d072ad4 · typeable/pandoc

git.iko.soy/typeable/pandoc

Added 'odt' output option to pandoc:

Not a writer, but a module that inserts the output of the OpenDocument writer into an ODT archive. This replaces markdown2odt. + Added odt output option to Main.hs. + Added default for .odt output file. + Changed defaults so that .xml and .sgml aren't automatically DocBook. + Added odt writer to Text.Pandoc exports. + Added Text.Pandoc.ODT and included in pandoc.cabal. + Added reference.odt as data-file in pandoc.cabal. + Handle picture links in OpenDocument files using xml library. + Removed markdown2odt and references from Makefile, README, man. git-svn-id: https://pandoc.googlecode.com/svn/trunk@1345 788f1e2b-df1e-0410-8736-df70ead52e1b

commitd072ad4b66b5cdf5a8f811ae1f78460e34270d58 authorfiddlosopher <fiddlosopher@788f1e2b-df1e-0410-8736-df70ead52e1b> dateThu 31 Jul 2008, 23:16 (18 years ago) parent504a61a

8 files changed+210−144

compare with master

unifiedsplit

Main.hs+31−15 Makefile+1−6 README+3−23 Text/Pandoc/ODT.hs+141 Text/Pandoc/Shared.hs+25−1 man/man1/markdown2odt.1.md−52 pandoc.cabal+9−6 wrappers/markdown2odt.in−41

Main.hs+31−15

@@ -31,11 +31,12 @@ writers.

3131module Main where

3232import Text.Pandoc

3333import Text.Pandoc.UTF8

34+import Text.Pandoc.ODT

3435import Text.Pandoc.Shared ( joinWithSep, HTMLMathMethod (..) )

3536import Text.Pandoc.Highlighting ( languages )

3637import System.Environment ( getArgs, getProgName, getEnvironment )

3738import System.Exit ( exitWith, ExitCode (..) )

38−import System.FilePath ( takeExtension )

39+import System.FilePath ( takeExtension, takeDirectory )

3940import System.Console.GetOpt

4041import System.IO

4142import Data.Maybe ( fromMaybe )

@@ -82,6 +83,7 @@ writers = [("native" , (writeDoc, ""))

8283 ,("s5" , (writeS5String, defaultS5Header))

8384 ,("docbook" , (writeDocbook, defaultDocbookHeader))

8485 ,("opendocument" , (writeOpenDocument, defaultOpenDocumentHeader))

86+ ,("odt" , (writeOpenDocument, defaultOpenDocumentHeader))

8587 ,("latex" , (writeLaTeX, defaultLaTeXHeader))

8688 ,("context" , (writeConTeXt, defaultConTeXtHeader))

8789 ,("texinfo" , (writeTexinfo, ""))

@@ -92,6 +94,10 @@ writers = [("native" , (writeDoc, ""))

9294 ,("rtf" , (writeRTF, defaultRTFHeader))

9395 ]

9496

97+isNonTextOutput :: String -> Bool

98+isNonTextOutput "odt" = True

99+isNonTextOutput _ = False

100+

95101-- | Writer for Pandoc native format.

96102writeDoc :: WriterOptions -> Pandoc -> String

97103writeDoc _ = prettyPandoc

@@ -392,8 +398,7 @@ defaultWriterName x =

392398 ".texi" -> "texinfo"

393399 ".texinfo" -> "texinfo"

394400 ".db" -> "docbook"

395− ".xml" -> "docbook"

396− ".sgml" -> "docbook"

401+ ".odt" -> "odt"

397402 ['.',y] | y `elem` ['1'..'9'] -> "man"

398403 _ -> "html"

399404

@@ -478,10 +483,6 @@ main = do

478483 Just (w,h) -> return (w, h)

479484 Nothing -> error ("Unknown writer: " ++ writerName')

480485

481− output <- if (outputFile == "-")

482− then return stdout

483− else openFile outputFile WriteMode

484−

485486 environment <- getEnvironment

486487 let columns = case lookup "COLUMNS" environment of

487488 Just cols -> read cols

@@ -501,11 +502,13 @@ main = do

501502 tabFilter spsToNextStop (x:xs) =

502503 x:(tabFilter (spsToNextStop - 1) xs)

503504

505+ let standalone' = (standalone && not strict) || writerName' == "odt"

506+

504507 let startParserState =

505508 defaultParserState { stateParseRaw = parseRaw,

506509 stateTabStop = tabStop,

507510 stateSanitizeHTML = sanitize,

508− stateStandalone = standalone && (not strict),

511+ stateStandalone = standalone',

509512 stateSmart = smart || writerName' `elem`

510513 ["latex", "context"],

511514 stateColumns = columns,

@@ -519,16 +522,15 @@ main = do

519522 let header = (if (customHeader == "DEFAULT")

520523 then defaultHeader

521524 else customHeader) ++ csslink ++ includeHeader

522− let writerOptions = WriterOptions { writerStandalone = standalone &&

523− (not strict),

525+ let writerOptions = WriterOptions { writerStandalone = standalone',

524526 writerHeader = header,

525527 writerTitlePrefix = titlePrefix,

526528 writerTabStop = tabStop,

527529 writerTableOfContents = toc &&

528530 (not strict) &&

529− writerName/="s5",

531+ writerName' /= "s5",

530532 writerHTMLMathMethod = mathMethod,

531− writerS5 = (writerName=="s5"),

533+ writerS5 = (writerName' == "s5"),

532534 writerIgnoreNotes = False,

533535 writerIncremental = incremental,

534536 writerNumberSections = numberSections,

@@ -538,11 +540,25 @@ main = do

538540 writerReferenceLinks = referenceLinks,

539541 writerWrapText = wrap }

540542

541− (readSources sources) >>= (hPutStrLn output . toUTF8 .

543+ let writeOutput = if writerName' == "odt"

544+ then if outputFile == "-"

545+ then \_ -> do

546+ hPutStrLn stderr ("Error: Cannot write " ++ writerName ++

547+ " output to stdout.\n" ++

548+ "Specify an output file using the -o option.")

549+ exitWith $ ExitFailure 5

550+ else let sourceDirRelative = if null sources

551+ then ""

552+ else takeDirectory (head sources)

553+ in saveOpenDocumentAsODT outputFile sourceDirRelative

554+ else if outputFile == "-"

555+ then putStrLn

556+ else writeFile outputFile . (++ "\n")

557+

558+ (readSources sources) >>= writeOutput . toUTF8 .

542559 (writer writerOptions) .

543560 (reader startParserState) . tabFilter tabStop .

544− fromUTF8 . (joinWithSep "\n")) >>

545− hClose output

561+ fromUTF8 . (joinWithSep "\n")

546562

547563 where

548564 readSources [] = mapM readSource ["-"]

Makefile+1−6

@@ -24,7 +24,7 @@ EXECSBASE := $(shell sed -ne 's/^[Ee]xecutable:\{0,1\}[[:space:]]*//p' $(CABAL))

2424#-------------------------------------------------------------------------------

2525# Install targets

2626#-------------------------------------------------------------------------------

27−WRAPPERS := html2markdown markdown2pdf hsmarkdown markdown2odt

27+WRAPPERS := html2markdown markdown2pdf hsmarkdown

2828# Add .exe extensions if we're running Windows/Cygwin.

2929EXTENSION := $(shell uname | tr '[:upper:]' '[:lower:]' | \

3030 sed -ne 's/^cygwin.*$$/\.exe/p')

@@ -114,11 +114,6 @@ $(ODTREF): $(addprefix $(ODTSTYLES)/, layout-cache meta.xml styles.xml content.x

114114 cd $(ODTSTYLES) ; \

115115 zip -9 -r $(notdir $@) * -x $(notdir $@)

116116

117−ODTREFSH=$(SRCDIR)/wrappers/odtref.sh

118−cleanup_files+=$(ODTREFSH)

119−$(ODTREFSH): $(ODTREF)

120− echo "REFERENCEODT='$(PREFIX)/share/$(DATADIR)/$(notdir $(ODTREF))'" > $@

121−

122117.PHONY: wrappers

123118wrappers: $(WRAPPERS)

124119cleanup_files+=$(WRAPPERS)

README+3−23

@@ -119,7 +119,7 @@ then convert the output back to the local encoding.

119119Shell scripts

120120=============

121121

122−Four shell scripts, `markdown2pdf`, `markdown2odt`, `html2markdown`, and

122+Three shell scripts, `markdown2pdf`, `html2markdown`, and

123123`hsmarkdown`, are included in the standard Pandoc installation. (They

124124are not included in the Windows binary package, as they require a POSIX

125125shell, but they may be used in Windows under Cygwin.)

@@ -149,27 +149,7 @@ shell, but they may be used in Windows under Cygwin.)

149149 included in your LaTeX distribution, you can get them from

150150 [CTAN].

151151

152−2. `markdown2odt` produces an ODT file from markdown-formatted

153− text, using `pandoc` and `pdflatex`. (ODT is "OpenDocument

154− Text," the default format for the OpenOffice.org Writer.)

155− The default behavior of `markdown2odt` is to create a file with the

156− same base name as the first argument and the extension `odt`; thus,

157− for example,

158−

159− markdown2odt sample.txt endnotes.txt

160−

161− will produce `sample.odt`. (If `sample.odt` exists already,

162− it will be backed up before being overwritten.) An output file

163− name can be specified explicitly using the `-o` option:

164−

165− markdown2odt -o book.odt chap1 chap2

166−

167− If no input file is specified, input will be taken from STDIN.

168− All of `pandoc`'s options will work with `markdown2odt` as well.

169−

170− `markdown2odt` requires `zip`, which must be in the path.

171−

172−3. `html2markdown` grabs a web page from a file or URL and converts

152+2. `html2markdown` grabs a web page from a file or URL and converts

173153 it to markdown-formatted text, using `tidy` and `pandoc`.

174154

175155 All of `pandoc`'s options will work with `html2markdown` as well.

@@ -198,7 +178,7 @@ shell, but they may be used in Windows under Cygwin.)

198178 It uses [`iconv`] for character encoding conversions; if `iconv`

199179 is absent, it will still work, but it will treat everything as UTF-8.

200180

201−4. `hsmarkdown` is designed to be used as a drop-in replacement for

181+3. `hsmarkdown` is designed to be used as a drop-in replacement for

202182 `Markdown.pl`. It forces `pandoc` to convert from markdown to

203183 HTML, and to use the `--strict` flag for maximal compliance with

204184 official markdown syntax. (All of Pandoc's syntax extensions and

Text/Pandoc/ODT.hsnew file+141

@@ -0,0 +1,141 @@

1+{-

4+This program is free software; you can redistribute it and/or modify

5+it under the terms of the GNU General Public License as published by

6+the Free Software Foundation; either version 2 of the License, or

7+(at your option) any later version.

9+This program is distributed in the hope that it will be useful,

10+but WITHOUT ANY WARRANTY; without even the implied warranty of

11+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the

12+GNU General Public License for more details.

13+

14+You should have received a copy of the GNU General Public License

15+along with this program; if not, write to the Free Software

16+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA

17+-}

18+

19+{- |

20+ Module : Text.Pandoc.ODT

22+ License : GNU GPL, version 2 or above

23+

24+ Maintainer : John MacFarlane <jgm@berkeley.edu>

25+ Stability : alpha

26+ Portability : portable

27+

28+Functions for producing an ODT file from OpenDocument XML.

29+-}

30+module Text.Pandoc.ODT ( saveOpenDocumentAsODT ) where

31+import Data.Maybe ( fromJust )

32+import Data.List ( partition, intersperse )

33+import Prelude hiding ( writeFile, readFile )

34+import System.IO.UTF8

35+import System.IO ( stderr )

36+import System.Directory

37+import System.FilePath ( (</>), takeDirectory, takeFileName, splitDirectories )

38+import System.Process ( runCommand, waitForProcess )

39+import System.Exit

40+import Text.XML.Light

41+import Text.XML.Light.Cursor

42+import Text.Pandoc.Shared ( withTempDir )

43+import Network.URI ( isURI )

44+import Paths_pandoc

45+

46+-- | Produce an ODT file from OpenDocument XML.

47+saveOpenDocumentAsODT :: FilePath -- ^ Pathname of ODT file to be produced.

48+ -> FilePath -- ^ Relative directory of source file.

49+ -> String -- ^ OpenDocument XML contents.

50+ -> IO ()

51+saveOpenDocumentAsODT destinationODTPath sourceDirRelative xml = do

52+ let zipCmd = "zip"

53+ -- check for zip in path:

54+ maybeZipPath <- findExecutable zipCmd

55+ let zipPath = case maybeZipPath of

56+ Just p -> p

57+ Nothing -> error $ "The '" ++ zipCmd ++ "' command, which is needed to build an ODT file, " ++

58+ "was not found.\n" ++

59+ "It can be obtained from http://www.info-zip.org/Zip.html\n" ++

60+ "Debian (and Debian-based) linux: apt-get install zip\n" ++

61+ "Windows: See http://gnuwin32.sourceforge.net/packages/zip.htm"

62+ referenceODTPath <- getDataFileName "reference.odt"

63+ withTempDir "pandoc-odt" $ \tempDir -> do

64+ let tempODT = tempDir </> "reference.odt"

65+ copyFile referenceODTPath tempODT

66+ createDirectory $ tempDir </> "Pictures"

67+ xml' <- handlePictures tempODT sourceDirRelative xml

68+ writeFile (tempDir </> "content.xml") xml'

69+ oldDir <- getCurrentDirectory

70+ setCurrentDirectory tempDir

71+ let zipCmdLine = zipPath ++ " -9 -q -r " ++ tempODT ++ " " ++ "content.xml Pictures"

72+ ec <- runCommand zipCmdLine >>= waitForProcess -- this requires compilation with -threaded

73+ setCurrentDirectory oldDir

74+ case ec of

75+ ExitSuccess -> copyFile tempODT destinationODTPath

76+ _ -> error "Error creating ODT." >> exitWith ec

77+

78+

79+-- | Find <draw:image ... /> elements and copy the file (xlink:href attribute) into Pictures/ in

80+-- the zip file. If filename is a URL, attempt to download it. Modify xlink:href attributes

81+-- to point to the new locations in Pictures/. Return modified XML.

82+handlePictures :: FilePath -- ^ Path of ODT file in temp directory

83+ -> FilePath -- ^ Directory (relative) containing source file

84+ -> String -- ^ OpenDocument XML string

85+ -> IO String -- ^ Modified XML

86+handlePictures tempODT sourceDirRelative xml = do

87+ let parsed = case parseXMLDoc xml of

88+ Nothing -> error "Could not parse OpenDocument XML."

89+ Just x -> x

90+ let cursor = case (fromForest $ elContent parsed) of

91+ Nothing -> error "ODT appears empty"

92+ Just x -> x

93+ cursor' <- scanPictures tempODT sourceDirRelative cursor

94+ let modified = parsed { elContent = toForest $ root cursor' }

95+ putStrLn $ showTopElement modified

96+ return $ showTopElement modified

97+

98+scanPictures :: FilePath -> FilePath -> Cursor -> IO Cursor

99+scanPictures tempODT sourceDirRelative cursor = do

100+ cursor' <- handleTree tempODT sourceDirRelative cursor

101+ case right cursor' of

102+ Just n -> scanPictures tempODT sourceDirRelative n

103+ Nothing -> return cursor'

104+

105+handleTree :: FilePath -> FilePath -> Cursor -> IO Cursor

106+handleTree tempODT sourceDirRelative cursor = do

107+ case firstChild cursor of

108+ Nothing -> modifyContentM (handleContent tempODT sourceDirRelative) cursor

109+ Just n -> scanPictures tempODT sourceDirRelative n >>= return . fromJust . parent

110+

111+-- | If content is an image link, handle it appropriately.

112+-- Otherwise, handle children if any.

113+handleContent :: FilePath -> FilePath -> Content -> IO Content

114+handleContent tempODT sourceDirRelative content@(Elem el) = do

115+ if qName (elName el) == "image"

116+ then do

117+ let (hrefs, rest) = partition (\a -> qName (attrKey a) == "href") $ elAttribs el

118+ let href = case hrefs of

119+ [] -> error $ "No href found in " ++ show el

120+ [x] -> x

121+ _ -> error $ "Multiple hrefs found in " ++ show el

122+ if isURI $ attrVal href

123+ then return content

124+ else do -- treat as filename

125+ let oldLoc = sourceDirRelative </> attrVal href

126+ fileExists <- doesFileExist oldLoc

127+ if fileExists

128+ then do

129+ let pref = take 230 $ concat $ intersperse "_" $

130+ splitDirectories $ takeDirectory $ attrVal href

131+ let newLoc = "Pictures" </> pref ++ "_" ++ (takeFileName $ attrVal href)

132+ let tempDir = takeDirectory tempODT

133+ copyFile oldLoc $ tempDir </> newLoc

134+ let newAttrs = (href { attrVal = newLoc }) : rest

135+ return $ Elem (el { elAttribs = newAttrs })

136+ else do

137+ hPutStrLn stderr $ "Warning: Unable to find image at " ++ oldLoc ++ " - ignoring."

138+ return content

139+ else return content

140+

141+handleContent _ _ c = return c -- not Element

Text/Pandoc/Shared.hs+25−1

@@ -98,7 +98,9 @@ module Text.Pandoc.Shared (

9898 -- * Writer options

9999 HTMLMathMethod (..),

100100 WriterOptions (..),

101− defaultWriterOptions

101+ defaultWriterOptions,

102+ -- * File handling

103+ withTempDir

102104 ) where

103105

104106import Text.Pandoc.Definition

@@ -110,6 +112,9 @@ import Data.Char ( toLower, toUpper, ord, isLower, isUpper )

110112import Data.List ( find, isPrefixOf )

111113import Control.Monad ( join )

112114import Network.URI ( parseURI, URI (..), isAllowedInURI )

115+import System.FilePath ( (</>), (<.>) )

116+import System.IO.Error ( catch, ioError, isAlreadyExistsError )

117+import System.Directory

113118

114119--

115120-- List processing

@@ -900,3 +905,22 @@ defaultWriterOptions =

900905 , writerReferenceLinks = False

901906 , writerWrapText = True

902907 }

908+

909+-- | Perform a function in a temporary directory and clean up.

910+withTempDir :: FilePath -> (FilePath -> IO a) -> IO a

911+withTempDir baseName func = do

912+ tempDir <- createTempDir 0 baseName

913+ result <- catch (func tempDir) $ \e -> removeDirectoryRecursive tempDir >> ioError e

914+ removeDirectoryRecursive tempDir

915+ return result

916+

917+-- | Create a temporary directory with a unique name.

918+createTempDir :: Integer -> FilePath -> IO FilePath

919+createTempDir num baseName = do

920+ sysTempDir <- getTemporaryDirectory

921+ let dirName = sysTempDir </> baseName <.> show num

922+ catch (createDirectory dirName >> return dirName) $

923+ \e -> if isAlreadyExistsError e

924+ then createTempDir (num + 1) baseName

925+ else ioError e

926+

man/man1/markdown2odt.1.mddeleted−52

@@ -1,52 +0,0 @@

1−% MARKDOWN2ODT(1) Pandoc User Manuals

2−% John MacFarlane and Recai Oktas

3−% March 14, 2008

4−

5−# NAME

6−

7−markdown2odt - converts markdown-formatted text to ODT

8−

9−# SYNOPSIS

10−

11−markdown2odt [*options*] [*input-file*]...

12−

13−# DESCRIPTION

14−

15−`markdown2odt` converts *input-file* (or text from standard

16−input) from markdown-formatted plain text to ODT (OpenDocument

17−Text) format. If no output filename is specified (using the `-o`

18−option), the name of the output file is derived from the input file;

19−thus, for example, if the input file is *hello.txt*, the output file

20−will be *hello.odt*. If the input is read from STDIN and no output

21−filename is specified, the output file will be named *stdin.odt*. If

22−multiple input files are specified, they will be concatenated before

23−conversion, and the name of the output file will be derived from the

24−first input file.

25−

26−Input is assumed to be in the UTF-8 character encoding. If your

27−local character encoding is not UTF-8, you should pipe input

28−through `iconv`:

29−

30− iconv -t utf-8 input.txt | markdown2odt

31−

32−# OPTIONS

33−

34−`markdown2odt` is a wrapper around `pandoc`, so all of

35−`pandoc`'s options can be used with `markdown2odt` as well.

36−See `pandoc`(1) for a complete list.

37−The following options are most relevant:

38−

39−-o *FILE*, \--output=*FILE*

40−: Write output to *FILE*.

41−

42−\--strict

43−: Use strict markdown syntax, with no extensions or variants.

44−

45−-S, \--smart

46−: Use smart quotes, dashes, and ellipses. (This option is significant

47− only when the input format is `markdown`. It is selected automatically

48− when the output format is `latex` or `context`.)

49−

50−# SEE ALSO

51−

52−`pandoc`(1)

pandoc.cabal+9−6

@@ -57,6 +57,7 @@ Extra-Source-Files: README, INSTALL, COPYRIGHT, COPYING,

5757Extra-Tmp-Files: Text/Pandoc/ASCIIMathML.hs,

5858 Text/Pandoc/DefaultHeaders.hs,

5959 Text/Pandoc/Writers/S5.hs

60+Data-Files: odt-styles/reference.odt

6061Flag splitBase

6162 Description: Choose the new, smaller, split-up base package.

6263 Default: True

@@ -72,7 +73,8 @@ Library

7273 if flag(highlighting)

7374 Build-depends: highlighting-kate

7475 cpp-options: -DHIGHLIGHTING

75− Build-Depends: parsec < 3, xhtml, mtl, network, filepath

76+ Build-Depends: parsec < 3, xhtml, mtl, network, filepath, process, directory, xml,

77+ bytestring, binary

7678 Hs-Source-Dirs: .

7779 Exposed-Modules: Text.Pandoc,

7880 Text.Pandoc.Blocks,

@@ -80,6 +82,7 @@ Library

8082 Text.Pandoc.CharacterReferences,

8183 Text.Pandoc.Shared,

8284 Text.Pandoc.UTF8,

85+ Text.Pandoc.ODT,

8386 Text.Pandoc.ASCIIMathML,

8487 Text.Pandoc.DefaultHeaders,

8588 Text.Pandoc.Highlighting,

@@ -101,11 +104,11 @@ Library

101104 Text.Pandoc.Writers.RTF,

102105 Text.Pandoc.Writers.S5

103106 Other-Modules: Text.Pandoc.XML

104− Ghc-Options: -O2 -Wall

107+ Ghc-Options: -O2 -Wall -threaded

105108 Ghc-Prof-Options: -auto-all

106109

107110Executable pandoc

108− Hs-Source-Dirs: .

109− Main-Is: Main.hs

110− Ghc-Options: -O2 -Wall

111− Ghc-Prof-Options: -auto-all

111+ Hs-Source-Dirs: .

112+ Main-Is: Main.hs

113+ Ghc-Options: -O2 -Wall -threaded

114+ Ghc-Prof-Options: -auto-all

wrappers/markdown2odt.indeleted−41

@@ -1,41 +0,0 @@

1−#!/bin/sh -e

2−

3−REQUIRED="zip"

4−SYNOPSIS="converts markdown-formatted text to ODT."

5−

6−### odtref.sh

7−

8−### common.sh

9−

10−### tempdir.sh

11−

12−if [ "$OUTPUT" = "-" ]; then

13− firstinfile="$(echo $ARGS | sed -ne '1p')"

14− firstinfilebase="${firstinfile%.*}"

15− destname="${firstinfilebase:-stdin}.odt"

16−else

17− destname="$OUTPUT"

18−fi

19−

20−(

21− cp $REFERENCEODT $THIS_TEMPDIR/new.odt

22− pandoc -s -r markdown -w opendocument "$@" -o $THIS_TEMPDIR/content.xml

23− zip -9 -j $THIS_TEMPDIR/new.odt $THIS_TEMPDIR/content.xml

24−) || exit $?

25−

26−is_target_exists=

27−if [ -f "$destname" ]; then

28− is_target_exists=1

29− mv "$destname" "$destname~"

30−fi

31−

32−mv -f $THIS_TEMPDIR/new.odt "$destname"

33−

34−errn "Created $destname"

35−[ -z "$is_target_exists" ] || {

36− errn " (previous file has been backed up as $destname~)"

37−}

38−

39−err .

40−

41−