5527465 · typeable/pandoc

git.iko.soy/typeable/pandoc

Docx reader: Handle dummy list items.

These come up when people create a list item and then delete the bullet. It doesn't refer to any real list item, and we used to ignore it. We handle it with a DummyListItem type, which, in Docx.hs, is turned into a normal paragraph with a "ListParagraph" class. If it follow another list item, it is folded as another paragraph into that item. If it doesn't, it's just its own (usually indented, and therefore block-quoted) paragraph.

commit5527465c778736e91e88fae414ea26558b966724 authorJesse Rosenthal <jrosenthal@jhu.edu> dateWed 18 Nov 2015, 13:02 (11 years ago) parent1f2c373

2 files changed+15−6

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Docx.hs+4 src/Text/Pandoc/Readers/Docx/Parse.hs+11−6

src/Text/Pandoc/Readers/Docx.hs+4

@@ -503,6 +503,10 @@ bodyPartToBlocks (ListItem pPr numId lvl levelInfo parparts) = do

503503 ]

504504 blks <- bodyPartToBlocks (Paragraph pPr parparts)

505505 return $ divWith ("", ["list-item"], kvs) blks

506+bodyPartToBlocks (DummyListItem pPr _ parparts) =

507+ let pPr' = pPr {pStyle = "ListParagraph": (pStyle pPr)}

508+ in

509+ bodyPartToBlocks $ Paragraph pPr' parparts

506510bodyPartToBlocks (Tbl _ _ _ []) =

507511 return $ para mempty

508512bodyPartToBlocks (Tbl cap _ look (r:rs)) = do

src/Text/Pandoc/Readers/Docx/Parse.hs+11−6

@@ -144,9 +144,6 @@ type Level = (String, String, String, Maybe Integer)

144144data DocumentLocation = InDocument | InFootnote | InEndnote

145145 deriving (Eq,Show)

146146

147−-- data RelationshipType = DocumentRel | FootnoteRel | EndnoteRel

148−-- deriving Show

149−

150147data Relationship = Relationship DocumentLocation RelId Target

151148 deriving Show

152149

@@ -181,6 +178,7 @@ defaultParagraphStyle = ParagraphStyle { pStyle = []

181178

182179data BodyPart = Paragraph ParagraphStyle [ParPart]

183180 | ListItem ParagraphStyle String String Level [ParPart]

181+ | DummyListItem ParagraphStyle String [ParPart]

184182 | Tbl String TblGrid TblLook [Row]

185183 | OMathPara [Exp]

186184 deriving Show

@@ -245,7 +243,6 @@ defaultRunStyle = RunStyle { isBold = Nothing

245243 , rUnderline = Nothing

246244 , rStyle = Nothing}

247245

248−

249246type Target = String

250247type Anchor = String

251248type URL = String

@@ -411,6 +408,13 @@ archiveToMedia :: Archive -> Media

411408archiveToMedia zf =

412409 mapMaybe (getMediaPair zf) (filter filePathIsMedia (filesInArchive zf))

413410

411+-- lookupLevel :: String -> String -> Numbering -> Maybe Level

412+-- lookupLevel numId ilvl (Numbering _ numbs absNumbs) = do

413+-- absNumId <- lookup numId $ map (\(Numb nid absnumid) -> (nid, absnumid)) numbs

414+-- lvls <- lookup absNumId $ map (\(AbstractNumb aid ls) -> (aid, ls)) absNumbs

415+-- lvl <- lookup ilvl $ map (\l@(i, _, _, _) -> (i, l)) lvls

416+-- return lvl

417+

414418lookupLevel :: String -> String -> Numbering -> Maybe Level

415419lookupLevel numId ilvl (Numbering _ numbs absNumbs) = do

416420 absNumId <- lookup numId $ map (\(Numb nid absnumid) -> (nid, absnumid)) numbs

@@ -418,6 +422,7 @@ lookupLevel numId ilvl (Numbering _ numbs absNumbs) = do

418422 lvl <- lookup ilvl $ map (\l@(i, _, _, _) -> (i, l)) lvls

419423 return lvl

420424

425+

421426numElemToNum :: NameSpaces -> Element -> Maybe Numb

422427numElemToNum ns element |

423428 qName (elName element) == "num" &&

@@ -569,7 +574,7 @@ elemToBodyPart ns element

569574 num <- asks envNumbering

570575 case lookupLevel numId lvl num of

571576 Just levelInfo -> return $ ListItem parstyle numId lvl levelInfo parparts

572− Nothing -> throwError WrongElem

577+ Nothing -> return $ DummyListItem parstyle lvl parparts

573578elemToBodyPart ns element

574579 | isElem ns "w" "p" element = do

575580 sty <- asks envParStyles

@@ -582,7 +587,7 @@ elemToBodyPart ns element

582587 Just levelInfo ->

583588 return $ ListItem parstyle numId lvl levelInfo parparts

584589 Nothing ->

585− throwError WrongElem

590+ return $ DummyListItem parstyle lvl parparts

586591 Nothing -> return $ Paragraph parstyle parparts

587592elemToBodyPart ns element

588593 | isElem ns "w" "tbl" element = do