Provides a configurable HTML sanitizer (`cleanify`) that walks the DOM tree, removes disallowed tags, strips disallowed attributes, and enforces iframe domain whitelisting. It is self-contained (only depends on `linkedom` for parsing) and can be extracted into any project needing safe HTML output. The implementation handles edge cases like nested elements and attribute removal efficiently.

A collection of URL utility functions including validation, purification (removing tracking parameters from a curated blacklist), absolutification of relative URLs, normalization (making all links absolute and adding target=_blank), and best URL selection via string similarity to the article title. The tracking parameter removal is particularly well-curated and reusable in any scraping or URL processing pipeline.

Implements a registration and execution system for per-site transformations. Users can add transformations (pre/post functions) that match URL patterns, allowing site-specific DOM manipulation before or after readability extraction. The module is self-contained and can be reused as a lightweight middleware pattern in other extraction or web scraping tools.

Extracts metadata (author, date, description, image, type) from JSON-LD `<script>` tags. It validates the schema type against a list of known article/webpage types, parses the JSON, and extracts the relevant fields. The logic handles arrays, objects, and nested structures gracefully. This module is reusable for any project needing to parse JSON-LD metadata from HTML.

A heuristic date extraction function that searches the DOM for date patterns in `<time>` elements, elements with `datetime` attribute, itemprop attributes, and then falls back to scanning text content of paragraphs, spans, and divs. It handles multiple date formats and converts them to ISO strings. This is a self-contained utility that can be reused in any HTML parsing project needing date discovery.