Hakyll - Converting file links to .html links when compiling
In the knowledge base I am creating (a collection of articles and thoughts that have helped me) I have links to other Org files to avoid clutter and keep these pages loading fast. Suppose we have the following link in an Org file
* [[file:Python.org][Python]]
By default when this is compiled by Pandoc to html, it will remain as a link to the org file.
<a href="Python.org">Python</a>
Which is not what I want.
But how do we modify “href” targets without modifying anything else? For example I want the following text: dontmodifyme.org to remain as such, not to be dontmodifyme.html, since it isn’t inside an Org link.
In order to accomplish this, I had to add a transform to the Pandoc compiler in my site.hs.
Here is the code for the compiler with the transform added
Here is where the compiler is used on the knowledge base
This compiler is also used in other places in site.hs, namely where blog posts are compiled. When writing the previous blog post, I linked to this blog post as an org file, then followed it to create this post. The compiler turns it into an .html link.
How the compiler works
It’s fairly simple. Even being extremely out of practice with Haskell, I was able to whip this up.
Create the compiler with pandocCompilerWithTransform
Supply some default options and a Pandoc -> Pandoc function.
pandocPostCompiler :: Compiler (Item String) pandocPostCompiler = pandocCompilerWithTransform defaultHakyllReaderOptions defaultHakyllWriterOptions orgToHtml
Walk the Pandoc AST
To parse the AST (Abstract Syntax Tree) or the initial compiler results, the `walk` function is used. This allows us to follow the structure until Links are found and modify the text directly.
orgToHtml :: Pandoc -> Pandoc orgToHtml = walk $ \inline -> case inline of Link attr inline (url, title) -> Link attr inline (pack(orgRegex (unpack url)), title) _ -> inline
Apply the substitution on link targets
This is the substitution code. It’s matching on any string that ends with .org and capturing what comes before .org, and then replacing it with that capture group with .html appended.
orgRegex :: String -> String orgRegex t = subRegex (mkRegex "^(.*?)\\.org$") t "\\1.html"
Done!
I have a feeling I am going to be doing more things like this as the site and my Org knowledge expand, but I think I will be ready!