0a768f1 · typeable/pandoc

git.iko.soy/typeable/pandoc

Added preliminary support for PDF creation via wkhtmltopdf.

To use this: pandoc -t html5 -o result.pdf (and add `--mathjax` if you have math.)

commit0a768f1cc568722973302554265efa414d097da3 authorJohn MacFarlane <jgm@berkeley.edu> dateMon 21 Dec 2015, 17:22 (11 years ago) parent32d2789

3 files changed+79−21

compare with master

unifiedsplit

README+6−5 pandoc.hs+15−13 src/Text/Pandoc/PDF.hs+58−3

README+6−5

@@ -24,7 +24,7 @@ markup], [Haddock markup], [OPML], [Emacs Org mode], [DocBook],

2424[FictionBook2], [Textile], [groff man] pages, [Emacs Org mode],

2525[AsciiDoc], [InDesign ICML], and [Slidy], [Slideous], [DZSlides],

2626[reveal.js] or [S5] HTML slide shows. It can also produce [PDF] output

27−on systems where LaTeX or ConTeXt is installed.

27+on systems where LaTeX, ConTeXt, or `wkhtmltopdf` is installed.

2828

2929Pandoc's enhanced version of Markdown includes syntax for [footnotes],

3030[tables], flexible [ordered lists], [definition lists], [fenced code blocks],

@@ -189,12 +189,13 @@ any header file. The [`natbib`], [`biblatex`], [`bibtex`], and [`biber`]

189189packages can optionally be used for [citation rendering]. These are

190190included with all recent versions of [TeX Live].

191191

192−Alternatively, pandoc can use ConTeXt to create a PDF.

192+Alternatively, pandoc can use ConTeXt or `wkhtmltopdf` to create a PDF.

193193To do this, specify an output file with a `.pdf` extension,

194−as before, but add `-t context` to the command line.

194+as before, but add `-t context` or `-t html5` to the command line.

195195

196−PDF output can be controlled using [variables for LaTeX]

197−or [variables for ConTeXt].

196+PDF output can be controlled using [variables for LaTeX] (if

197+LaTeX is used), [variables for ConTeXt] (if ConTeXt is used),

198+and `--css` (if `wkhtmltopdf` is used).

198199

199200[`amsfonts`]: https://ctan.org/pkg/amsfonts

200201[`amsmath`]: https://ctan.org/pkg/amsmath

pandoc.hs+15−13

@@ -1168,6 +1168,7 @@ convertWithOpts opts args = do

11681168 let laTeXOutput = "latex" `isPrefixOf` writerName' ||

11691169 "beamer" `isPrefixOf` writerName'

11701170 let conTeXtOutput = "context" `isPrefixOf` writerName'

1171+ let html5Output = "html5" `isPrefixOf` writerName'

11711172

11721173 let laTeXInput = "latex" `isPrefixOf` readerName' ||

11731174 "beamer" `isPrefixOf` readerName'

@@ -1367,27 +1368,28 @@ convertWithOpts opts args = do

13671368 IOByteStringWriter f -> f writerOptions doc' >>= writeBinary

13681369 PureStringWriter f

13691370 | pdfOutput -> do

1370− -- make sure writer is latex or beamer

1371− unless (laTeXOutput || conTeXtOutput) $

1371+ -- make sure writer is latex or beamer or context or html5

1372+ unless (laTeXOutput || conTeXtOutput || html5Output) $

13721373 err 47 $ "cannot produce pdf output with " ++ writerName' ++

13731374 " writer"

13741375

1375− let texprog = if conTeXtOutput

1376− then "context"

1377− else latexEngine

1378− -- check for latex program

1379− mbLatex <- findExecutable texprog

1380− when (isNothing mbLatex) $

1381− err 41 $ texprog ++ " not found. " ++

1382− texprog ++ " is needed for pdf output."

1383−

1384− res <- makePDF texprog f writerOptions doc'

1376+ let pdfprog = case () of

1377+ _ | conTeXtOutput -> "context"

1378+ _ | html5Output -> "wkhtmltopdf"

1379+ _ -> latexEngine

1380+ -- check for pdf creating program

1381+ mbPdfProg <- findExecutable pdfprog

1382+ when (isNothing mbPdfProg) $

1383+ err 41 $ pdfprog ++ " not found. " ++

1384+ pdfprog ++ " is needed for pdf output."

1385+

1386+ res <- makePDF pdfprog f writerOptions doc'

13851387 case res of

13861388 Right pdf -> writeBinary pdf

13871389 Left err' -> do

13881390 B.hPutStr stderr err'

13891391 B.hPut stderr $ B.pack [10]

1390− err 43 "Error producing PDF from TeX source"

1392+ err 43 "Error producing PDF"

13911393 | otherwise -> selfcontain (f writerOptions doc' ++

13921394 ['\n' | not standalone'])

13931395 >>= writerFn outputFile . handleEntities

src/Text/Pandoc/PDF.hs+58−3

@@ -38,6 +38,7 @@ import Text.Pandoc.Compat.Monoid ((<>))

3838import System.Exit (ExitCode (..))

3939import System.FilePath

4040import System.IO (stderr, stdout)

41+import System.IO.Temp (withTempFile)

4142import System.Directory

4243import Data.Digest.Pure.SHA (showDigest, sha1)

4344import System.Environment

@@ -49,7 +50,7 @@ import qualified Text.Pandoc.UTF8 as UTF8

4950import Text.Pandoc.Definition

5051import Text.Pandoc.Walk (walkM)

5152import Text.Pandoc.Shared (fetchItem', warn, withTempDir, inDirectory)

52−import Text.Pandoc.Options (WriterOptions(..))

53+import Text.Pandoc.Options (WriterOptions(..), HTMLMathMethod(..))

5354import Text.Pandoc.MIME (extensionFromMimeType, getMimeType)

5455import Text.Pandoc.Process (pipeProcess)

5556import qualified Data.ByteString.Lazy as BL

@@ -63,18 +64,29 @@ changePathSeparators :: FilePath -> FilePath

6364changePathSeparators = intercalate "/" . splitDirectories

6465#endif

6566

66−makePDF :: String -- ^ pdf creator (pdflatex, lualatex, xelatex)

67+makePDF :: String -- ^ pdf creator (pdflatex, lualatex,

68+ -- xelatex, context, wkhtmltopdf)

6769 -> (WriterOptions -> Pandoc -> String) -- ^ writer

6870 -> WriterOptions -- ^ options

6971 -> Pandoc -- ^ document

7072 -> IO (Either ByteString ByteString)

73+makePDF "wkhtmltopdf" writer opts doc =

74+ wkhtml2pdf (writerVerbose opts) args source

75+ where args = case writerHTMLMathMethod opts of

76+ -- with MathJax, wait til all math is rendered:

77+ MathJax _ -> ["--run-script", "MathJax.Hub.Register.StartupHook('End Typeset', function() { window.status = 'mathjax_loaded' });",

78+ "--window-status", "mathjax_loaded"]

79+ _ -> []

80+ source = writer opts doc

7181makePDF program writer opts doc = withTempDir "tex2pdf." $ \tmpdir -> do

7282 doc' <- handleImages opts tmpdir doc

7383 let source = writer opts doc'

7484 args = writerLaTeXArgs opts

7585 case program of

7686 "context" -> context2pdf (writerVerbose opts) tmpdir source

77− _ -> tex2pdf' (writerVerbose opts) args tmpdir program source

87+ _ | program `elem` ["pdflatex", "lualatex", "xelatex"]

88+ -> tex2pdf' (writerVerbose opts) args tmpdir program source

89+ _ -> return $ Left $ UTF8.fromStringLazy $ "Unknown program " ++ program

7890

7991handleImages :: WriterOptions

8092 -> FilePath -- ^ temp dir to store images

@@ -235,6 +247,49 @@ runTeXProgram verbose program args runNumber numRuns tmpDir source = do

235247 else return Nothing

236248 return (exit, out <> err, pdf)

237249

250+wkhtml2pdf :: Bool -- ^ Verbose output

251+ -> [String] -- ^ Args to wkhtmltopdf

252+ -> String -- ^ HTML5 source

253+ -> IO (Either ByteString ByteString)

254+wkhtml2pdf verbose args source = do

255+ file <- withTempFile "." "wkhtml2pdf.html" $ \fp _ -> return fp

256+ pdfFile <- withTempFile "." "wkhtml2pdf.pdf" $ \fp _ -> return fp

257+ UTF8.writeFile file source

258+ let programArgs = args ++ [file, pdfFile]

259+ env' <- getEnvironment

260+ when verbose $ do

261+ putStrLn "[makePDF] Command line:"

262+ putStrLn $ "wkhtmltopdf" ++ " " ++ unwords (map show programArgs)

263+ putStr "\n"

264+ putStrLn "[makePDF] Environment:"

265+ mapM_ print env'

266+ putStr "\n"

267+ putStrLn $ "[makePDF] Contents of " ++ file ++ ":"

268+ B.readFile file >>= B.putStr

269+ putStr "\n"

270+ (exit, out, err) <- pipeProcess (Just env') "wkhtmltopdf"

271+ programArgs BL.empty

272+ removeFile file

273+ when verbose $ do

274+ B.hPutStr stdout out

275+ B.hPutStr stderr err

276+ putStr "\n"

277+ pdfExists <- doesFileExist pdfFile

278+ mbPdf <- if pdfExists

279+ -- We read PDF as a strict bytestring to make sure that the

280+ -- temp directory is removed on Windows.

281+ -- See https://github.com/jgm/pandoc/issues/1192.

282+ then do

283+ res <- (Just . B.fromChunks . (:[])) `fmap` BS.readFile pdfFile

284+ removeFile pdfFile

285+ return res

286+ else return Nothing

287+ let log' = out <> err

288+ return $ case (exit, mbPdf) of

289+ (ExitFailure _, _) -> Left log'

290+ (ExitSuccess, Nothing) -> Left ""

291+ (ExitSuccess, Just pdf) -> Right pdf

292+

238293context2pdf :: Bool -- ^ Verbose output

239294 -> FilePath -- ^ temp directory for output

240295 -> String -- ^ ConTeXt source