// +----------------------------------------------------------------------
declare (strict_types = 1);
namespace addon\articles\service;
/**
* 正文渲染服务（输出期增强，不改动入库内容）
*
* 三件事：
* 1. 图片懒加载：给 ![]()
加 loading="lazy" decoding="async"
* 2. WebP：浏览器支持时把站内图片替换为 .webp 副本（见 ImageService）
* 3. 内链：把命中的关键词（如本文标签）替换为站内链接，利于收录与导流
*
* 全部基于 DOM 操作，不会破坏原有 HTML 结构；DOM 不可用时原样返回。
*/
class ContentRenderService
{
/** 单篇文章最多生成的内链数 */
const MAX_LINKS = 5;
/**
* 渲染正文
*
* @param string $html 已净化的正文
* @param array $linkKeywords [关键词 => 链接地址]
* @param bool $useWebp 是否启用 WebP 替换
*/
public static function render(string $html, array $linkKeywords = [], bool $useWebp = false): string
{
$html = trim($html);
if ($html === '' || !class_exists('DOMDocument')) {
return $html;
}
$prev = libxml_use_internal_errors(true);
$doc = new \DOMDocument('1.0', 'UTF-8');
$loaded = $doc->loadHTML('' . $html);
libxml_clear_errors();
libxml_use_internal_errors($prev);
if (!$loaded) {
return $html;
}
$body = $doc->getElementsByTagName('body')->item(0);
if (!$body) {
return $html;
}
self::enhanceImages($doc, $useWebp);
if (!empty($linkKeywords)) {
self::linkKeywords($doc, $linkKeywords);
}
$out = '';
foreach ($body->childNodes as $child) {
$out .= $doc->saveHTML($child);
}
return trim($out);
}
/**
* 图片增强：懒加载 + WebP
*/
private static function enhanceImages(\DOMDocument $doc, bool $useWebp): void
{
$images = [];
foreach ($doc->getElementsByTagName('img') as $img) {
$images[] = $img;
}
foreach ($images as $img) {
$img->setAttribute('loading', 'lazy');
$img->setAttribute('decoding', 'async');
if (!$useWebp) {
continue;
}
$src = (string) $img->getAttribute('src');
$webp = ImageService::webpUrl($src);
if ($webp !== '') {
$img->setAttribute('src', $webp);
}
}
}
/**
* 内链替换（跳过已在 内的文本，避免链接套链接）
*/
private static function linkKeywords(\DOMDocument $doc, array $keywords): void
{
$xpath = new \DOMXPath($doc);
$nodes = $xpath->query('//text()[not(ancestor::a)]');
if (!$nodes || $nodes->length === 0) {
return;
}
// 先快照，避免替换过程中结构变化影响遍历
$list = [];
foreach ($nodes as $node) {
$list[] = $node;
}
$used = [];
$count = 0;
foreach ($list as $node) {
if ($count >= self::MAX_LINKS) {
break;
}
$text = (string) $node->nodeValue;
if (mb_strlen(trim($text), 'UTF-8') < 2) {
continue;
}
foreach ($keywords as $kw => $url) {
$kw = (string) $kw;
if ($count >= self::MAX_LINKS || isset($used[$kw]) || mb_strlen($kw, 'UTF-8') < 2) {
continue;
}
$pos = mb_stripos($text, $kw, 0, 'UTF-8');
if ($pos === false) {
continue;
}
$before = mb_substr($text, 0, $pos, 'UTF-8');
$after = mb_substr($text, $pos + mb_strlen($kw, 'UTF-8'), null, 'UTF-8');
$a = $doc->createElement('a', htmlspecialchars($kw, ENT_QUOTES, 'UTF-8'));
$a->setAttribute('href', (string) $url);
$a->setAttribute('class', 'article-inner-link');
$frag = $doc->createDocumentFragment();
$frag->appendChild($doc->createTextNode($before));
$frag->appendChild($a);
$frag->appendChild($doc->createTextNode($after));
$parent = $node->parentNode;
if (!$parent) {
continue;
}
$parent->replaceChild($frag, $node);
$used[$kw] = true;
$count++;
break; // 每个文本节点只处理一个关键词
}
}
}
}