article

Documentation

Overview

Package article contains the bounded, text-only rendering used by article extraction. It deliberately accepts a DOM instead of an HTML string so the extractor does not need to serialize and parse its result again.

Index

• Constants
• Variables
• func Convert(root *html.Node) (string, error)
• func ReadLimited(source io.Reader) ([]byte, error)
• func Render(source []byte, contentType, pageURL string) ([]byte, error)

Constants

const (
	// MaxMarkdownDepth is separate from readability's element limit. A deeply
	// nested, otherwise small document must not consume an unbounded Go stack.
	MaxMarkdownDepth = core.MaxNestingDepth
)

Variables

var (
	// ErrNilNode reports an invalid conversion root.
	ErrNilNode = errors.New("article markdown: nil HTML node")
	// ErrMarkdownTooLarge reports output that cannot be retained safely.
	ErrMarkdownTooLarge = errors.New("article markdown: rendered output exceeds the size limit")
)

var ErrUnsupportedContentType = errors.New("article mode requires HTML or Markdown content")

ErrUnsupportedContentType identifies a response that article mode cannot extract. Article mode accepts HTML/XHTML and Markdown responses.

Functions

func Convert

func Convert(root *html.Node) (string, error)

Convert renders root and its descendants as deterministic CommonMark-style Markdown. root may be a document node or any content node. The result has a single trailing newline when it contains content.

The renderer is intentionally conservative. It omits executable and presentation-only elements, treats unknown elements according to their block/inline role, and skips descendants beyond MaxMarkdownDepth.

func ReadLimited

func ReadLimited(source io.Reader) ([]byte, error)

ReadLimited reads at most the configured decoded article-body limit. It does not retain a sentinel byte when the input is exactly at the limit.

func Render

func Render(source []byte, contentType, pageURL string) ([]byte, error)

Render extracts readable HTML or adds article frontmatter to Markdown. pageURL is the final response URL. The returned bytes are uncolored and are suitable for files, pipes, and terminal presentation.