48cef91 · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

[Docx Reader] Refactor/update smushInlines

48cef91Nikolay Yakimov6 years ago

src/Text/Pandoc/Readers/Docx/Combine.hs+32−44

compare with master

unifiedsplit

@@ -4,7 +4,8 @@

4{- |

5 Module : Text.Pandoc.Readers.Docx.Combine

7− 2014-2020 John MacFarlane <jgm@berkeley.edu>

7+ 2014-2020 John MacFarlane <jgm@berkeley.edu>,

8+ 2020 Nikolay Yakimov <root@livid.pp.ru>

8 License : GNU GPL, version 2 or above

9 License : GNU GPL, version 2 or above

10 Maintainer : Jesse Rosenthal <jrosenthal@jhu.edu>

11 Maintainer : Jesse Rosenthal <jrosenthal@jhu.edu>

@@ -59,7 +60,9 @@ module Text.Pandoc.Readers.Docx.Combine ( smushInlines

59 where

60 where

61import Data.List

62import Data.List

63+import Data.Bifunctor

62−import Data.Sequence (ViewL (..), ViewR (..), viewl, viewr, (><), (|>))

64+import Data.Sequence ( ViewL (..), ViewR (..), viewl, viewr, spanr, spanl

65+ , (><), (|>) )

63import qualified Data.Sequence as Seq (null)

66import qualified Data.Sequence as Seq (null)

64import Text.Pandoc.Builder

67import Text.Pandoc.Builder

@@ -80,14 +83,14 @@ spaceOutInlinesR ms = (stackInlines fs (l <> m'), r)

80spaceOutInlines :: Inlines -> (Inlines, Inlines, Inlines)

83spaceOutInlines :: Inlines -> (Inlines, Inlines, Inlines)

81spaceOutInlines ils =

84spaceOutInlines ils =

82 let (fs, ils') = unstackInlines ils

85 let (fs, ils') = unstackInlines ils

83− contents = unMany ils'

86+ (left, (right, contents')) = second (spanr isSpace) $ spanl isSpace $ unMany ils'

84− left = case viewl contents of

87+ -- NOTE: spanr counterintuitively returns suffix as the FIRST tuple element

85− (Space :< _) -> space

86− _ -> mempty

87− right = case viewr contents of

88− (_ :> Space) -> space

89− _ -> mempty in

90− (left, stackInlines fs $ trimInlines . Many $ contents, right)

88+ in (Many left, stackInlines fs $ Many contents', Many right)

89+

90+isSpace :: Inline -> Bool

91+isSpace Space = True

92+isSpace SoftBreak = True

93+isSpace _ = False

92stackInlines :: [Modifier Inlines] -> Inlines -> Inlines

95stackInlines :: [Modifier Inlines] -> Inlines -> Inlines

93stackInlines [] ms = ms

96stackInlines [] ms = ms

@@ -99,39 +102,24 @@ stackInlines (Modifier f : fs) ms =

99stackInlines (AttrModifier f attr : fs) ms = f attr $ stackInlines fs ms

102stackInlines (AttrModifier f attr : fs) ms = f attr $ stackInlines fs ms

100

103

101unstackInlines :: Inlines -> ([Modifier Inlines], Inlines)

104unstackInlines :: Inlines -> ([Modifier Inlines], Inlines)

102−unstackInlines ms = case ilModifier ms of

105+unstackInlines ms = case ilModifierAndInnards ms of

103− NullModifier -> ([], ms)

106+ (NullModifier, _) -> ([], ms)

104− _ -> (f : fs, ms') where

105− f = ilModifier ms

106− (fs, ms') = unstackInlines $ ilInnards ms

107+ (f, innards ) -> first (f :) $ unstackInlines innards

107−

108+

108−ilModifier :: Inlines -> Modifier Inlines

109+ilModifierAndInnards :: Inlines -> (Modifier Inlines, Inlines)

109−ilModifier ils = case viewl (unMany ils) of

110+ilModifierAndInnards ils = case viewl $ unMany ils of

110− (x :< xs) | Seq.null xs -> case x of

111+ x :< xs | Seq.null xs -> case x of

111− (Emph _) -> Modifier emph

112+ Emph lst -> (Modifier emph, fromList lst)

112− (Strong _) -> Modifier strong

113+ Strong lst -> (Modifier strong, fromList lst)

113− (SmallCaps _) -> Modifier smallcaps

114+ SmallCaps lst -> (Modifier smallcaps, fromList lst)

114− (Strikeout _) -> Modifier strikeout

115+ Strikeout lst -> (Modifier strikeout, fromList lst)

116+ Underline lst -> (Modifier underline, fromList lst)

115− (Superscript _) -> Modifier superscript

117+ Superscript lst -> (Modifier superscript, fromList lst)

116− (Subscript _) -> Modifier subscript

118+ Subscript lst -> (Modifier subscript, fromList lst)

117− (Link attr _ tgt) -> Modifier $ linkWith attr (fst tgt) (snd tgt)

119+ Link attr lst tgt -> (Modifier $ linkWith attr (fst tgt) (snd tgt), fromList lst)

118− (Span attr _) -> AttrModifier spanWith attr

120+ Span attr lst -> (AttrModifier spanWith attr, fromList lst)

119− _ -> NullModifier

121+ _ -> (NullModifier, ils)

120− _ -> NullModifier

122+ _ -> (NullModifier, ils)

121−

122−ilInnards :: Inlines -> Inlines

123−ilInnards ils = case viewl (unMany ils) of

124− (x :< xs) | Seq.null xs -> case x of

125− (Emph lst) -> fromList lst

126− (Strong lst) -> fromList lst

127− (SmallCaps lst) -> fromList lst

128− (Strikeout lst) -> fromList lst

129− (Superscript lst) -> fromList lst

130− (Subscript lst) -> fromList lst

131− (Link _ lst _) -> fromList lst

132− (Span _ lst) -> fromList lst

133− _ -> ils

134− _ -> ils

135

123

136inlinesL :: Inlines -> (Inlines, Inlines)

124inlinesL :: Inlines -> (Inlines, Inlines)

137inlinesL ils = case viewl $ unMany ils of

125inlinesL ils = case viewl $ unMany ils of

@@ -162,7 +150,7 @@ combineSingletonInlines x y =

162 in

150 in

163 case null shared of

151 case null shared of

164 True | isEmpty xs && isEmpty ys ->

152 True | isEmpty xs && isEmpty ys ->

165− stackInlines (x_rem_attr ++ y_rem_attr) mempty

153+ stackInlines (x_rem_attr <> y_rem_attr) mempty

166 | isEmpty xs ->

154 | isEmpty xs ->

167 let (sp, y') = spaceOutInlinesL y in

155 let (sp, y') = spaceOutInlinesL y in

168 stackInlines x_rem_attr mempty <> sp <> y'

156 stackInlines x_rem_attr mempty <> sp <> y'

test/docx/golden/inline_formatting.docx

Binary file not shown.

test/docx/inline_formatting.native+1−1

@@ -1,6 +1,6 @@

1Pandoc (Meta {unMeta = fromList []})

2[Para [Str "Regular",Space,Str "text",Space,Emph [Str "italics"],Space,Strong [Str "bold",Space,Emph [Str "bold",Space,Str "italics"]],Str "."]

3,Para [Str "This",Space,Str "is",Space,SmallCaps [Str "Small",Space,Str "Caps"],Str ",",Space,Str "and",Space,Str "this",Space,Str "is",Space,Strikeout [Str "strikethrough"],Str "."]

4−,Para [Str "Some",Space,Str "people",Space,Str "use",Space,Underline [Str "single",Space,Str "underlines",Space,Str "for",Space],Emph [Underline [Str "emphasis"]],Str "."]

4+,Para [Str "Some",Space,Str "people",Space,Str "use",Space,Underline [Str "single",Space,Str "underlines",Space,Str "for",Space,Emph [Str "emphasis"]],Str "."]

5,Para [Str "Above",Space,Str "the",Space,Str "line",Space,Str "is",Space,Superscript [Str "superscript"],Space,Str "and",Space,Str "below",Space,Str "the",Space,Str "line",Space,Str "is",Space,Subscript [Str "subscript"],Str "."]

6,Para [Str "A",Space,Str "line",LineBreak,Str "break."]]