html-cleaning

HTML cleaning, sanitization, and text processing utilities for Rust.

Features

HTML Cleaning: Remove unwanted elements (scripts, styles, forms)
Tag Stripping: Remove tags while preserving text content
Text Normalization: Collapse whitespace, trim text
Link Processing: Make URLs absolute, filter links
Content Deduplication: LRU-based duplicate detection
Markdown Output: Convert HTML to Markdown with structure preservation
Presets: Ready-to-use configurations for common scenarios

Quick Start

use html_cleaning::{HtmlCleaner, presets};
use dom_query::Document;

// Use a preset for quick setup
let cleaner = HtmlCleaner::with_options(presets::standard());

let html = "<html><body><script>bad</script><p>Hello!</p></body></html>";
let doc = Document::from(html);

cleaner.clean(&doc);
// Scripts removed, paragraph content preserved

Installation

Add to your Cargo.toml:

[dependencies]
html-cleaning = "0.3"

With all features:

[dependencies]
html-cleaning = { version = "0.3", features = ["full"] }

Usage Examples

Basic Cleaning

use html_cleaning::{HtmlCleaner, CleaningOptions};

let options = CleaningOptions {
    tags_to_remove: vec!["script".into(), "style".into()],
    prune_empty: true,
    normalize_whitespace: true,
    ..Default::default()
};

let cleaner = HtmlCleaner::with_options(options);

Using the Builder Pattern

use html_cleaning::CleaningOptions;

let options = CleaningOptions::builder()
    .remove_tags(&["script", "style", "noscript"])
    .remove_selectors(&[".advertisement", "#cookie-banner"])
    .prune_empty(true)
    .normalize_whitespace(true)
    .build();

Using Presets

use html_cleaning::presets;

// Minimal: Just scripts and styles
let minimal = presets::minimal();

// Standard: + forms, iframes, objects
let standard = presets::standard();

// Aggressive: + nav, header, footer, aside
let aggressive = presets::aggressive();

// Article extraction: Optimized for content extraction
let article = presets::article_extraction();

Text Processing

use html_cleaning::text;

let has_content = text::has_content("  hello  ");  // true
let normalized = text::normalize("  multiple   spaces  ");  // "multiple spaces"
let words = text::word_count("hello world");  // 2

HTML to Markdown

use html_cleaning::markdown::html_to_markdown;

let html = "<h1>Title</h1><p>Content with <strong>bold</strong></p>";
let md = html_to_markdown(html);
// Output: "# Title\n\nContent with **bold**\n"

Feature Flags

Feature	Default	Description
`presets`	Yes	Include prebuilt cleaning configurations
`regex`	No	Enable regex-based selectors
`url`	No	Enable URL processing with the `url` crate
`markdown`	No	Enable HTML to Markdown conversion
`full`	No	Enable all features

Modules

Module	Description
`cleaner`	Core `HtmlCleaner` and cleaning operations
`text`	Text processing utilities
`tree`	lxml-style text/tail tree manipulation
`dom`	DOM helper utilities
`dedup`	Content deduplication
`presets`	Ready-to-use cleaning configurations
`links`	URL and link processing (feature: `url`)
`markdown`	HTML to Markdown conversion (feature: `markdown`)

Presets Reference

`minimal()`

Removes: script, style, noscript
Best for: Quick sanitization

`standard()`

Removes: script, style, noscript, form, iframe, object, embed, svg, canvas, video, audio
Enables: prune_empty, normalize_whitespace
Best for: General web scraping

`aggressive()`

Includes all of standard() plus:
Removes: nav, header, footer, aside, figure, figcaption
Enables: strip_attributes (preserves href, src, alt)
Best for: Maximum content extraction

`article_extraction()`

Optimized for article content extraction
Removes navigation and layout elements
Strips wrapper tags (div, span) while preserving content
Best for: News articles, blog posts

`trafilatura()`

Full web content extraction cleaning (50 tags removed, 18 stripped)
Comment removal and attribute whitelist cleaning
Empty element pruning across 22 tag types
Best for: Use with rs-trafilatura extraction pipeline

Related Projects

rs-trafilatura - Web content extraction library (uses html-cleaning)
dom_query - DOM manipulation library

License

Licensed under either of:

Apache License, Version 2.0 (LICENSE-APACHE)
MIT license (LICENSE-MIT)

at your option.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

html-cleaning 0.3.0