Is there a way to only return the TOC

[jonathanbriehl]

Hello ~

I've been struggling trying to figure out a way to get only the table of contents. I am using the Table of Contents extension and when return the markdown, the TOC is there. What I guess I'm wondering is if there's a way to passing the markdown and instead of returning everything along with the TOC is if it's possible to only return the TOC?

Thanks so much for the help!

[colinodell]

The Table of Contents is stored within a TableOfContents node in the AST. There are a few ways to accomplish what you're looking to do, but all of them will center around locating that node. Given a Document object named $document, you can grab the TOC with code like this:

$toc = (new Query())
    ->where(Query::type(TableOfContents::class))
    ->findOne($document);

if ($toc === null) {
    // No TOC is present
} else {
    // FOUND IT!
}

Or without using the fancy Query functionality:

function getTOC(Document $document): ?TableOfContents
{
    foreach ($document->iterator() as $node) {
        if ($node instanceof TableOfContents) {
            return $node;
        }
    }

    return null;
}

There are two main ways you could use this:

Manually parse and render the TOC

With this approach, you have to wire some of the parsing/rendering logic yourself, but it's not too bad:

$environment = new Environment();
// TODO: add any/all extensions you want

$document = (new MarkdownParser($environment))->parse($markdown);

$toc = (new Query())
    ->where(Query::type(TableOfContents::class))
    ->findOne($document);

if ($toc === null) {
    // No TOC is present; you might want to handle this somehow
}

$tocHtml = (new HtmlRenderer($environment))->renderNodes([$toc]);

Use an event listener to remove everything except the TOC

If you'd like to avoid that wiring, you can create a custom event listener that removes everything from the Document except for the TableOfContents:

final class RemoveEverythingExceptTOCListener
{
    public function __invoke(DocumentParsedEvent $event): void
    {
        $document = $event->getDocument();

        $toc = (new Query())
            ->where(Query::type(TableOfContents::class))
            ->findOne($document);

        if ($toc === null) {
            // No TOC is present; remove all children
            $event->getDocument()->replaceChildren([]);
        } else {
            $event->getDocument()->replaceChildren([$toc]);
        }
    }
}

You can then register that with the environment:

$environment->addEventListener(DocumentParsedEvent::class, new RemoveEverythingExceptTOCListener(), -200); // I'm not 100% sure about the priority here

Or bundle it into a custom extension that registers it.

---

I have not tested any of this code but I'm confident that it should at least be very very close to what you need.

Hope that helps!

[CapitaineToinon]

Thanks, works like a charm!

Personally I'm also detaching the toc from the document because I also need the rest of the document. I'm doing something like this

        $converter = new GithubFlavoredMarkdownConverter($options);
        $environment = $converter->getEnvironment();

        foreach ($extensions as $extension) {
            $environment->addExtension($extension);
        }

        $converted = $converter->convert($page->content);

        $document = $converted->getDocument();

        $toc = (new Query())
            ->where(Query::type(TableOfContents::class))
            ->findOne($document);

        $toc->detach();

        $renderer = new HtmlRenderer($environment);
        $content = $renderer->renderDocument($document);
        $toc = $renderer->renderNodes([$toc]);

And that way I have my content I can put in the main container and the toc I can put aside, with different css and html structure.