jcpaik / sofa-mdbook Goto Github PK

• No Ext_wikilinks_title_after_pipe in Pandoc < 3.0

In md,

![[wiki]]

![wiki](image.png)

In pandoc,

[ Para
    [ Str "!"
    , Link
        ( "" , [] , [] ) [ Str "wiki" ] ( "wiki" , "wikilink" )
    ]
, Para
    [ Image ( "" , [] , [] ) [ Str "wiki" ] ( "image.png" , "" )
    ]
]

Current implementation assumes a fixed depth of .md files.

[ ] Remove tags
[ ] Query links through all files

Right now the preprocessor is a Cabal script. I would have prefered to use a stack script like this

#!/usr/bin/env stack
{- stack script --resolver=nightly-2023-02-14 --verbosity=info
  --package process
  --package pandoc
  --package pandoc-types
  --package text
-}

but it just don't give me the needed Pandoc >= 3.0 saying that it is hidden.
LTS resolvers work with no 'hidden package' error, but it doesn't have >= 3.0
and I already wasted a full working day on it just to make it work.
I might better make a full project, but this is just a script and I don't want all the extra steps/files to compile and run.

Possible options is <div style="line-height:150%;"> <br> </div> or anything in the following.
https://stackoverflow.com/questions/1409649/how-to-change-the-height-of-a-br
CSS would be great too.

Parsing with Markdown reader, Pandoc does not recognize lists in a quote.

Using CommonMark reader, we can express list in a quotation but doing so leads to multilple errors:

• - inside multiline equations, if started in a line, is considered as a list not a part of equation.
• Even after fixing a few corners, parsing the file as commonmark does not work (WHY?)

Use Turtle

• Transverse a directory recursively.
• Check if filename matches either starts with [0-9A-Z][0-9].
  • If not, ignore.
• If it is a directory, add this as a section name
• If it is a file, add a file section

We need an abstract tree structure that represents the content of the document. Right now we use file paths for guessing the depth.