How can I extend Linguify for language-dependent images? - Converting linguify context to string

[SillyFreak]

Hi @TobiOneCenobi, welcome and thank you for your question! I have changed your post’s title to bring it in line with the question guidelines and thus make it easier to understand from the title:

Good titles are questions you would ask your friend about Typst.

---

First of all, the basic problem you’re facing is the following:

Why is the value I receive from context always content?

When working with context, you need to put everything that depends on the contextual information into the context block/expression itself. The explicit context expression is a trade-off: You get to know where in the document you are (accessible via here()) and which set rules are active there (accessible via things like text.lang). In return, the context value itself becomes opaque. You cannot peek into it, so everything that depends on the contextual information must happen within it. Let’s lo…

Linguify can’t change that, but it can offer an option that requires you to provide context instead of doing that itself. I also help maintain that package, so I can give a bit of insight. The linguify() function is defined here, and the problematic code is the following:

  if from == auto or lang == auto {
    // context is needed to use the default database or current language
    context impl()
  } else {
    impl()
  }

As you can see, you can work around this problem by providing the config and language explicitly, but that’s of course no fun. To support your use case (and have a github account), you could make a PR adding a parameter outside-context: false that, when set to true goes to the else branch. This enables you to get actual strings from the function, at the cost of having to call it as context do-something(linguify(...)).