Docx reader: handle `\b`, `\i`, `\y` modifiers in `XE` index entries.

See #10171.

3 files changed, 32 insertions, 16 deletions

diff --git a/src/Text/Pandoc/Readers/Docx.hs b/src/Text/Pandoc/Readers/Docx.hs
index 8bd4aa8b5..a9e53eb39 100644
--- a/src/Text/Pandoc/Readers/Docx.hs
+++ b/src/Text/Pandoc/Readers/Docx.hs
@@ -472,10 +472,13 @@ parPartToInlines' (OMathPara exps) =
 parPartToInlines' (Field info children) =
   case info of
     HyperlinkField url -> parPartToInlines' $ ExternalHyperLink url children
-    IndexrefField entry mbsee ->
+    IndexrefField ie ->
       pure $ spanWith ("",["indexref"],
-                           (("entry",entry) :
-                             maybe [] (\x -> [("crossref",x)]) mbsee)) mempty
+                           (("entry", entryTitle ie) :
+                             maybe [] (\x -> [("crossref",x)]) (entrySee ie)
+                          ++ maybe [] (\x -> [("yomi",x)]) (entryYomi ie)
+                          ++ [("bold","") | entryBold ie]
+                          ++ [("italic","") | entryItalic ie])) mempty
     PagerefField fieldAnchor True -> parPartToInlines' $ InternalHyperLink fieldAnchor children
     EndNoteCite t -> do
       formattedCite <- smushInlines <$> mapM parPartToInlines' children
diff --git a/src/Text/Pandoc/Readers/Docx/Fields.hs b/src/Text/Pandoc/Readers/Docx/Fields.hs
index f21ad425e..8b1837394 100644
--- a/src/Text/Pandoc/Readers/Docx/Fields.hs
+++ b/src/Text/Pandoc/Readers/Docx/Fields.hs
@@ -12,20 +12,30 @@ For parsing Field definitions in instText tags, as described in
 ECMA-376-1:2016, §17.16.5 -}
 
 module Text.Pandoc.Readers.Docx.Fields ( FieldInfo(..)
+                                       , IndexEntry(..)
                                        , parseFieldInfo
                                        ) where
 
 import Data.Functor (($>), void)
 import qualified Data.Text as T
 import Text.Pandoc.Parsing
+import Data.Maybe (isJust)
 
 type URL = T.Text
 type Anchor = T.Text
 
+data IndexEntry = IndexEntry
+  { entryTitle :: T.Text
+  , entrySee :: Maybe T.Text
+  , entryYomi :: Maybe T.Text
+  , entryBold :: Bool
+  , entryItalic :: Bool }
+  deriving (Show)
+
 data FieldInfo = HyperlinkField URL
                 -- The boolean indicates whether the field is a hyperlink.
                | PagerefField Anchor Bool
-               | IndexrefField T.Text (Maybe T.Text) -- second is optional 'see'
+               | IndexrefField IndexEntry
                | CslCitation T.Text
                | CslBibliography
                | EndNoteCite T.Text
@@ -41,11 +51,11 @@ parseFieldInfo = parse fieldInfo ""
 fieldInfo :: Parser FieldInfo
 fieldInfo = do
   spaces
-  (HyperlinkField <$> hyperlink)
+  hyperlink
     <|>
-    ((uncurry PagerefField) <$> pageref)
+    pageref
     <|>
-    ((uncurry IndexrefField) <$> indexref)
+    indexref
     <|>
     addIn
     <|>
@@ -101,7 +111,7 @@ fieldArgument = do
   notFollowedBy (char '\\') -- switch
   quotedString <|> unquotedString
 
-hyperlink :: Parser URL
+hyperlink :: Parser FieldInfo
 hyperlink = do
   string "HYPERLINK"
   spaces
@@ -110,7 +120,7 @@ hyperlink = do
   let url = case [s | ('l',s) <- switches] of
               [s] -> farg <> "#" <> s
               _   -> farg
-  return url
+  return $ HyperlinkField url
 
 -- See §17.16.5.45
 fieldSwitch :: Parser (Char, T.Text)
@@ -119,25 +129,27 @@ fieldSwitch = try $ do
   char '\\'
   c <- anyChar
   spaces
-  farg <- fieldArgument
+  farg <- option mempty fieldArgument
   return (c, farg)
 
-pageref :: Parser (Anchor, Bool)
+pageref :: Parser FieldInfo
 pageref = do
   string "PAGEREF"
   spaces
   farg <- fieldArgument
   switches <- many fieldSwitch
   let isLink = any ((== 'h') . fst) switches
-  return (farg, isLink)
+  return $ PagerefField farg isLink
 
 -- second element of tuple is optional "see".
-indexref :: Parser (T.Text, Maybe T.Text)
+indexref :: Parser FieldInfo
 indexref = do
   string "XE"
   spaces
   farg <- fieldArgument
   switches <- spaces *> many fieldSwitch
-  case [see | ('t', see) <- switches] of
-     [see] -> pure (farg, Just see)
-     _     -> pure (farg, Nothing)
+  return $ IndexrefField $ IndexEntry{ entryTitle = farg
+                                     , entrySee = lookup 't' switches
+                                     , entryYomi = lookup 'y' switches
+                                     , entryBold = isJust (lookup 'b' switches)
+                                     , entryItalic = isJust (lookup 'i' switches) }
diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index 3b0c5fdc6..20e510d8d 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -43,6 +43,7 @@ module Text.Pandoc.Readers.Docx.Parse ( Docx(..)
                                       , ChangeType(..)
                                       , ChangeInfo(..)
                                       , FieldInfo(..)
+                                      , IndexEntry(..)
                                       , Level(..)
                                       , ParaStyleName
                                       , CharStyleName