<?php
/**
 * Accessibility checks on post content: problems the site owner can fix in the editor.
 * They complement (and never replace) a manual review: an automatic check finds only part of the problems.
 *
 * @package Agevole
 */

if ( ! defined( 'ABSPATH' ) ) {
	exit;
}

/**
 * Link texts that do not say where the link goes (WCAG 2.4.4).
 *
 * @return string[]
 */
function agevole_a11y_generic_link_texts() {
	return array(
		'clicca qui',
		'clicca',
		'qui',
		'leggi di più',
		'leggi tutto',
		'scopri di più',
		'scopri',
		'continua',
		'continua a leggere',
		'questo link',
		'link',
		'click here',
		'click',
		'here',
		'read more',
		'more',
		'learn more',
		'this link',
		'continue',
		'continue reading',
	);
}

/**
 * Value of an HTML attribute in a tag, or null when the attribute is missing.
 *
 * @param string $tag  Opening tag.
 * @param string $name Attribute name.
 * @return string|null
 */
function agevole_a11y_attr( $tag, $name ) {
	if ( preg_match( '/\s' . preg_quote( $name, '/' ) . '\s*=\s*(?:"([^"]*)"|\'([^\']*)\'|([^\s>]+))/i', $tag, $m ) ) {
		$value = '';
		foreach ( array( 1, 2, 3 ) as $i ) {
			if ( isset( $m[ $i ] ) && '' !== $m[ $i ] ) {
				$value = $m[ $i ];
				break;
			}
		}
		return html_entity_decode( $value, ENT_QUOTES, 'UTF-8' );
	}
	if ( preg_match( '/\s' . preg_quote( $name, '/' ) . '(?=[\s>\/])/i', $tag ) ) {
		return '';
	}
	return null;
}

/**
 * Checks a fragment of HTML (post content).
 *
 * @param string $html Post content.
 * @return array[] List of issues: array( 'code' => string, 'level' => 'error'|'warning', 'detail' => string ).
 */
function agevole_a11y_check_html( $html ) {
	$issues = array();
	$html   = (string) $html;

	// Images: missing alt attribute (error) or empty alt (fine only for decorative images).
	if ( preg_match_all( '/<img\b[^>]*>/i', $html, $imgs ) ) {
		foreach ( $imgs[0] as $img ) {
			$alt = agevole_a11y_attr( $img, 'alt' );
			$src = (string) agevole_a11y_attr( $img, 'src' );
			$src = wp_basename( (string) wp_parse_url( $src, PHP_URL_PATH ) );
			if ( null === $alt ) {
				$issues[] = array(
					'code'   => 'img_no_alt',
					'level'  => 'error',
					'detail' => $src,
				);
			} elseif ( '' === trim( $alt ) && 'presentation' !== agevole_a11y_attr( $img, 'role' ) ) {
				$issues[] = array(
					'code'   => 'img_empty_alt',
					'level'  => 'warning',
					'detail' => $src,
				);
			}
		}
	}

	// Headings: levels must not be skipped going down (h2 -> h4); h1 is usually the page title.
	if ( preg_match_all( '/<h([1-6])\b/i', $html, $hs ) ) {
		$prev = 1;
		foreach ( $hs[1] as $level ) {
			$level = (int) $level;
			if ( 1 === $level ) {
				$issues[] = array(
					'code'   => 'heading_h1',
					'level'  => 'warning',
					'detail' => 'H1',
				);
			} elseif ( $level > $prev + 1 ) {
				$issues[] = array(
					'code'   => 'heading_skip',
					'level'  => 'error',
					'detail' => 'H' . $prev . ' → H' . $level,
				);
			}
			$prev = $level;
		}
	}

	// Links: empty links and texts that do not describe the destination.
	if ( preg_match_all( '/<a\b([^>]*)>(.*?)<\/a>/is', $html, $links, PREG_SET_ORDER ) ) {
		$generic = agevole_a11y_generic_link_texts();
		foreach ( $links as $link ) {
			$tag = '<a' . $link[1] . '>';
			if ( null === agevole_a11y_attr( $tag, 'href' ) ) {
				continue;
			}
			$label = trim( (string) agevole_a11y_attr( $tag, 'aria-label' ) );
			$text  = trim( preg_replace( '/\s+/u', ' ', html_entity_decode( wp_strip_all_tags( $link[2] ), ENT_QUOTES, 'UTF-8' ) ) );
			if ( '' === $text && preg_match_all( '/<img\b[^>]*>/i', $link[2], $inner ) ) {
				foreach ( $inner[0] as $img ) {
					$text .= trim( (string) agevole_a11y_attr( $img, 'alt' ) );
				}
			}
			if ( '' === $text && '' === $label ) {
				$issues[] = array(
					'code'   => 'link_empty',
					'level'  => 'error',
					'detail' => (string) agevole_a11y_attr( $tag, 'href' ),
				);
				continue;
			}
			$normalized = trim( function_exists( 'mb_strtolower' ) ? mb_strtolower( $text, 'UTF-8' ) : strtolower( $text ), " \t\n\r\0\x0B.:!?»›→>…" );
			if ( '' === $label && in_array( $normalized, $generic, true ) ) {
				$issues[] = array(
					'code'   => 'link_generic',
					'level'  => 'warning',
					'detail' => $text,
				);
			}
		}
	}

	// Data tables without header cells.
	if ( preg_match_all( '/<table\b.*?<\/table>/is', $html, $tables ) ) {
		foreach ( $tables[0] as $table ) {
			if ( ! preg_match( '/<th\b/i', $table ) && 'presentation' !== agevole_a11y_attr( strtok( $table, '>' ) . '>', 'role' ) ) {
				$issues[] = array(
					'code'   => 'table_no_th',
					'level'  => 'warning',
					'detail' => '',
				);
			}
		}
	}

	return $issues;
}

/**
 * Human-readable message for an issue.
 *
 * @param array $issue Issue from agevole_a11y_check_html().
 * @return string
 */
function agevole_a11y_issue_message( $issue ) {
	switch ( $issue['code'] ) {
		case 'img_no_alt':
			/* translators: %s: image file name. */
			return sprintf( __( 'Image without alternative text: %s', 'agevole-accessibility' ), $issue['detail'] );
		case 'img_empty_alt':
			/* translators: %s: image file name. */
			return sprintf( __( 'Image with empty alternative text (fine only if it is decorative): %s', 'agevole-accessibility' ), $issue['detail'] );
		case 'heading_h1':
			return __( 'H1 heading in the content: the page title is usually already the H1.', 'agevole-accessibility' );
		case 'heading_skip':
			/* translators: %s: heading levels, for example "H2 → H4". */
			return sprintf( __( 'Skipped heading level: %s', 'agevole-accessibility' ), $issue['detail'] );
		case 'link_empty':
			/* translators: %s: link address. */
			return sprintf( __( 'Link without text: %s', 'agevole-accessibility' ), $issue['detail'] );
		case 'link_generic':
			/* translators: %s: link text. */
			return sprintf( __( 'Link text does not say where it goes: "%s"', 'agevole-accessibility' ), $issue['detail'] );
		case 'table_no_th':
			return __( 'Table without header cells.', 'agevole-accessibility' );
	}
	return $issue['code'];
}