aboutsummaryrefslogtreecommitdiff
path: root/src/Text/Pandoc/Readers/EPUB.hs
blob: 851d4771fd30b427b700bc32ff167dc01e476eab (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
{-# LANGUAGE
   ViewPatterns
 , StandaloneDeriving
 , TupleSections
 , FlexibleContexts  #-}

module Text.Pandoc.Readers.EPUB
  (readEPUB)
  where

import Text.XML.Light
import Text.Pandoc.Definition hiding (Attr)
import Text.Pandoc.Readers.HTML (readHtml)
import Text.Pandoc.Walk (walk, query)
import Text.Pandoc.Options ( ReaderOptions(..), Verbosity(..))
import Text.Pandoc.Extensions (enableExtension, Extension(Ext_raw_html))
import Text.Pandoc.Shared (escapeURI, collapseFilePath, addMetaField)
import Network.URI (unEscapeString)
import Text.Pandoc.MediaBag (MediaBag, insertMedia)
import Control.Monad.Except (throwError)
import Text.Pandoc.MIME (MimeType)
import qualified Text.Pandoc.Builder as B
import Codec.Archive.Zip ( Archive (..), toArchiveOrFail, fromEntry
                         , findEntryByPath, Entry)
import qualified Data.ByteString.Lazy as BL (ByteString)
import System.FilePath ( takeFileName, (</>), dropFileName, normalise
                       , dropFileName
                       , splitFileName )
import qualified Text.Pandoc.UTF8 as UTF8 (toStringLazy)
import Control.Monad (guard, liftM)
import Data.List (isPrefixOf, isInfixOf)
import Data.Maybe (mapMaybe, fromMaybe)
import qualified Data.Map as M (Map, lookup, fromList, elems)
import Data.Monoid ((<>))
import Control.DeepSeq (deepseq, NFData)
import Text.Pandoc.Error
import Text.Pandoc.Class (PandocMonad, report)
import qualified Text.Pandoc.Class as P

type Items = M.Map String (FilePath, MimeType)

readEPUB :: PandocMonad m => ReaderOptions -> BL.ByteString -> m Pandoc
readEPUB opts bytes = case toArchiveOrFail bytes of
  Right archive -> archiveToEPUB opts $ archive
  Left  _       -> throwError $ PandocParseError "Couldn't extract ePub file"

-- runEPUB :: Except PandocError a -> Either PandocError a
-- runEPUB = runExcept

-- Note that internal reference are aggresively normalised so that all ids
-- are of the form "filename#id"
--
archiveToEPUB :: (PandocMonad m) => ReaderOptions -> Archive -> m Pandoc
archiveToEPUB os archive = do
  -- root is path to folder with manifest file in
  (root, content) <- getManifest archive
  meta  <- parseMeta content
  (cover, items) <- parseManifest content
  -- No need to collapse here as the image path is from the manifest file
  let coverDoc = fromMaybe mempty (imageToPandoc <$> cover)
  spine <- parseSpine items content
  let escapedSpine = map (escapeURI . takeFileName . fst) spine
  Pandoc _ bs <-
      foldM' (\a b -> ((a <>) . walk (prependHash escapedSpine))
        `liftM` parseSpineElem root b) mempty spine
  let ast = coverDoc <> (Pandoc meta bs)
  P.setMediaBag $ fetchImages (M.elems items) root archive ast
  return ast
  where
    os' = os {readerExtensions = enableExtension Ext_raw_html (readerExtensions os)}
    parseSpineElem :: PandocMonad m => FilePath -> (FilePath, MimeType) -> m Pandoc
    parseSpineElem (normalise -> r) (normalise -> path, mime) = do
      report DEBUG ("parseSpineElem called with path " ++ show path)
      doc <- mimeToReader mime r path
      let docSpan = B.doc $ B.para $ B.spanWith (takeFileName path, [], []) mempty
      return $ docSpan <> doc
    mimeToReader :: PandocMonad m => MimeType -> FilePath -> FilePath -> m Pandoc
    mimeToReader "application/xhtml+xml" (unEscapeString -> root)
                                         (unEscapeString -> path) = do
      fname <- findEntryByPathE (root </> path) archive
      html <- readHtml os' . UTF8.toStringLazy $ fromEntry fname
      return $ fixInternalReferences path html
    mimeToReader s _ (unEscapeString -> path)
      | s `elem` imageMimes = return $ imageToPandoc path
      | otherwise = return $ mempty

-- paths should be absolute when this function is called
-- renameImages should do this
fetchImages :: [(FilePath, MimeType)]
            -> FilePath -- ^ Root
            -> Archive
            -> Pandoc
            -> MediaBag
fetchImages mimes root arc (query iq -> links) =
    foldr (uncurry3 insertMedia) mempty
      (mapMaybe getEntry links)
  where
    getEntry link =
        let abslink = normalise (root </> link) in
        (link , lookup link mimes, ) . fromEntry
          <$> findEntryByPath abslink arc

iq :: Inline -> [FilePath]
iq (Image _ _ (url, _)) = [url]
iq _ = []

-- Remove relative paths
renameImages :: FilePath -> Inline -> Inline
renameImages root img@(Image attr a (url, b))
  | "data:" `isPrefixOf` url = img
  | otherwise                = Image attr a (collapseFilePath (root </> url), b)
renameImages _ x = x

imageToPandoc :: FilePath -> Pandoc
imageToPandoc s = B.doc . B.para $ B.image s "" mempty

imageMimes :: [MimeType]
imageMimes = ["image/gif", "image/jpeg", "image/png"]

type CoverImage = FilePath

parseManifest :: (PandocMonad m) => Element -> m (Maybe CoverImage, Items)
parseManifest content = do
  manifest <- findElementE (dfName "manifest") content
  let items = findChildren (dfName "item") manifest
  r <- mapM parseItem items
  let cover = findAttr (emptyName "href") =<< filterChild findCover manifest
  return (cover, (M.fromList r))
  where
    findCover e = maybe False (isInfixOf "cover-image")
                  (findAttr (emptyName "properties") e)
    parseItem e = do
      uid <- findAttrE (emptyName "id") e
      href <- findAttrE (emptyName "href") e
      mime <- findAttrE (emptyName "media-type") e
      return (uid, (href, mime))

parseSpine :: PandocMonad m => Items -> Element -> m [(FilePath, MimeType)]
parseSpine is e = do
  spine <- findElementE (dfName "spine") e
  let itemRefs = findChildren (dfName "itemref") spine
  mapM (mkE "parseSpine" . (flip M.lookup is)) $ mapMaybe parseItemRef itemRefs
  where
    parseItemRef ref = do
      let linear = maybe True (== "yes") (findAttr (emptyName "linear") ref)
      guard linear
      findAttr (emptyName "idref") ref

parseMeta :: PandocMonad m => Element -> m Meta
parseMeta content = do
  meta <- findElementE (dfName "metadata") content
  let dcspace (QName _ (Just "http://purl.org/dc/elements/1.1/") (Just "dc")) = True
      dcspace _ = False
  let dcs = filterChildrenName dcspace meta
  let r = foldr parseMetaItem nullMeta dcs
  return r

-- http://www.idpf.org/epub/30/spec/epub30-publications.html#sec-metadata-elem
parseMetaItem :: Element -> Meta -> Meta
parseMetaItem e@(stripNamespace . elName -> field) meta =
  addMetaField (renameMeta field) (B.str $ strContent e) meta

renameMeta :: String -> String
renameMeta "creator" = "author"
renameMeta s = s

getManifest :: PandocMonad m => Archive -> m (String, Element)
getManifest archive = do
  metaEntry <- findEntryByPathE ("META-INF" </> "container.xml") archive
  docElem <- (parseXMLDocE . UTF8.toStringLazy . fromEntry) metaEntry
  let namespaces = mapMaybe attrToNSPair (elAttribs docElem)
  ns <- mkE "xmlns not in namespaces" (lookup "xmlns" namespaces)
  as <- liftM ((map attrToPair) . elAttribs)
    (findElementE (QName "rootfile" (Just ns) Nothing) docElem)
  manifestFile <- mkE "Root not found" (lookup "full-path" as)
  let rootdir = dropFileName manifestFile
  --mime <- lookup "media-type" as
  manifest <- findEntryByPathE manifestFile archive
  liftM ((,) rootdir) (parseXMLDocE . UTF8.toStringLazy . fromEntry $ manifest)

-- Fixup

fixInternalReferences :: FilePath -> Pandoc -> Pandoc
fixInternalReferences pathToFile =
  (walk $ renameImages root)
  . (walk $ fixBlockIRs filename)
  . (walk $ fixInlineIRs filename)
  where
    (root, escapeURI -> filename) = splitFileName pathToFile

fixInlineIRs :: String -> Inline -> Inline
fixInlineIRs s (Span as v) =
  Span (fixAttrs s as) v
fixInlineIRs s (Code as code) =
  Code (fixAttrs s as) code
fixInlineIRs s (Link as is ('#':url, tit)) =
  Link (fixAttrs s as) is (addHash s url, tit)
fixInlineIRs s (Link as is t) =
  Link (fixAttrs s as) is t
fixInlineIRs _ v = v

prependHash :: [String] -> Inline -> Inline
prependHash ps l@(Link attr is (url, tit))
  | or [s `isPrefixOf` url | s <- ps] =
    Link attr is ('#':url, tit)
  | otherwise = l
prependHash _ i = i

fixBlockIRs :: String -> Block -> Block
fixBlockIRs s (Div as b) =
  Div (fixAttrs s as) b
fixBlockIRs s (Header i as b) =
  Header i (fixAttrs s as) b
fixBlockIRs s (CodeBlock as code) =
  CodeBlock (fixAttrs s as) code
fixBlockIRs _ b = b

fixAttrs :: FilePath -> B.Attr -> B.Attr
fixAttrs s (ident, cs, kvs) = (addHash s ident, filter (not . null) cs, removeEPUBAttrs kvs)

addHash :: String -> String -> String
addHash _ "" = ""
addHash s ident = takeFileName s ++ "#" ++ ident

removeEPUBAttrs :: [(String, String)] -> [(String, String)]
removeEPUBAttrs kvs = filter (not . isEPUBAttr) kvs

isEPUBAttr :: (String, String) -> Bool
isEPUBAttr (k, _) = "epub:" `isPrefixOf` k

-- Library

-- Strict version of foldM
foldM' :: (Monad m, NFData a) => (a -> b -> m a) -> a -> [b] -> m a
foldM' _ z [] = return z
foldM' f z (x:xs) = do
  z' <- f z x
  z' `deepseq` foldM' f z' xs

uncurry3 :: (a -> b -> c -> d) -> (a, b, c) -> d
uncurry3 f (a, b, c) = f a b c

-- Utility

stripNamespace :: QName -> String
stripNamespace (QName v _ _) = v

attrToNSPair :: Attr -> Maybe (String, String)
attrToNSPair (Attr (QName "xmlns" _ _) val) = Just ("xmlns", val)
attrToNSPair _ = Nothing

attrToPair :: Attr -> (String, String)
attrToPair (Attr (QName name _ _) val) = (name, val)

defaultNameSpace :: Maybe String
defaultNameSpace = Just "http://www.idpf.org/2007/opf"

dfName :: String -> QName
dfName s = QName s defaultNameSpace Nothing

emptyName :: String -> QName
emptyName s = QName s Nothing Nothing

-- Convert Maybe interface to Either

findAttrE :: PandocMonad m => QName -> Element -> m String
findAttrE q e = mkE "findAttr" $ findAttr q e

findEntryByPathE :: PandocMonad m => FilePath -> Archive -> m Entry
findEntryByPathE (normalise -> path) a =
  mkE ("No entry on path: " ++ path) $ findEntryByPath path a

parseXMLDocE :: PandocMonad m => String -> m Element
parseXMLDocE doc = mkE "Unable to parse XML doc" $ parseXMLDoc doc

findElementE :: PandocMonad m => QName -> Element -> m Element
findElementE e x = mkE ("Unable to find element: " ++ show e) $ findElement e x

mkE :: PandocMonad m => String -> Maybe a -> m a
mkE s = maybe (throwError . PandocParseError $ s) return