1{-# LANGUAGE OverloadedStrings #-}
2{-# LANGUAGE FlexibleContexts #-}
3{-# LANGUAGE ScopedTypeVariables #-}
4{- |
5   Module      : Text.Pandoc.Readers.CommonMark
6   Copyright   : Copyright (C) 2015-2021 John MacFarlane
7   License     : GNU GPL, version 2 or above
8
9   Maintainer  : John MacFarlane <jgm@berkeley.edu>
10   Stability   : alpha
11   Portability : portable
12
13Conversion of CommonMark-formatted plain text to 'Pandoc' document.
14
15CommonMark is a strongly specified variant of Markdown: http://commonmark.org.
16-}
17module Text.Pandoc.Readers.CommonMark (readCommonMark)
18where
19
20import Commonmark
21import Commonmark.Extensions
22import Commonmark.Pandoc
23import Data.Text (Text)
24import Text.Pandoc.Class.PandocMonad (PandocMonad)
25import Text.Pandoc.Definition
26import Text.Pandoc.Builder as B
27import Text.Pandoc.Options
28import Text.Pandoc.Error
29import Text.Pandoc.Readers.Metadata (yamlMetaBlock)
30import Control.Monad.Except
31import Data.Functor.Identity (runIdentity)
32import Data.Typeable
33import Text.Pandoc.Parsing (runParserT, getInput,
34                            runF, defaultParserState, option, many1, anyChar,
35                            Sources(..), ToSources(..), ParserT, Future,
36                            sourceName)
37import qualified Data.Text as T
38
39-- | Parse a CommonMark formatted string into a 'Pandoc' structure.
40readCommonMark :: (PandocMonad m, ToSources a)
41               => ReaderOptions -> a -> m Pandoc
42readCommonMark opts s
43  | isEnabled Ext_yaml_metadata_block opts = do
44    let sources = toSources s
45    let toks = concatMap sourceToToks (unSources sources)
46    res <- runParserT (do meta <- yamlMetaBlock (metaValueParser opts)
47                          rest <- getInput
48                          return (meta, rest))
49                      defaultParserState "YAML metadata" (toSources s)
50    case res of
51      Left _ -> readCommonMarkBody opts sources toks
52      Right (meta, rest) -> do
53        -- strip off metadata section and parse body
54        let body = concatMap sourceToToks (unSources rest)
55        Pandoc _ bs <- readCommonMarkBody opts sources body
56        return $ Pandoc (runF meta defaultParserState) bs
57  | otherwise = do
58    let sources = toSources s
59    let toks = concatMap sourceToToks (unSources sources)
60    readCommonMarkBody opts sources toks
61
62sourceToToks :: (SourcePos, Text) -> [Tok]
63sourceToToks (pos, s) = tokenize (sourceName pos) s
64
65metaValueParser :: Monad m
66                => ReaderOptions -> ParserT Sources st m (Future st MetaValue)
67metaValueParser opts = do
68  inp <- option "" $ T.pack <$> many1 anyChar
69  let toks = concatMap sourceToToks (unSources (toSources inp))
70  case runIdentity (parseCommonmarkWith (specFor opts) toks) of
71     Left _ -> mzero
72     Right (Cm bls :: Cm () Blocks) -> return $ return $ B.toMetaValue bls
73
74readCommonMarkBody :: PandocMonad m => ReaderOptions -> Sources -> [Tok] -> m Pandoc
75readCommonMarkBody opts s toks
76  | isEnabled Ext_sourcepos opts =
77    case runIdentity (parseCommonmarkWith (specFor opts) toks) of
78      Left err -> throwError $ PandocParsecError s err
79      Right (Cm bls :: Cm SourceRange Blocks) -> return $ B.doc bls
80  | otherwise =
81    case runIdentity (parseCommonmarkWith (specFor opts) toks) of
82      Left err -> throwError $ PandocParsecError s err
83      Right (Cm bls :: Cm () Blocks) -> return $ B.doc bls
84
85specFor :: (Monad m, Typeable m, Typeable a,
86            Rangeable (Cm a Inlines), Rangeable (Cm a Blocks))
87        => ReaderOptions -> SyntaxSpec m (Cm a Inlines) (Cm a Blocks)
88specFor opts = foldr ($) defaultSyntaxSpec exts
89 where
90  exts = [ (hardLineBreaksSpec <>) | isEnabled Ext_hard_line_breaks opts ] ++
91         [ (smartPunctuationSpec <>) | isEnabled Ext_smart opts ] ++
92         [ (strikethroughSpec <>) | isEnabled Ext_strikeout opts ] ++
93         [ (superscriptSpec <>) | isEnabled Ext_superscript opts ] ++
94         [ (subscriptSpec <>) | isEnabled Ext_subscript opts ] ++
95         [ (mathSpec <>) | isEnabled Ext_tex_math_dollars opts ] ++
96         [ (fancyListSpec <>) | isEnabled Ext_fancy_lists opts ] ++
97         [ (fencedDivSpec <>) | isEnabled Ext_fenced_divs opts ] ++
98         [ (bracketedSpanSpec <>) | isEnabled Ext_bracketed_spans opts ] ++
99         [ (rawAttributeSpec <>) | isEnabled Ext_raw_attribute opts ] ++
100         [ (attributesSpec <>) | isEnabled Ext_attributes opts ] ++
101         [ (<> pipeTableSpec) | isEnabled Ext_pipe_tables opts ] ++
102            -- see #6739
103         [ (autolinkSpec <>) | isEnabled Ext_autolink_bare_uris opts ] ++
104         [ (emojiSpec <>) | isEnabled Ext_emoji opts ] ++
105         [ (autoIdentifiersSpec <>)
106           | isEnabled Ext_gfm_auto_identifiers opts
107           , not (isEnabled Ext_ascii_identifiers opts) ] ++
108         [ (autoIdentifiersAsciiSpec <>)
109           | isEnabled Ext_gfm_auto_identifiers opts
110           , isEnabled Ext_ascii_identifiers opts ] ++
111         [ (implicitHeadingReferencesSpec <>)
112           | isEnabled Ext_implicit_header_references opts ] ++
113         [ (footnoteSpec <>) | isEnabled Ext_footnotes opts ] ++
114         [ (definitionListSpec <>) | isEnabled Ext_definition_lists opts ] ++
115         [ (taskListSpec <>) | isEnabled Ext_task_lists opts ] ++
116         [ (rebaseRelativePathsSpec <>)
117           | isEnabled Ext_rebase_relative_paths opts ]
118
119