{-# LANGUAGE OverloadedStrings   #-}
{-# LANGUAGE MonoLocalBinds      #-}
{-# LANGUAGE RankNTypes          #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TupleSections       #-}
{- |
   Module      : Text.Pandoc.Readers
   Copyright   : Copyright (C) 2006-2023 John MacFarlane
   License     : GNU GPL, version 2 or above

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

This helper module exports the readers.

Note:  all of the readers assume that the input text has @'\n'@
line endings.  So if you get your input text from a web form,
you should remove @'\r'@ characters using @filter (/='\r')@.

-}

module Text.Pandoc.Readers
  (
    -- * Readers: converting /to/ Pandoc format
    Reader (..)
  , readers
  , readDocx
  , readODT
  , readMarkdown
  , readCommonMark
  , readCreole
  , readDokuWiki
  , readMediaWiki
  , readVimwiki
  , readRST
  , readOrg
  , readLaTeX
  , readHtml
  , readJATS
  , readJira
  , readTextile
  , readDocBook
  , readOPML
  , readHaddock
  , readNative
  , readJSON
  , readTWiki
  , readTikiWiki
  , readTxt2Tags
  , readEPUB
  , readMuse
  , readFB2
  , readIpynb
  , readCSV
  , readTSV
  , readCslJson
  , readBibTeX
  , readBibLaTeX
  , readEndNoteXML
  , readRIS
  , readRTF
  , readTypst
  -- * Miscellaneous
  , getReader
  , getDefaultExtensions
  ) where

import Control.Monad.Except (throwError)
import Data.Aeson
import qualified Data.ByteString.Lazy as BL
import Data.Text (Text)
import qualified Data.Text as T
import Text.Pandoc.Class
import Text.Pandoc.Definition
import Text.Pandoc.Error
import Text.Pandoc.Extensions
import qualified Text.Pandoc.Format as Format
import Text.Pandoc.Options
import Text.Pandoc.Readers.CommonMark
import Text.Pandoc.Readers.Markdown
import Text.Pandoc.Readers.Creole
import Text.Pandoc.Readers.DocBook
import Text.Pandoc.Readers.Docx
import Text.Pandoc.Readers.DokuWiki
import Text.Pandoc.Readers.EPUB
import Text.Pandoc.Readers.FB2
import Text.Pandoc.Readers.Ipynb
import Text.Pandoc.Readers.Haddock
import Text.Pandoc.Readers.HTML (readHtml)
import Text.Pandoc.Readers.JATS (readJATS)
import Text.Pandoc.Readers.Jira (readJira)
import Text.Pandoc.Readers.LaTeX
import Text.Pandoc.Readers.MediaWiki
import Text.Pandoc.Readers.Muse
import Text.Pandoc.Readers.Native
import Text.Pandoc.Readers.ODT
import Text.Pandoc.Readers.OPML
import Text.Pandoc.Readers.Org
import Text.Pandoc.Readers.RST
import Text.Pandoc.Readers.Textile
import Text.Pandoc.Readers.TikiWiki
import Text.Pandoc.Readers.TWiki
import Text.Pandoc.Readers.Txt2Tags
import Text.Pandoc.Readers.Vimwiki
import Text.Pandoc.Readers.Man
import Text.Pandoc.Readers.CSV
import Text.Pandoc.Readers.CslJson
import Text.Pandoc.Readers.BibTeX
import Text.Pandoc.Readers.EndNote
import Text.Pandoc.Readers.RIS
import Text.Pandoc.Readers.RTF
import Text.Pandoc.Readers.Typst
import qualified Text.Pandoc.UTF8 as UTF8
import Text.Pandoc.Sources (ToSources(..), sourcesToText)

data Reader m = TextReader (forall a . ToSources a =>
                                ReaderOptions -> a -> m Pandoc)
              | ByteStringReader (ReaderOptions -> BL.ByteString -> m Pandoc)

-- | Association list of formats and readers.
readers :: PandocMonad m => [(Text, Reader m)]
readers :: forall (m :: * -> *). PandocMonad m => [(Text, Reader m)]
readers = [(Text
"native"       , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readNative)
          ,(Text
"json"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readJSON)
          ,(Text
"markdown"     , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMarkdown)
          ,(Text
"markdown_strict" , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMarkdown)
          ,(Text
"markdown_phpextra" , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMarkdown)
          ,(Text
"markdown_github" , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMarkdown)
          ,(Text
"markdown_mmd",  forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMarkdown)
          ,(Text
"commonmark"   , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCommonMark)
          ,(Text
"commonmark_x" , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCommonMark)
          ,(Text
"creole"       , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCreole)
          ,(Text
"dokuwiki"     , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readDokuWiki)
          ,(Text
"gfm"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCommonMark)
          ,(Text
"rst"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readRST)
          ,(Text
"mediawiki"    , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMediaWiki)
          ,(Text
"vimwiki"      , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readVimwiki)
          ,(Text
"docbook"      , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readDocBook)
          ,(Text
"opml"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readOPML)
          ,(Text
"org"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readOrg)
          ,(Text
"textile"      , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTextile) -- TODO : textile+lhs
          ,(Text
"html"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readHtml)
          ,(Text
"jats"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readJATS)
          ,(Text
"jira"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readJira)
          ,(Text
"latex"        , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readLaTeX)
          ,(Text
"haddock"      , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readHaddock)
          ,(Text
"twiki"        , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTWiki)
          ,(Text
"tikiwiki"     , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTikiWiki)
          ,(Text
"docx"         , forall (m :: * -> *).
(ReaderOptions -> ByteString -> m Pandoc) -> Reader m
ByteStringReader forall (m :: * -> *).
PandocMonad m =>
ReaderOptions -> ByteString -> m Pandoc
readDocx)
          ,(Text
"odt"          , forall (m :: * -> *).
(ReaderOptions -> ByteString -> m Pandoc) -> Reader m
ByteStringReader forall (m :: * -> *).
PandocMonad m =>
ReaderOptions -> ByteString -> m Pandoc
readODT)
          ,(Text
"t2t"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTxt2Tags)
          ,(Text
"epub"         , forall (m :: * -> *).
(ReaderOptions -> ByteString -> m Pandoc) -> Reader m
ByteStringReader forall (m :: * -> *).
PandocMonad m =>
ReaderOptions -> ByteString -> m Pandoc
readEPUB)
          ,(Text
"muse"         , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMuse)
          ,(Text
"man"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readMan)
          ,(Text
"fb2"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readFB2)
          ,(Text
"ipynb"        , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readIpynb)
          ,(Text
"csv"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCSV)
          ,(Text
"tsv"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTSV)
          ,(Text
"csljson"      , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readCslJson)
          ,(Text
"bibtex"       , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readBibTeX)
          ,(Text
"biblatex"     , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readBibLaTeX)
          ,(Text
"endnotexml"   , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readEndNoteXML)
          ,(Text
"ris"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readRIS)
          ,(Text
"rtf"          , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readRTF)
          ,(Text
"typst"        , forall (m :: * -> *).
(forall a. ToSources a => ReaderOptions -> a -> m Pandoc)
-> Reader m
TextReader forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readTypst)
           ]

-- | Retrieve reader, extensions based on format spec (format+extensions).
getReader :: PandocMonad m => Format.FlavoredFormat -> m (Reader m, Extensions)
getReader :: forall (m :: * -> *).
PandocMonad m =>
FlavoredFormat -> m (Reader m, Extensions)
getReader FlavoredFormat
flvrd = do
  let readerName :: Text
readerName = FlavoredFormat -> Text
Format.formatName FlavoredFormat
flvrd
  case forall a b. Eq a => a -> [(a, b)] -> Maybe b
lookup Text
readerName forall (m :: * -> *). PandocMonad m => [(Text, Reader m)]
readers of
    Maybe (Reader m)
Nothing  -> forall e (m :: * -> *) a. MonadError e m => e -> m a
throwError forall a b. (a -> b) -> a -> b
$ Text -> PandocError
PandocUnknownReaderError Text
readerName
    Just  Reader m
r  -> (Reader m
r,) forall (f :: * -> *) a b. Functor f => (a -> b) -> f a -> f b
<$>
      forall (m :: * -> *).
PandocMonad m =>
ExtensionsConfig -> FlavoredFormat -> m Extensions
Format.applyExtensionsDiff (Text -> ExtensionsConfig
Format.getExtensionsConfig Text
readerName) FlavoredFormat
flvrd

-- | Read pandoc document from JSON format.
readJSON :: (PandocMonad m, ToSources a)
         => ReaderOptions
         -> a
         -> m Pandoc
readJSON :: forall (m :: * -> *) a.
(PandocMonad m, ToSources a) =>
ReaderOptions -> a -> m Pandoc
readJSON ReaderOptions
_ a
s =
  case forall a. FromJSON a => ByteString -> Either String a
eitherDecode' forall b c a. (b -> c) -> (a -> b) -> a -> c
. ByteString -> ByteString
BL.fromStrict forall b c a. (b -> c) -> (a -> b) -> a -> c
. Text -> ByteString
UTF8.fromText
                     forall b c a. (b -> c) -> (a -> b) -> a -> c
. Sources -> Text
sourcesToText forall b c a. (b -> c) -> (a -> b) -> a -> c
. forall a. ToSources a => a -> Sources
toSources forall a b. (a -> b) -> a -> b
$ a
s of
       Right Pandoc
doc -> forall (m :: * -> *) a. Monad m => a -> m a
return Pandoc
doc
       Left String
e    -> forall e (m :: * -> *) a. MonadError e m => e -> m a
throwError forall a b. (a -> b) -> a -> b
$ Text -> PandocError
PandocParseError (Text
"JSON parse error: "
                                                   forall a. Semigroup a => a -> a -> a
<> String -> Text
T.pack String
e)