LaTeX reader: add Support for `glossaries` and `acronym` package (#3589)

Acronyms are not resolved by the reader, but acronym and glossary information is put into attributes on Spans so that they can be processed in filters.

2 files changed, 84 insertions, 0 deletions

diff --git a/src/Text/Pandoc/Readers/LaTeX.hs b/src/Text/Pandoc/Readers/LaTeX.hs
index 986547f42..5627e4ee1 100644
--- a/src/Text/Pandoc/Readers/LaTeX.hs
+++ b/src/Text/Pandoc/Readers/LaTeX.hs
@@ -702,6 +702,21 @@ enquote = do
      then singleQuoted <$> withQuoteContext InSingleQuote tok
      else doubleQuoted <$> withQuoteContext InDoubleQuote tok
 
+doAcronym :: PandocMonad m => String -> LP m Inlines
+doAcronym form = do
+  acro <- braced
+  return . mconcat $ [spanWith ("",[],[("acronym-label", toksToString acro), 
+    ("acronym-form", "singular+" ++ form)]) 
+    $ str $ toksToString acro]
+
+doAcronymPlural :: PandocMonad m => String -> LP m Inlines
+doAcronymPlural form = do
+  acro <- braced
+  plural <- lit "s"
+  return . mconcat $ [spanWith ("",[],[("acronym-label", toksToString acro), 
+    ("acronym-form", "plural+" ++ form)]) $ mconcat 
+    $ [str $ toksToString acro, plural]]
+
 doverb :: PandocMonad m => LP m Inlines
 doverb = do
   Tok _ Symbol t <- anySymbol
@@ -1371,6 +1386,30 @@ inlineCommands = M.fromList $
   , ("nocite", mempty <$ (citation "nocite" NormalCitation False >>=
                           addMeta "nocite"))
   , ("hypertarget", braced >> tok)
+  -- glossaries package
+  , ("gls", doAcronym "short")
+  , ("Gls", doAcronym "short")
+  , ("glsdesc", doAcronym "long")
+  , ("Glsdesc", doAcronym "long")
+  , ("GLSdesc", doAcronym "long")
+  , ("acrlong", doAcronym "long")
+  , ("Acrlong", doAcronym "long")
+  , ("acrfull", doAcronym "full")
+  , ("Acrfull", doAcronym "full")
+  , ("acrshort", doAcronym "abbrv")
+  , ("Acrshort", doAcronym "abbrv")
+  , ("glspl", doAcronymPlural "short")
+  , ("Glspl", doAcronymPlural "short")
+  , ("glsdescplural", doAcronymPlural "long")
+  , ("Glsdescplural", doAcronymPlural "long")
+  , ("GLSdescplural", doAcronymPlural "long")
+  -- acronyms package
+  , ("ac", doAcronym "short")
+  , ("acf", doAcronym "full")
+  , ("acs", doAcronym "abbrv")
+  , ("acp", doAcronymPlural "short")
+  , ("acfp", doAcronymPlural "full")
+  , ("acsp", doAcronymPlural "abbrv")
   -- siuntix
   , ("SI", dosiunitx)
   -- hyphenat
diff --git a/test/command/3539.md b/test/command/3539.md
new file mode 100644
index 000000000..6ef9ffac7
--- /dev/null
+++ b/test/command/3539.md
@@ -0,0 +1,45 @@
+# Commands of [glossaries package](ftp://ftp.tu-chemnitz.de/pub/tex/macros/latex/contrib/glossaries/glossaries-code.pdf)
+
+```
+% pandoc -f latex -t native
+Many programming languages provide \glspl{API}. Each \gls{API} should provide a documentation.
+^D
+[Para [Str "Many",Space,Str "programming",Space,Str "languages",Space,Str "provide",Space,Span ("",[],[("acronym-label","API"),("acronym-form","plural+short")]) [Str "APIs"],Str ".",Space,Str "Each",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+short")]) [Str "API"],Space,Str "should",Space,Str "provide",Space,Str "a",Space,Str "documentation."]]
+```
+
+```
+% pandoc -f latex -t native
+\Glsdesc{API} XYZ ist not as performant as \glsdesc{API} ZXY.
+^D
+[Para [Span ("",[],[("acronym-label","API"),("acronym-form","singular+long")]) [Str "API"],Space,Str "XYZ",Space,Str "ist",Space,Str "not",Space,Str "as",Space,Str "performant",Space,Str "as",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+long")]) [Str "API"],Space,Str "ZXY."]]
+```
+
+```
+% pandoc -f latex -t native
+\Acrlong{API} XYZ ist not as performant as \acrlong{API} ZXY.
+^D
+[Para [Span ("",[],[("acronym-label","API"),("acronym-form","singular+long")]) [Str "API"],Space,Str "XYZ",Space,Str "ist",Space,Str "not",Space,Str "as",Space,Str "performant",Space,Str "as",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+long")]) [Str "API"],Space,Str "ZXY."]]
+```
+
+```
+% pandoc -f latex -t native
+\Acrfull{API} XYZ ist not as performant as \acrfull{API} ZXY.
+^D
+[Para [Span ("",[],[("acronym-label","API"),("acronym-form","singular+full")]) [Str "API"],Space,Str "XYZ",Space,Str "ist",Space,Str "not",Space,Str "as",Space,Str "performant",Space,Str "as",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+full")]) [Str "API"],Space,Str "ZXY."]]
+```
+
+```
+% pandoc -f latex -t native
+\Acrshort{API} XYZ ist not as performant as \acrshort{API} ZXY.
+^D
+[Para [Span ("",[],[("acronym-label","API"),("acronym-form","singular+abbrv")]) [Str "API"],Space,Str "XYZ",Space,Str "ist",Space,Str "not",Space,Str "as",Space,Str "performant",Space,Str "as",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+abbrv")]) [Str "API"],Space,Str "ZXY."]]
+```
+
+# Commands of [acronym package](ftp://ftp.mpi-sb.mpg.de/pub/tex/mirror/ftp.dante.de/pub/tex/macros/latex/contrib/acronym/acronym.pdf)
+
+```
+% pandoc -f latex -t native
+Many programming languages provide \acp{API}. Each \ac{API} should provide a documentation.
+^D
+[Para [Str "Many",Space,Str "programming",Space,Str "languages",Space,Str "provide",Space,Span ("",[],[("acronym-label","API"),("acronym-form","plural+short")]) [Str "APIs"],Str ".",Space,Str "Each",Space,Span ("",[],[("acronym-label","API"),("acronym-form","singular+short")]) [Str "API"],Space,Str "should",Space,Str "provide",Space,Str "a",Space,Str "documentation."]]
+```