a26b3a2 · typeable/pandoc

git.iko.soy/typeable/pandoc

Org reader: Add partial support for `#+EXCLUDE_TAGS` option. (#4950)

Closes #4284. Headers with the corresponding tags should not appear in the output. If one or more of the specified tags contains a non-tag character like `+`, Org-mode will not treat that as a valid tag, but will nonetheless continue scanning for valid tags. That behavior is not replicated in this patch; entering `cat+dog` as one of the entries in `#+EXCLUDE_TAGS` and running the file through Pandoc will cause the parser to fail and result in the only excluded tag being the default, `noexport`.

commita26b3a2d6af8614e13299bbf477e28c5932ef680 authorBrian Leung <bkleung89@gmail.com> dateFri 5 Oct 2018, 14:28 (8 years ago) parente4ca51c

5 files changed+61−7

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Org/DocumentTree.hs+7−7 src/Text/Pandoc/Readers/Org/Meta.hs+11 src/Text/Pandoc/Readers/Org/ParserState.hs+6 src/Text/Pandoc/Readers/Org/Parsing.hs+8 test/command/4284.md+29

src/Text/Pandoc/Readers/Org/DocumentTree.hs+7−7

@@ -43,6 +43,7 @@ import Text.Pandoc.Readers.Org.BlockStarts

4343import Text.Pandoc.Readers.Org.ParserState

4444import Text.Pandoc.Readers.Org.Parsing

4545

46+import qualified Data.Set as Set

4647import qualified Text.Pandoc.Builder as B

4748

4849--

@@ -73,9 +74,6 @@ documentTree blocks inline = do

7374 , headlineChildren = headlines'

7475 }

7576

76−newtype Tag = Tag { fromTag :: String }

77− deriving (Show, Eq)

78−

7977-- | Create a tag containing the given string.

8078toTag :: String -> Tag

8179toTag = Tag

@@ -153,7 +151,7 @@ headline blocks inline lvl = try $ do

153151

154152 headerTags :: Monad m => OrgParser m [Tag]

155153 headerTags = try $

156− let tag = many1 (alphaNum <|> oneOf "@%#_") <* char ':'

154+ let tag = orgTagWord <* char ':'

157155 in map toTag <$> (skipSpaces *> char ':' *> many1 tag <* skipSpaces)

158156

159157-- | Convert an Org mode headline (i.e. a document tree) into pandoc's Blocks

@@ -163,15 +161,17 @@ headlineToBlocks hdln = do

163161 let tags = headlineTags hdln

164162 let text = headlineText hdln

165163 let level = headlineLevel hdln

164+ shouldNotExport <- hasDoNotExportTag tags

166165 case () of

167− _ | any isNoExportTag tags -> return mempty

166+ _ | shouldNotExport -> return mempty

168167 _ | any isArchiveTag tags -> archivedHeadlineToBlocks hdln

169168 _ | isCommentTitle text -> return mempty

170169 _ | maxLevel <= level -> headlineToHeaderWithList hdln

171170 _ | otherwise -> headlineToHeaderWithContents hdln

172171

173−isNoExportTag :: Tag -> Bool

174−isNoExportTag = (== toTag "noexport")

172+hasDoNotExportTag :: Monad m => [Tag] -> OrgParser m Bool

173+hasDoNotExportTag tags = containsExcludedTag . orgStateExcludedTags <$> getState

174+ where containsExcludedTag s = any (`Set.member` s) tags

175175

176176isArchiveTag :: Tag -> Bool

177177isArchiveTag = (== toTag "ARCHIVE")

src/Text/Pandoc/Readers/Org/Meta.hs+11

@@ -52,6 +52,7 @@ import Data.Char (toLower)

5252import Data.List (intersperse)

5353import Data.Maybe (fromMaybe)

5454import qualified Data.Map as M

55+import qualified Data.Set as Set

5556import Network.HTTP (urlEncode)

5657

5758-- | Returns the current meta, respecting export options.

@@ -158,6 +159,7 @@ optionLine = try $ do

158159 "seq_todo" -> todoSequence >>= updateState . registerTodoSequence

159160 "typ_todo" -> todoSequence >>= updateState . registerTodoSequence

160161 "macro" -> macroDefinition >>= updateState . registerMacro

162+ "exclude_tags" -> excludedTagSet >>= updateState . setExcludedTags

161163 "pandoc-emphasis-pre" -> emphChars >>= updateState . setEmphasisPreChar

162164 "pandoc-emphasis-post" -> emphChars >>= updateState . setEmphasisPostChar

163165 _ -> mzero

@@ -190,6 +192,15 @@ parseFormat = try $ replacePlain <|> replaceUrl <|> justAppend

190192 rest = manyTill anyChar (eof <|> () <$ oneOf "\n\r")

191193 tillSpecifier c = manyTill (noneOf "\n\r") (try $ string ('%':c:""))

192194

195+excludedTagSet :: Monad m => OrgParser m (Set.Set Tag)

196+excludedTagSet = do

197+ skipSpaces

198+ Set.fromList . map Tag <$>

199+ many (orgTagWord <* skipSpaces) <* newline

200+

201+setExcludedTags :: Set.Set Tag -> OrgParserState -> OrgParserState

202+setExcludedTags tagSet st = st { orgStateExcludedTags = tagSet }

203+

193204setEmphasisPreChar :: Maybe [Char] -> OrgParserState -> OrgParserState

194205setEmphasisPreChar csMb st =

195206 let preChars = fromMaybe (orgStateEmphasisPostChars defaultOrgParserState) csMb

src/Text/Pandoc/Readers/Org/ParserState.hs+6

@@ -33,6 +33,7 @@ module Text.Pandoc.Readers.Org.ParserState

3333 , defaultOrgParserState

3434 , OrgParserLocal (..)

3535 , OrgNoteRecord

36+ , Tag(..)

3637 , HasReaderOptions (..)

3738 , HasQuoteContext (..)

3839 , HasMacros (..)

@@ -88,6 +89,9 @@ type OrgNoteTable = [OrgNoteRecord]

8889type OrgLinkFormatters = M.Map String (String -> String)

8990-- | Macro expander function

9091type MacroExpander = [String] -> String

92+-- | Tag

93+newtype Tag = Tag { fromTag :: String }

94+ deriving (Show, Eq, Ord)

9195

9296-- | The states in which a todo item can be

9397data TodoState = Todo | Done

@@ -113,6 +117,7 @@ data OrgParserState = OrgParserState

113117 -- specified here.

114118 , orgStateEmphasisPostChars :: [Char] -- ^ Chars allowed at after emphasis

115119 , orgStateEmphasisNewlines :: Maybe Int

120+ , orgStateExcludedTags :: Set.Set Tag

116121 , orgStateExportSettings :: ExportSettings

117122 , orgStateHeaderMap :: M.Map Inlines String

118123 , orgStateIdentifiers :: Set.Set String

@@ -183,6 +188,7 @@ defaultOrgParserState = OrgParserState

183188 , orgStateEmphasisCharStack = []

184189 , orgStateEmphasisNewlines = Nothing

185190 , orgStateExportSettings = def

191+ , orgStateExcludedTags = Set.singleton $ Tag "noexport"

186192 , orgStateHeaderMap = M.empty

187193 , orgStateIdentifiers = Set.empty

188194 , orgStateIncludeFiles = []

src/Text/Pandoc/Readers/Org/Parsing.hs+8

@@ -46,6 +46,8 @@ module Text.Pandoc.Readers.Org.Parsing

4646 , orgArgKey

4747 , orgArgWord

4848 , orgArgWordChar

49+ , orgTagWord

50+ , orgTagWordChar

4951 -- * Re-exports from Text.Pandoc.Parser

5052 , ParserContext (..)

5153 , many1Till

@@ -220,3 +222,9 @@ orgArgWord = many1 orgArgWordChar

220222-- | Chars treated as part of a word in plists.

221223orgArgWordChar :: Monad m => OrgParser m Char

222224orgArgWordChar = alphaNum <|> oneOf "-_"

225+

226+orgTagWord :: Monad m => OrgParser m String

227+orgTagWord = many1 orgTagWordChar

228+

229+orgTagWordChar :: Monad m => OrgParser m Char

230+orgTagWordChar = alphaNum <|> oneOf "@%#_"

test/command/4284.mdnew file+29

@@ -0,0 +1,29 @@

1+```

2+% pandoc -f org -t native

3+#+EXCLUDE_TAGS:apple cat bye dog %

5+* This should not appear :apple:

6+* NOEXPORT should appear if not specified in EXCLUDE_TAGS :noexport:

7+* This should not appear :cat:hi:laptop:

8+** Children of headers with excluded tags should not appear :xylophone:

9+* This should not appear :%:

10+^D

11+[Header 1 ("noexport-should-appear-if-not-specified-in-excludetags",[],[]) [Str "NOEXPORT",Space,Str "should",Space,Str "appear",Space,Str "if",Space,Str "not",Space,Str "specified",Space,Str "in",Space,Str "EXCLUDE",Subscript [Str "TAGS"],Space,Span ("",["tag"],[("tag-name","noexport")]) [SmallCaps [Str "noexport"]]]]

12+```

13+

14+```

15+% pandoc -f org -t native

16+#+EXCLUDE_TAGS:elephant

17+* This should not appear :elephant:

18+* This should appear :fawn:

19+^D

20+[Header 1 ("this-should-appear",[],[]) [Str "This",Space,Str "should",Space,Str "appear",Space,Span ("",["tag"],[("tag-name","fawn")]) [SmallCaps [Str "fawn"]]]]

21+```

22+

23+```

24+% pandoc -f org -t native

25+#+EXCLUDE_TAGS:

26+* NOEXPORT should appear if not specified in EXCLUDE_TAGS :noexport:

27+^D

28+[Header 1 ("noexport-should-appear-if-not-specified-in-excludetags",[],[]) [Str "NOEXPORT",Space,Str "should",Space,Str "appear",Space,Str "if",Space,Str "not",Space,Str "specified",Space,Str "in",Space,Str "EXCLUDE",Subscript [Str "TAGS"],Space,Span ("",["tag"],[("tag-name","noexport")]) [SmallCaps [Str "noexport"]]]]

29+```