Readability++

Readability++ is a C++23 port of Mozilla Readability focused on behavioral parity. It has no JavaScript or Node.js dependency, and the algorithm is independent of any concrete HTML parser or DOM implementation.

The current reference is Mozilla Readability commit ab4027a8b37669745016869a37a504727992b2ba. Its complete 130-page fixture corpus passes for content, metadata, and readerability.

Public API

#include <readability/lightweight_dom.hpp>
#include <readability/readability.hpp>

readability::LightweightDomParser parser;
auto document = parser.parse(html, page_url);

readability::Readability reader(*document);
if (auto article = reader.parse()) {
  // article->title, byline, dir, lang, content, text_content, length,
  // excerpt, site_name, and published_time
}

bool likely = readability::is_probably_readerable(*document);

Applications with an existing DOM implement readability::dom::Document and readability::dom::Node, then pass that adapter directly to the algorithm. In particular, a Lexbor-backed application can supply its own adapter without Readability++ linking Lexbor. See the architecture and the audited DOM surface. The complete contracts, options, result fields, and adapter requirements are listed in the API reference.

Readability::parse() mutates the supplied document, matching Mozilla's API. The document must therefore outlive the reader and parse call. Nodes are owned by the document; pointers exposed by the abstraction are non-owning and stable for the document lifetime.

Lightweight parser

LightweightDomParser is the MPL-2.0 port of Mozilla's JSDOMParser.js. It is a reference DOM adapter and allows standalone use for suitable input. Like the upstream parser, it is deliberately not a general HTML5 error-recovering parser:

• input must be properly formed HTML/XML (serialized DOM input is ideal);
• node-list results are snapshots, not live collections;
• it does not execute scripts, load resources, apply CSS, or implement a full browser DOM.

Use a production HTML parser through an adapter for arbitrary web input.

Build and test

cmake -B build
cmake --build build
ctest --test-dir build --output-on-failure

Standalone tests compile with strict warnings and, by default, AddressSanitizer and UndefinedBehaviorSanitizer. The project-specific options are:

READABILITY_BUILD_TESTS        ON when built standalone, OFF as a dependency
READABILITY_BUILD_EXAMPLES     ON when built standalone, OFF as a dependency
READABILITY_ENABLE_SANITIZERS  follows READABILITY_BUILD_TESTS by default

This keeps FetchContent integration lightweight while retaining strict standalone development defaults.

To consume the tagged release directly:

include(FetchContent)
FetchContent_Declare(
  readability
  SYSTEM
  GIT_REPOSITORY https://github.com/L-A-Marchetti/readability-cpp.git
  GIT_TAG v0.1.0
)
FetchContent_MakeAvailable(readability)

target_link_libraries(your_target PRIVATE readability::readability)

Example

The readability_extract_example executable parses a complete HTML document, stores it behind std::unique_ptr<dom::Document>, runs the parser-independent API, and prints the extracted metadata, readable HTML, and readable text:

cmake -B build
cmake --build build --target readability_extract_example
./build/readability_extract_example

For installation:

cmake -B build-release \
  -DREADABILITY_BUILD_TESTS=OFF \
  -DREADABILITY_BUILD_EXAMPLES=OFF \
  -DREADABILITY_ENABLE_SANITIZERS=OFF
cmake --build build-release
cmake --install build-release --prefix /your/prefix

Consumers may then use find_package(readability CONFIG REQUIRED) and link readability::readability.

Parity details and the current result are in docs/parity.md.

Licensing

The algorithm and most of the project are Apache-2.0; see LICENSE.md and NOTICE. The lightweight DOM port is MPL-2.0 and retains its file-level notice. See THIRD_PARTY_NOTICES.md.