Chunking for formats other than HTML?

[iandol]

For HTML / EPub, there is an internal chunking mechanism to split a single document to multiple outputs. There are cases where it would be useful for output formats other than HTML (tools that require multiple files for chapters). For example split a single markdown document to multiple markdown files:

pandoc --chunked -t markdown input.md

Chapter links would get modified to include the file.md#id and reference links would be converted to inline. I don't know how feasible this is (depending where the chunking code sits in the processing chain) but as the problems are already solved for EPub, I would hope this is just about making the existing code accessible to other writers?

[jgm]

I guess the thing to do would be to look at the Text.Pandoc.Writers.ChunkedHTML module and see how much of it is HTML-specific. It's possible that we could abstract out a higher-order function into which specific writer functions could be plugged, to yield a "chunked" version.

[iandol]

I'd be happy to help, though Haskell looks like Klingon to me 🤣 — let me see what I can glean…

[iandol]

There is also Text.Pandoc.Chunks that seems more generic:

https://github.com/jgm/pandoc/blob/main/src/Text/Pandoc/Chunks.hs

[jgm]

Yes, that is generic, and T.P.Writers.ChunkedHTML uses it.

That's why T.P.Writers.ChunkedHTML will show what else needs to be done for a writer, beyond chunking.