n ts'
        go _ [] = mzero

hasTagWarning :: [Tag a] -> Bool
hasTagWarning (TagWarning _:_) = True
hasTagWarning _                = False

-- | Matches a tag meeting a certain condition.
htmlTag :: (HasReaderOptions st, Monad m)
        => (Tag Text -> Bool)
        -> ParserT Text st m (Tag Text, Text)
htmlTag f = try $ do
  lookAhead (char '<')
  startpos <- getPosition
  inp <- getInput
  let ts = canonicalizeTags $ parseTagsOptions
                               parseOptions{ optTagWarning = False
                                           , optTagPosition = True }
                               (inp <> " ") -- add space to ensure that
                               -- we get a TagPosition after the tag
  (next, ln, col) <- case ts of
                      (TagPosition{} : next : TagPosition ln col : _)
                        | f next -> return (next, ln, col)
                      _ -> mzero

  -- <www.boe.es/buscar/act.php?id=BOE-A-1996-8930#a66>
  -- should NOT be parsed as an HTML tag, see #2277,
  -- so we exclude . even though it's a valid character
  -- in XML element names
  let isNameChar c = isAlphaNum c || c == ':' || c == '-' || c == '_'
  let isName s = case T.uncons s of
                   Nothing      -> False
                   Just (c, cs) -> isLetter c && T.all isNameChar cs
  let isPI s = case T.uncons s of
                 Just ('?', _) -> True -- processing instruction
                 _             -> False
  let endpos = if ln == 1
                  then setSourceColumn startpos
                         (sourceColumn startpos + (col - 1))
                  else setSourceColumn (setSourceLine startpos
                                        (sourceLine startpos + (ln - 1)))
                         col
  let endAngle = try $
        do char '>'
           pos <- getPosition
           guard $ pos >= endpos

  let handleTag tagname = do
       -- basic sanity check, since the parser is very forgiving
       -- and finds tags in stuff like x<y)
       guard $ isName tagname || isPI tagname
       guard $ not $ T.null tagname
       -- <https://example.org> should NOT be a tag either.
       -- tagsoup will parse it as TagOpen "https:" [("example.org","")]
       guard $ T.last tagname /= ':'
       char '<'
       rendered <- manyTill anyChar endAngle
       return (next, T.pack $ "<" ++ rendered ++ ">")
  case next of
       TagComment s
         | "<!--" `T.isPrefixOf` inp -> do
          string "<!--"
          count (T.length s) anyChar
          string "-->"
          stripComments <- getOption readerStripComments
          if stripComments
             then return (next, "")
             else return (next, "<!--" <> s <> "-->")
         | otherwise -> Prelude.fail "bogus comment mode, HTML5 parse error"
       TagOpen tagname attr -> do
         guard $ isPI tagname || all (isName . fst) attr
         handleTag tagname
       TagClose tagname ->
         handleTag tagname
       _ -> mzero

-- Strip namespace prefixes
stripPrefixes :: [Tag Text] -> [Tag Text]
stripPrefixes = map stripPrefix

stripPrefix :: Tag Text -> Tag Text
stripPrefix (TagOpen s as) =
    TagOpen (stripPrefix' s) (map (first stripPrefix') as)
stripPrefix (TagClose s) = TagClose (stripPrefix' s)
stripPrefix x = x

stripPrefix' :: Text -> Text
stripPrefix' s =
  if T.null t then s else T.drop 1 t
  where (_, t) = T.span (/= ':') s

-- Utilities

-- | Adjusts a url according to the document's base URL.
canonicalizeUrl :: PandocMonad m => Text -> TagParser m Text
canonicalizeUrl url = do
  mbBaseHref <- baseHref <$> getState
  return $ case (parseURIReference (T.unpack url), mbBaseHref) of
                (Just rel, Just bs) -> tshow (rel `nonStrictRelativeTo` bs)
                _                   -> url

-- For now we need a special version here; the one in Shared has String type
renderTags' :: [Tag Text] -> Text
renderTags' = renderTagsOptions
               renderOptions{ optMinimize = matchTags ["hr", "br", "img",
                                                       "meta", "link"]
                            , optRawTag   = matchTags ["script", "style"] }
              where matchTags tags = flip elem tags . T.toLower


-- EPUB Specific
--
--
{-

types :: [(String, ([String], Int))]
types =  -- Document divisions
   map (\s -> (s, (["section", "body"], 0)))
    ["volume", "part", "chapter", "division"]
  <> -- Document section and components
  [
    ("abstract",  ([], 0))]
-}