Collecting Links

Input Sources

The Collector extracts links from four input types:

import io.mvnpm.raclette.collector.Collector;
import io.mvnpm.raclette.collector.Input;
import io.mvnpm.raclette.types.Uri;

Set<Uri> links = Collector.builder()
    .build()
    .collectLinks(Set.of(
        new Input.FsPath(Path.of("target/site")),        // local file or directory
        new Input.FsGlob("docs/**/*.html"),               // glob pattern
        new Input.RemoteUrl("https://example.com"),        // remote URL
        new Input.StringContent("<a href='...'>link</a>") // raw HTML string
    ));

Input.FsPath

Checks a single file or walks a directory recursively. Hidden files and directories are skipped by default.

Input.FsGlob

Matches files using glob patterns like docs/**/*.html.

Input.RemoteUrl

Fetches the page over HTTP and extracts links from the response body. The URL itself is used as the base for resolving relative links.

Input.StringContent

Parses a raw HTML string. Useful for testing or checking dynamically generated content.

Base URL Resolution

Relative links need a base URL to resolve against. Set it on the builder:

Collector.builder()
    .base("https://example.com/docs/")
    .build()
    .collectLinks(Set.of(new Input.StringContent(html)));

With base https://example.com/docs/:

• ../index.html resolves to https://example.com/index.html
• /about resolves to https://example.com/about
• page.html resolves to https://example.com/docs/page.html

For RemoteUrl inputs, the fetched URL is automatically used as the base.

Collector Options

What Gets Extracted

The collector uses JSoup to extract links from HTML elements:

• <a href="...">
• <img src="...">
• <link href="...">
• <script src="...">
• <source src="...">
• <img srcset="..."> (all URLs in srcset)
• <form action="...">
• Bare email addresses in text content