Docx reader: omit "Table NN" from caption.

Closes #9002.
jgm · Aug 19, 2023 · 068fce4 · 068fce4
1 parent bd4de14
commit 068fce4
Show file tree

Hide file tree

Showing 5 changed files with 50 additions and 23 deletions.
diff --git a/pandoc.cabal b/pandoc.cabal
@@ -207,6 +207,7 @@ extra-source-files:
                  test/command/*.md
                  test/command/*.csl
                  test/command/*.svg
+                 test/command/9002.docx
                  test/command/biblio.bib
                  test/command/averroes.bib
                  test/command/A.txt

diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -86,13 +86,15 @@ import Text.Pandoc.XML.Light
       strContent,
       showElement,
       findAttr,
+      filterChild,
       filterChildrenName,
       filterElementName,
+      lookupAttrBy,
       parseXMLElement,
       elChildren,
       QName(QName, qName),
       Content(Elem),
-      Element(elContent, elName),
+      Element(..),
       findElements )
 
 data ReaderEnv = ReaderEnv { envNotes         :: Notes
@@ -725,7 +727,25 @@ elemToBodyPart ns element
       parstyle <- elemToParagraphStyle ns element
                   <$> asks envParStyles
                   <*> asks envNumbering
-      parparts' <- mconcat <$> mapD (elemToParPart ns) (elChildren element)
+
+      let hasCaptionStyle = elem "Caption" (pStyleId <$> pStyle parstyle)
+
+      let isTableNumberElt el@(Element name attribs _ _) =
+           (qName name == "fldSimple" &&
+             case lookupAttrBy ((== "instr") . qName) attribs of
+               Nothing -> False
+               Just instr -> "Table" `elem` T.words instr) ||
+           (qName name == "instrText" && "Table" `elem` T.words (strContent el))
+
+      let isTable = hasCaptionStyle &&
+                      isJust (filterChild isTableNumberElt element)
+
+      let stripOffLabel = dropWhile (not . isTableNumberElt)
+
+      let children = (if isTable
+                          then stripOffLabel
+                          else id) $ elChildren element
+      parparts' <- mconcat <$> mapD (elemToParPart ns) children
       fldCharState <- gets stateFldCharState
       modify $ \st -> st {stateFldCharState = emptyFldCharContents fldCharState}
       -- Word uses list enumeration for numbered headings, so we only
@@ -734,21 +754,9 @@ elemToBodyPart ns element
       case pHeading parstyle of
         Nothing | Just (numId, lvl) <- pNumInfo parstyle -> do
                     mkListItem parstyle numId lvl parparts
-        _ -> let
-          hasCaptionStyle = elem "Caption" (pStyleId <$> pStyle parstyle)
-
-          hasSimpleTableField = fromMaybe False $ do
-            fldSimple <- findChildByName ns "w" "fldSimple" element
-            instr <- findAttrByName ns "w" "instr" fldSimple
-            pure ("Table" `elem` T.words instr)
-
-          hasComplexTableField = fromMaybe False $ do
-            instrText <- findElementByName ns "w" "instrText" element
-            pure ("Table" `elem` T.words (strContent instrText))
-
-          in if hasCaptionStyle && (hasSimpleTableField || hasComplexTableField)
-            then return $ TblCaption parstyle parparts
-            else return $ Paragraph parstyle parparts
+        _ -> if isTable
+                then return $ TblCaption parstyle parparts
+                else return $ Paragraph parstyle parparts
 
 elemToBodyPart ns element
   | isElem ns "w" "tbl" element = do

diff --git a/test/command/9002.docx b/test/command/9002.docx
diff --git a/test/command/9002.md b/test/command/9002.md
@@ -0,0 +1,20 @@
+```
+% pandoc command/9002.docx -t html
+^D
+<table>
+<caption><p>This is my table!</p></caption>
+<colgroup>
+<col style="width: 50%" />
+<col style="width: 50%" />
+</colgroup>
+<thead>
+<tr class="header">
+<th>a</th>
+<th>b</th>
+</tr>
+</thead>
+<tbody>
+</tbody>
+</table>
+<p>See Table 1 This is my table!</p>
+```
diff --git a/test/docx/table_captions_with_field.native b/test/docx/table_captions_with_field.native
@@ -1,7 +1,6 @@
 [Para [Str "See",Space,Str "Table",Space,Str "1."]
-,Para []
-,Table ("",[],[]) (Caption Nothing
- [Para [Str "Table",Space,Str "1"]])
+,Para [Str "Table",Space,Str "1"]
+,Table ("",[],[]) (Caption Nothing [])
  [(AlignDefault,ColWidth 0.7605739372523825)
  ,(AlignDefault,ColWidth 0.11971303137380876)
  ,(AlignDefault,ColWidth 0.11971303137380876)]
@@ -32,8 +31,7 @@
  (TableFoot ("",[],[])
  [])
 ,Header 2 ("section", [], []) []
-,Table ("",[],[]) (Caption Nothing
- [Para [Str "Table",Space,Str "2"]])
+,Table ("",[],[]) (Caption Nothing [])
  [(AlignDefault,ColWidth 0.3332963620230701)
  ,(AlignDefault,ColWidth 0.3332963620230701)
  ,(AlignDefault,ColWidth 0.3334072759538598)]
@@ -50,5 +48,5 @@
   [])]
  (TableFoot ("",[],[])
  [])
-,Para []
+,Para [Str "Table",Space,Str "2"]
 ,Para [Str "See",Space,Str "Table",Space,Str "2."]]