2677e84 · typeable/pandoc

git.iko.soy/typeable/pandoc

Revert "Revert "Added `--filter` option.""

This reverts commit 2e5edbb27837372f658b1abbe05371be57415847.

commit2677e8466311e2becdd7f65f1f6c23d559db14aa authorJohn MacFarlane <jgm@berkeley.edu> dateThu 8 Aug 2013, 11:09 (13 years ago) parent9aa9d5c

3 files changed+30−4

compare with master

unifiedsplit

README+8 pandoc.cabal+1 pandoc.hs+21−4

README+8

@@ -259,6 +259,14 @@ Reader options

259259 require different kinds of images. Currently this option only affects

260260 the markdown and LaTeX readers.

261261

262+`--filter=`*PATH*

263+: Specify an executable to be used as a filter transforming the

264+ Pandoc AST after the input is parsed and before the output is

265+ written. The executable should read JSON from stdin and write

266+ JSON to stdout. The JSON must be formatted like pandoc's own

267+ JSON input and output. Filters may be most easily created in Haskell,

268+ using the utility function `toJsonFilter` from `Text.Pandoc`.

269+

262270`--normalize`

263271: Normalize the document after reading: merge adjacent

264272 `Str` or `Emph` elements, for example, and remove repeated `Space`s.

pandoc.cabal+1

@@ -352,6 +352,7 @@ Executable pandoc

352352 extensible-exceptions >= 0.1 && < 0.2,

353353 highlighting-kate >= 0.5.5 && < 0.6,

354354 HTTP >= 4000.0.5 && < 4000.3,

355+ process >= 1 && < 1.2,

355356 citeproc-hs >= 0.3.7 && < 0.4

356357 Ghc-Options: -rtsopts -with-rtsopts=-K16m -Wall -fno-warn-unused-do-bind

357358 Ghc-Prof-Options: -auto-all -caf-all -rtsopts -with-rtsopts=-K16m

pandoc.hs+21−4

@@ -43,6 +43,7 @@ import System.Environment ( getArgs, getProgName )

4343import System.Exit ( exitWith, ExitCode (..) )

4444import System.FilePath

4545import System.Console.GetOpt

46+import System.Process (readProcess)

4647import Data.Char ( toLower )

4748import Data.List ( intercalate, isPrefixOf, sort )

4849import System.Directory ( getAppUserDataDirectory, doesFileExist, findExecutable )

@@ -53,6 +54,7 @@ import Control.Exception.Extensible ( throwIO )

5354import qualified Text.Pandoc.UTF8 as UTF8

5455import qualified Text.CSL as CSL

5556import Control.Monad (when, unless, liftM)

57+import Data.Foldable (foldrM)

5658import Network.HTTP (simpleHTTP, mkRequest, getResponseBody, RequestMethod(..))

5759import Network.URI (parseURI, isURI, URI(..))

5860import qualified Data.ByteString.Lazy as B

@@ -86,6 +88,12 @@ wrapWords indent c = wrap' (c - indent) (c - indent)

8688isTextFormat :: String -> Bool

8789isTextFormat s = takeWhile (`notElem` "+-") s `notElem` ["odt","docx","epub","epub3"]

8890

91+externalFilter :: FilePath -> Pandoc -> IO Pandoc

92+externalFilter f d = E.catch

93+ (readJSON def `fmap` readProcess f [] (writeJSON def d))

94+ (\e -> let _ = (e :: E.SomeException)

95+ in err 83 $ "Error running filter `" ++ f ++ "'")

96+

8997-- | Data structure for command line options.

9098data Opt = Opt

9199 { optTabStop :: Int -- ^ Number of spaces per tab

@@ -272,6 +280,13 @@ options =

272280 "STRING")

273281 "" -- "Classes (whitespace- or comma-separated) to use for indented code-blocks"

274282

283+ , Option "" ["filter"]

284+ (ReqArg

285+ (\arg opt -> return opt { optPlugins = externalFilter arg :

286+ optPlugins opt })

287+ "PROGRAM")

288+ "" -- "External JSON filter"

289+

275290 , Option "" ["normalize"]

276291 (NoArg

277292 (\opt -> return opt { optTransforms =

@@ -876,6 +891,7 @@ main = do

876891 , optReferenceLinks = referenceLinks

877892 , optWrapText = wrap

878893 , optColumns = columns

894+ , optPlugins = plugins

879895 , optEmailObfuscation = obfuscationMethod

880896 , optIdentifierPrefix = idPrefix

881897 , optIndentedCodeClasses = codeBlockClasses

@@ -1099,6 +1115,7 @@ main = do

10991115 reader readerOpts

11001116

11011117 let doc0 = foldr ($) doc transforms

1118+ doc1 <- foldrM ($) doc0 plugins

11021119

11031120 let writeBinary :: B.ByteString -> IO ()

11041121 writeBinary = B.writeFile (UTF8.encodePath outputFile)

@@ -1109,15 +1126,15 @@ main = do

11091126

11101127 case getWriter writerName' of

11111128 Left e -> err 9 e

1112− Right (IOStringWriter f) -> f writerOptions doc0 >>= writerFn outputFile

1113− Right (IOByteStringWriter f) -> f writerOptions doc0 >>= writeBinary

1129+ Right (IOStringWriter f) -> f writerOptions doc1 >>= writerFn outputFile

1130+ Right (IOByteStringWriter f) -> f writerOptions doc1 >>= writeBinary

11141131 Right (PureStringWriter f)

11151132 | pdfOutput -> do

1116− res <- makePDF latexEngine f writerOptions doc0

1133+ res <- makePDF latexEngine f writerOptions doc1

11171134 case res of

11181135 Right pdf -> writeBinary pdf

11191136 Left err' -> err 43 $ UTF8.toStringLazy err'

1120− | otherwise -> selfcontain (f writerOptions doc0 ++

1137+ | otherwise -> selfcontain (f writerOptions doc1 ++

11211138 ['\n' | not standalone'])

11221139 >>= writerFn outputFile . handleEntities

11231140 where htmlFormat = writerName' `elem`