Re: FIguring out pandoc AST and types

[John MacFarlane <jgm-TVLZxgkOlNX2fBVCVOL8/A@public.gmane.org>]

Harsh Donga <harsh-7+aFW328pE6p1wGUEcWPqti2O/JbrIOy@public.gmane.org> writes:

> Thank you for your response,
>
> It is getting really hard for me to parse and build pandoc using pandoc 
> types, as I am Haskell beginner.
>
> Please correct me if I am wrong, but the only way to see the JSON output is 
> by trying out pandoc? I mean for covering all the pandoc-types, I will have 
> to put everything in a document then parse using pandoc?
> And then figure out by looking at output how actually all types are stored 
> in JSON?

There isn't complete documentation of the JSON format we use,
but it's all very regular and can be predicted from the AST given
a few examples. Pandoc can be used to give you those examples.

Or, you can look at the ToJSON and FromJSON instance definitions
in the source code for Text.Pandoc.Definition and figure it out
from that.

However, if what you want to do is to use Haskell to create a
pandoc AST, then you don't need to worry about JSON at all. JSON
is just a serialization format we use for communicating with
non-Haskell programs. For building a document, what you want is
Text.Pandoc.Builder, as I said. The documentation contains an
example to get you started.