Markdown to Typst: Typographic substitutions

[jsx97]

Test file input.md:

= Lorem ipsum

Dolor sit --- amet. Dolor "sit" amet.

• If I convert it to Typst using pandoc -o output.typ input.md, the output is

  = Lorem ipsum
  <lorem-ipdum>
  Dolor sit — amet. Dolor "sit" amet.
  

  (--- is replaced with em dash.)

• If I convert it to Typst using pandoc -f markdown-smart -o output.typ input.md, the output is

  = Lorem ipsum
  <lorem-ipdum>
  Dolor sit --- amet. Dolor \"sit\" amet.
  

Questions:

• How is it possible to convert from Markdown to Typst so that the output will be

  = Lorem ipsum
  <lorem-ipdum>
  Dolor sit — amet. Dolor “sit” amet.
  

  (--- is replaced with em dash. Straight quotation marks are replaced with curly ones.)

• How is it possible to convert from Markdown to Typst so that the output will be

  = Lorem ipsum
  <lorem-ipdum>
  Dolor sit --- amet. Dolor "sit" amet.
  

  (Quotation marks are straight and are not escaped. --- is not replaced with em dash.)

[jgm]

First, you probably want # instead of = for the markdown heading.

Pandoc parses this as

[ Header
    1
    ( "lorem-ipsum" , [] , [] )
    [ Str "Lorem" , Space , Str "ipsum" ]
, Para
    [ Str "Dolor"
    , Space
    , Str "sit"
    , Space
    , Str "\8212"
    , Space
    , Str "amet."
    , Space
    , Str "Dolor"
    , Space
    , Quoted DoubleQuote [ Str "sit" ]
    , Space
    , Str "amet."
    ]
]

but the Quoted element is rendered to typst using straight quotes, because typst will interpret them as if they're curly. (You'll get curly quotes after you compile with typst.)

[jsx97]

First, you probably want # instead of = for the markdown heading.

Yes, sorry. My bad, I was in a hurry.

Not sure I understand the rest part...

[cskeeters]

Quotes

Here, I get why you'd want Typst to handle. The user may have a custom template with #smartquote configured.

Smart	Markdown	Typst	Comment
+smart	I said "no"	I said "no".	Nice, let Typst handle
+smart	I said 'no'	I said 'no'.	Nice, let Typst handle
-smart	I said "no"	I said "no".	Fine, let template handle
-smart	I said 'no'	I said 'no'.	Fine, let template handle

Apostrophes

When smart is on, why not let Typst handle apostrophes?

Smart	Markdown	Typst	Comment
+smart	Don't do it	Don’t do it	Why not let Typst handle?
+smart	Don\'t do it	Don\'t do it	Makes sense
-smart	Don't do it	Don\'t do it	Why not let the template handle?
-smart	Don\'t do it	Don\'t do it	Makes sense

[thjbdvlt]

Absolutely ugly, yes! I was actually thinking about Typst as --pdf-engine; you're right that it would make the .typ unreadable.

(An alias function may be used, like #q[...] but it would probably be weird to define aliases function there.)

[jgm]

That's true, #q is not a bad idea, and it could be defined in our default template. It makes the typst less portable though.

[cskeeters]

I've been testing main since you put in the fix for me. I love it! Thank you.

I have concluded that you can use it two ways. Re-explaining helps me flush out the detail...

Say you have smart.md:

First -- Last

First --- Last

I said "no"

Don't do it

NOTE: This is valid Typst code and will utilize smart quotes on compile, so I want the output to be the same as the input, disregarding stuff related to making a standalone document.

If one runs pandoc smart.md -t typst, they are now effectively running pandoc -f markdown+smart smart.md -t typst+smart and the output will be unchanged from the markdown source. Yay! With or without -s (standalone) the dashes and quotes will get converted by typst compile. When run like this, the markdown reader does converts dashes and strait quotes into curly versions in the AST, but typst+smart converts them back before being output.

(Made with pandoc smart.md -t native)

[ Para
    [ Str "First" , Space , Str "\8211" , Space , Str "Last" ]
, Para
    [ Str "First" , Space , Str "\8212" , Space , Str "Last" ]
, Para
    [ Str "I"
    , Space
    , Str "said"
    , Space
    , Quoted DoubleQuote [ Str "no" ]
    ]
, Para
    [ Str "Don\8217t" , Space , Str "do" , Space , Str "it" ]
]

NOTE: \8211 used for en-dash, \8212 used for em-dash, Quoted DoubleQuote and \8217 for apostrophe.

If one runs pandoc -f markdown-smart smart.md -t typst-smart, the same output is generated without -s, but the native AST does not have unicode representations in the AST.

(Made with pandoc -f markdown-smart smart.md -t native)

[ Para
    [ Str "First" , Space , Str "--" , Space , Str "Last" ]
, Para
    [ Str "First" , Space , Str "---" , Space , Str "Last" ]
, Para
    [ Str "I" , Space , Str "said" , Space , Str "\"no\"" ]
, Para [ Str "Don't" , Space , Str "do" , Space , Str "it" ]
]

NOTE: Mostly no translation.
NOTE: Escaped double strait quotes. (? -- See question below)

The new default template for typst will output #set smartquote(enabled: false) when typst-smart is specified which makes quotes strait in the compiled PDF when run translated with -s and compiled with typst compile.

It appears a second way to achieve curly quotes it to run pandoc -s -f markdown-smart smart.md -t typst-smart --template <custom> where the custom template has the if block with #set smartquote(enabled: false) removed. Since I have my own templates already, I just have to not add it. Both ways work with the latter option being closer to what I intuitively want to happen.

Question: Why are the double quotes escaped in the AST for markdown-smart input?

[jgm]

This escaping only relates to the Haskell representation (because " is used for strings, you need \" for a literal quote inside a string). Don't worry, you won't get a backslash in rendered output.

[cskeeters]

Ah, of course. Thx for the reply.